Skip to content

AN INTERACTIVE HISTORY

Attention, technology
and meaning.

From early psychology to AI, explore the ideas and tools behind how we focus, find meaning and make choices.

Explore the timeline
1879–2025 · 3 connected histories · 51 milestonesSources & credits ↓

Reading the connections

Three histories, connected.

Cognitive psychology is unscatter’s primary scientific reference: the study of mental processes such as attention, memory and thinking. Cognitive ergonomics and human–computer interaction (HCI) connect that research with the design of tools, tasks and interfaces.

These tracks are interconnected histories, rather than competing measures of progress. Cognitive science already includes both psychology and AI. Technology can create new capabilities and introduce new demands; its effects depend on the design, the task and the person.

Language, meaning and motivation add a third perspective: how people describe experience, form goals and understand what matters to them.

Understanding attention and agencyDigital technology and AILanguage, meaning and motivation

Follow the threads

Explore the timeline

Three connected histories. Select a milestone to explore it, or follow a connection between the histories.

Drag the graph, or scroll or swipe right. Select a milestone to explore it. Key milestone

Ludwig Wittgenstein, 1922

“We’re going to use our fingers.”

  1. The Principles of Psychology William James
  2. Psychodiagnostik: Methodik und Ergebnisse eines wahrnehmungsdiagnostischen Experiments (Deutenlassen von Zufallsformen) Hermann Rorschach
  3. Tractatus Logico-Philosophicus Ludwig WittgensteinEnglish–German edition
  4. Studies of Interference in Serial Verbal Reactions J. Ridley Stroop
  5. On Computable Numbers, with an Application to the Entscheidungsproblem Alan M. Turing
  6. A Theory of Human Motivation Abraham H. Maslow
  7. A logical calculus of the ideas immanent in nervous activity Warren S. McCulloch and Walter Pitts
  8. Man’s Search for Meaning Viktor E. FranklOriginal German publication; English title, 1962.
  9. Cybernetics Norbert Wiener
  10. A Mathematical Theory of Communication Claude E. Shannon
  11. Philosophical Investigations (first published 1953; linked fourth edition) Ludwig Wittgenstein
  12. A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence John McCarthy, Marvin Minsky, Nathaniel Rochester and Claude Shannon
  13. The Magical Number Seven, Plus or Minus Two George A. Miller
  14. Syntactic Structures Noam Chomsky
  15. Perception and Communication Donald E. Broadbent
  16. Man-Computer Symbiosis J. C. R. Licklider
  17. How To Do Things With Words: The William James Lectures delivered at Harvard University in 1955 (first published 1962; linked second edition) J. L. Austin
  18. Augmenting Human Intellect: A Conceptual Framework Douglas C. Engelbart
  19. Understanding Media: The Extensions of Man Marshall McLuhanRead the opening chapter
  20. Touch display—a novel input/output device for computers E. A. Johnson
  21. ELIZA—A Computer Program for the Study of Natural Language Communication Between Man and Machine Joseph Weizenbaum
  22. Cognitive Psychology Ulric Neisser
  23. Designing Organizations for an Information-Rich World Herbert A. Simon
  24. Effects of Externally Mediated Rewards on Intrinsic Motivation Edward L. Deci
  25. Working Memory Alan D. Baddeley and Graham Hitch
  26. Judgment under Uncertainty: Heuristics and Biases Amos Tversky and Daniel Kahneman
  27. Metacognition and Cognitive Monitoring John H. Flavell
  28. The Psychology of Human-Computer Interaction Stuart K. Card, Thomas P. Moran and Allen Newell
  29. Ironies of Automation Lisanne Bainbridge
  30. Intrinsic Motivation and Self-Determination in Human Behavior Edward L. Deci and Richard M. Ryan
  31. Learning representations by back-propagating errors David E. Rumelhart, Geoffrey E. Hinton and Ronald J. Williams
  32. Cognitive Load During Problem Solving: Effects on Learning John Sweller
  33. Information Management: A Proposal Tim Berners-Lee
  34. Flow: The Psychology of Optimal Experience Mihaly Csikszentmihalyi
  35. Implementation Intentions: Strong Effects of Simple Plans Peter M. Gollwitzer
  36. Social Cognitive Theory: An Agentic Perspective Albert Bandura
  37. Company Info: Launch of News Feed Facebook / Meta
  38. Apple Reinvents the Phone with iPhone (product announcement) Apple
  39. The Cost of Interrupted Work: More Speed and Stress Gloria Mark, Daniela Gudith and Ulrich Klocke
  40. Why is it so hard to do my work? The challenge of attention residue when switching between work tasks Sophie Leroy
  41. ImageNet Classification with Deep Convolutional Neural Networks Alex Krizhevsky, Ilya Sutskever and Geoffrey E. Hinton
  42. Deep Neural Networks for YouTube Recommendations Paul Covington, Jay Adams and Emre Sargin
  43. The three meanings of meaning in life: Distinguishing coherence, purpose, and significance Frank Martela and Michael F. Steger
  44. Cognitive Offloading Evan F. Risko and Sam J. Gilbert
  45. Attention Is All You Need Ashish Vaswani and colleagues
  46. iOS 12 introduces new features to reduce interruptions and manage Screen Time Apple
  47. Guidelines for Human-AI Interaction Saleema Amershi and colleagues
  48. Climbing towards NLU: On Meaning, Form, and Understanding in the Age of Data Emily M. Bender and Alexander Koller
  49. Introducing ChatGPT (public research-preview announcement) OpenAI
  50. Introducing ChatGPT agent: bridging research and action OpenAI
Early foundations1870–1950

Dates mark selected milestones, not the origins of whole disciplines. Spacing groups ideas for reading and is not proportional to elapsed time. Lines indicate sequence only. Image credits

Where the histories meet
Select a milestone above, or open the full comparison below.Read all milestones and connections
01

1870–1950

Early foundations

Experimental psychology develops alongside behavioural, Gestalt and developmental approaches. Formal computation, information theory and cybernetics emerge as language, meaning and human needs are studied through distinct traditions.

  1. Laboratory development

    Attention and agency

    Experimental psychology

    Experimental research begins in Wundt’s Leipzig laboratory

    ★ Key milestone

    Wilhelm Wundt’s Leipzig laboratory helps establish systematic experimental study of mental processes.

    Read context and sources

    Students began experimental work in Leipzig in autumn 1879, including timing studies of apperception. The laboratory developed gradually and gained formal university recognition in 1883–1884. This is an institutional anchor for experimental psychology, whose methods also had earlier physiological and psychophysical roots.

    Why it’s a key milestone Wundt’s Leipzig laboratory helped make the study of mental processes a sustained experimental research programme. Its students carried laboratory methods to other countries, helping establish psychology as an international research discipline. Historical reference ↗

    Related concepts: attention · perception · measurement

    Why this matters What can be learned about attention through observation and careful measurement, rather than intuition alone?

    Sources
  2. Book publication

    Attention and agency

    Attention as selection

    William James, The Principles of Psychology

    William James describes how attending to something means selecting it from other possible experiences.

    Read context and sources

    A chapter in The Principles of Psychology discusses voluntary and involuntary attention, interest and distraction. It is an influential historical account, rather than the beginning of attention research or a complete modern theory.

    unscatter relevance: focus and attention. This is an editorial connection.

    Related concepts: focus and attention

    Why this matters Selection is a useful starting point: making room for one thing often means leaving something else in the background.

    Sources
  3. Book publication / historical psychiatric assessment

    Language, meaning and motivation

    Rorschach inkblot test

    Hermann Rorschach, Psychodiagnostik

    Psychodiagnostik introduced an inkblot-based assessment using ten plates and people’s interpretations of ambiguous forms.

    Read context and sources

    Psychodiagnostik appeared in 1921 in an applied-psychiatry series. The University of Bern preserves the ten cards developed around 1918 that served as masters for the first edition. This is a historical development in psychiatry and psychological assessment. Its connection here with language and meaning is an editorial interpretation through perception and interpretation, rather than a documented influence on linguistic theory or experimental attention research. Viewing this historical image here does not support diagnostic, personality or unconscious inferences.

    Related concepts: perception · interpretation · psychiatry · psychological assessment

    Why this matters How can the same ambiguous form invite different descriptions and interpretations?

    Sources
  4. Book publication

    Language, meaning and motivation

    Language and meaning

    Ludwig Wittgenstein, Tractatus Logico-Philosophicus

    Ludwig Wittgenstein examines how language can represent the world within a philosophical account of logic.

    Read context and sources

    Tractatus Logico-Philosophicus appeared in an English–German edition in 1922, following the German publication of 1921. In proposition 5.6, the Ogden translation states: “The limits of my language mean the limits of my world.” The passage belongs to a philosophical argument about logic and representation.

    unscatter relevance: language and meaning. This is an editorial connection.

    Related concepts: language and meaning

    Why this matters The passage invites reflection on the language used to frame a question or describe a goal.

    Sources
  5. Research publication

    Attention and agency

    Cognitive control

    Stroop’s color–word interference experiments

    ★ Key milestone

    J. Ridley Stroop finds that conflicting written colour names interfere with naming the ink colour.

    Read context and sources

    Studies of Interference in Serial Verbal Reactions compares word reading, colour naming and practice. Conflicting words markedly slow ink-colour naming. The task later becomes a major tool for investigating selective attention and cognitive control; it is not a complete measure of executive function or evidence about every everyday distraction.

    Why it’s a key milestone The Stroop task made conflict between reading a word and naming its ink colour visible and measurable. It became an enduring research tool for investigating selective attention, automatic processing and control over competing responses. Historical reference ↗

    unscatter relevance: executive function, cognitive control. This is an editorial connection.

    Related concepts: executive function · selective attention · interference · cognitive control

    Why this matters Which parts of a task compete with the response someone intends to make?

    Sources
  6. Mathematical research publication

    Technology and AI

    Formal computation

    Turing’s model of a general computing machine

    ★ Key milestone

    Alan Turing formalises calculation through a machine operating on symbols and examines limits on what can be computed.

    Read context and sources

    On Computable Numbers, with an Application to the Entscheidungsproblem describes a mathematical machine, including a universal machine able to simulate other machines. It is a foundation for computability, not a working electronic computer or evidence of human-like understanding. The paper was received and read in 1936; the publisher dates its volume to 1937.

    Dated stages
    1. · Manuscript receivedOn Computable Numbers manuscript received
    2. · Paper readPaper read to the London Mathematical Society
    3. · PublicationProceedings volume 42 publication

    Why it’s a key milestone Turing gave a precise mathematical account of a machine following rules over symbols, including a universal machine that could simulate other such machines. This helped establish the foundations of computer science and clarify the limits of what can be computed. Historical reference ↗

    Related concepts: computation · automation

    Why this matters A specified procedure can support routine work, while deciding what should be computed remains a human and design question.

    Sources
  7. Theoretical publication

    Language, meaning and motivation

    Motivation and needs

    Abraham Maslow, A Theory of Human Motivation

    Abraham Maslow proposes an account of human motivation organized around basic needs.

    Read context and sources

    A Theory of Human Motivation describes physiological needs, safety, love, esteem and self-actualisation as related motives with differing priorities. Maslow discusses partial satisfaction and exceptions to their ordering. The familiar pyramid is a later teaching convention, not a diagram drawn in his 1943 paper; influential historical theories should not be treated as current consensus.

    unscatter relevance: needs, motivation. This is an editorial connection.

    Related concepts: needs · motivation

    Why this matters Attention takes place within wider needs for security, belonging and purpose.

    Sources
  8. Mathematical model / research publication

    Technology and AI

    Early neural modelling

    McCulloch and Pitts’ logical model of neural networks

    McCulloch and Pitts use simplified all-or-none units and propositional logic to model networks of neural activity.

    Read context and sources

    A logical calculus of the ideas immanent in nervous activity describes how networks of idealised units can implement logical relationships. It contributed to early neural-network research; it is not a complete biological model of cognition or a demonstration of modern trained deep learning.

    Related concepts: neural networks · models

    Why this matters Models deliberately simplify a system. Their usefulness should be judged alongside the assumptions and aspects they leave out.

    Sources
  9. Book / memoir

    Language, meaning and motivation

    Meaning and purpose

    Viktor Frankl’s memoir and account of meaning

    Viktor Frankl reflects on imprisonment, human response and the search for meaning.

    Read context and sources

    The German book later known in English as Man’s Search for Meaning appeared in 1946. Frankl reflects on his experiences in Nazi concentration camps and the place of meaning in human life, a concern central to his logotherapy. The expanded English work received this title in 1962.

    unscatter relevance: meaning, purpose. This is an editorial connection.

    Related concepts: meaning · purpose

    Why this matters Purpose can frame what someone chooses to attend to and act on.

    Sources
  10. Book publication

    Technology and AI

    Cybernetics

    Norbert Wiener’s study of feedback, control and communication

    Norbert Wiener brings feedback and communication into a shared study of control in animals and machines.

    Read context and sources

    Cybernetics: Or Control and Communication in the Animal and the Machine connects engineering and biological examples through feedback. A feedback loop returns information about an output so subsequent actions can be adjusted.

    Related concepts: feedback · control · automation

    Why this matters Tools need feedback that lets people notice what happened after an action.

    Sources
  11. Research publication

    Technology and AI

    Information theory

    Claude Shannon, A Mathematical Theory of Communication

    ★ Key milestone

    Claude Shannon develops a mathematical account of information, noise and the capacity of communication channels.

    Read context and sources

    A Mathematical Theory of Communication measures information through the uncertainty among possible messages. It analyses limits on reliable communication in the presence of noise. This engineering use of information concerns message transmission rather than semantic meaning.

    Why it’s a key milestone Shannon established a mathematical framework for measuring information and determining how reliably messages can pass through a noisy channel. The work founded information theory and gave engineers limits and principles for digital communication and error-correcting codes. Historical reference ↗

    Related concepts: information theory · communication

    Why this matters A recipient’s goals still matter when deciding what information is useful.

    Sources
02

1950–1970

The cognitive turn

Information-processing models of attention and memory gain influence alongside early AI and interactive computing. Generative grammar, language in use, media and human motivation broaden the questions being explored.

  1. Posthumous book publication

    Language, meaning and motivation

    Language in use

    Philosophical Investigations

    Ludwig Wittgenstein examines language through its uses in shared activities and varied contexts.

    Read context and sources

    Published posthumously in 1953, Philosophical Investigations investigates language-games, rules and ordinary uses of words. Section 43 relates meaning to use for a large class of cases, with an explicit qualification. This is distinct from the earlier Tractatus and is a philosophical investigation, not an experimental model of attention.

    unscatter relevance: language in context. This is an editorial connection.

    Related concepts: meaning · language-games · context · communication · language in context

    Why this matters How does the situation in which a word is used change what someone understands by it?

    Sources
  2. Research proposal and meeting

    Technology and AI

    Early artificial intelligence

    AI research proposal and Dartmouth summer meeting

    ★ Key milestone

    A proposal by McCarthy, Minsky, Rochester and Shannon names artificial intelligence; the following summer’s Dartmouth project helps establish a research community.

    Read context and sources

    The proposal is dated 31 August 1955, before the summer 1956 gathering. Its conjecture concerns describing aspects of intelligence precisely enough for machines to simulate them. The meeting became an influential institutional anchor for AI, not a product release or the origin of every idea about intelligent machines.

    Dated stages
    1. · Research proposalMcCarthy, Minsky, Rochester and Shannon submit the proposal naming artificial intelligence
    2. · Research meetingDartmouth Summer Research Project takes place

    Why it’s a key milestone The proposal gave artificial intelligence a shared name and a research ambition; the 1956 gathering helped form a community around it. This became an influential foundation for AI as an organised field of research in universities and laboratories. Historical reference 1 ↗Historical reference 2 ↗

    Related concepts: artificial intelligence · research communities

    Why this matters An ambition to model intelligent activity is not evidence that human and machine cognition are the same.

    Sources
  3. Research publication

    Attention and agency

    Memory and capacity

    George Miller, The Magical Number Seven, Plus or Minus Two

    George Miller examines limits in particular judgement and immediate-memory tasks, and the role of grouping information.

    Read context and sources

    The Magical Number Seven, Plus or Minus Two distinguishes absolute judgement from immediate memory. A chunk is a meaningful grouping learned through experience. Miller warns against treating different tasks as one underlying limit; seven is not a universal limit on attention or a rule for every interface.

    unscatter relevance: cognitive load, working memory. This is an editorial connection.

    Related concepts: cognitive load · working memory

    Why this matters How information is organised and what someone already knows can matter as much as how many items are presented.

    Sources
  4. Book publication

    Language, meaning and motivation

    Generative grammar

    Noam Chomsky, Syntactic Structures

    Noam Chomsky models language through rules that generate grammatical sentence structures.

    Read context and sources

    Syntactic Structures develops formal grammars and transformations for analysing sentences. It asks how a finite set of rules can account for an open-ended range of grammatical sentences.

    Related concepts: generative grammar · formal models of language

    Why this matters Understanding a message involves relationships between words, alongside the words themselves.

    Sources
  5. Book / historical model

    Attention and agency

    Selective attention

    Broadbent’s filter model, with later attenuation research

    Donald Broadbent’s Perception and Communication develops an information-processing account of selective attention.

    Read context and sources

    Broadbent uses a filtering account to address how people select information when processing is limited. It is an influential historical model, developed in a particular research context, rather than a final explanation of every form of attention.

    Dated stages
    1. · Book / historical modelBroadbent develops a filter account in Perception and Communication.
    2. · Research publicationTreisman’s selective-listening work contributes to an attenuation account, rather than an all-or-nothing filter.

    unscatter relevance: selective attention. This is an editorial connection.

    Related concepts: selective attention

    Why this matters The idea of selection invites a practical question: what does a person need to notice to carry out their task?

    Sources
  6. Design proposal

    Technology and AI

    Interactive computing

    J. C. R. Licklider, Man-Computer Symbiosis

    J. C. R. Licklider proposes close cooperation between people and computers in Man-Computer Symbiosis.

    Read context and sources

    Licklider envisages people setting goals and evaluating results while computers handle routine preparation and calculation. This is a proposal for future interactive systems, with explicit technical obstacles, rather than evidence that the envisioned partnership had already been achieved.

    unscatter relevance: human agency, tools as partners. This is an editorial connection.

    Related concepts: human agency · tools as partners

    Why this matters A tool can create room for judgement when it supports the work around a decision and keeps its purpose understandable.

    Sources
  7. Posthumous book publication

    Language, meaning and motivation

    Speech acts

    Austin’s How to Do Things with Words

    J. L. Austin examines how speaking can perform acts such as promising, requesting and warning.

    Read context and sources

    How to Do Things with Words was published in 1962 from Austin’s 1955 William James Lectures. It develops distinctions between saying something, the act performed in saying it, and its consequences. These acts depend on conventions and circumstances. The linked Oxford edition is the revised second edition of 1975, not the first publication.

    unscatter relevance: language, intentional action. This is an editorial connection.

    Related concepts: communication · language in use · context · intentional action · language

    Why this matters Is an interface merely describing something, or also requesting, committing, warning or giving permission?

    Sources
  8. Research framework and public demonstration

    Technology and AI

    Human augmentation

    Augmentation framework and NLS demonstration

    ★ Key milestone

    Douglas Engelbart links people, methods, language and tools in an augmentation framework, later demonstrated through NLS.

    Read context and sources

    Augmenting Human Intellect describes an interacting system of human capabilities, practices and electronic aids in 1962. The 9 December 1968 demonstration combines text editing, hypertext, mouse input and remote collaboration. These are dated stages of a research programme, not proof that every later interface improves attention.

    Dated stages
    1. · Research frameworkAugmenting Human Intellect conceptual framework
    2. · Public demonstrationNLS demonstration combines text editing, hypertext, mouse input and remote collaboration

    Why it’s a key milestone Engelbart presented computers as interactive tools for extending intellectual work. The 1968 NLS demonstration brought mouse input, on-screen text editing, hypertext and remote collaboration together in an influential vision of how people could work with computers. Historical reference ↗

    unscatter relevance: cognitive offloading, human capability. This is an editorial connection.

    Related concepts: cognitive offloading · human capability

    Why this matters Useful technology depends on the surrounding practices and the person’s goals, not only on the capabilities of the machine.

    Sources
  9. Book publication

    Language, meaning and motivation

    Understanding Media

    Marshall McLuhan, Understanding Media: The Extensions of Man

    Marshall McLuhan argues that the form of a medium changes perception and social life alongside its content.

    Read context and sources

    Understanding Media: The Extensions of Man opens with “the medium is the message.” McLuhan examines how technologies change the scale, pace and patterns of human activity. Print, radio and television create different conditions for communication, even when carrying similar subject matter. The book offers a framework for interpreting media, rather than an experimental measure of their effects.

    Light is information without “content,” much as the missile is a vehicle without the additions of wheel or highway. As the missile is a self-contained transportation system that consumes not only its fuel but its engine, so light is a self-contained communication system in which the medium is the message.

    Related concepts: media · communication · perception

    Why this matters Look at how a medium organises attention: its timing, format and demands can shape an experience alongside the information it carries.

    Sources
  10. Research publication / input device

    Technology and AI

    Touchscreens

    E. A. Johnson’s published touch-display research

    E. A. Johnson describes a capacitive touch display as an input/output device for computers.

    Read context and sources

    Touch display—a novel input/output device for computers describes a mechanism in 1965, followed by a fuller account in 1967. This is an early published touchscreen development, decades before the iPhone; it does not establish that later touchscreen products caused a particular psychological effect.

    Related concepts: interfaces · direct manipulation

    Why this matters Interaction methods shape the actions a tool makes possible. Whether they support a task depends on the interface, context and person’s goal.

    Sources
  11. Research publication / conversational program

    Technology and AI

    Early conversational interfaces

    Weizenbaum’s ELIZA paper

    ELIZA uses keyword-triggered decomposition and reassembly rules to produce certain kinds of typed conversation.

    Read context and sources

    ELIZA—A Computer Program for the Study of Natural Language Communication Between Man and Machine describes scripts for transforming input sentences into responses; the DOCTOR script imitates aspects of a nondirective therapeutic conversation. This early conversational interface did not demonstrate human-like understanding. MIT’s archive also preserves a 1965 source-code listing, separate from the January 1966 publication.

    Related concepts: conversation · language understanding

    Why this matters Conversational fluency can encourage people to attribute understanding. Clear limits and appropriate reliance matter alongside an interface’s style.

    Sources
  12. Book publication

    Attention and agency

    Cognitive psychology

    Ulric Neisser’s synthesis of perception, attention and memory

    Ulric Neisser brings perception, attention and memory together in an influential account of mental processing.

    Read context and sources

    Cognitive Psychology surveys how information is transformed and used in thought and memory. It helps establish a shared account of an already developing field. Its historical importance does not make every model in it current scientific consensus.

    unscatter relevance: cognitive psychology. This is an editorial connection.

    Related concepts: cognitive psychology

    Why this matters Cognitive psychology offers a scientific reference for asking how people select, interpret and use information.

    Sources
03

1970–1990

Cognition and interactive tools

Working memory, metacognition and cognitive load become design concerns as personal computing and interactive tools develop. Research on language, goals and psychological needs continues alongside these changes.

  1. Published design argument

    Attention and agency

    Information and scarce attention

    Herbert Simon’s design argument about information-rich organisations

    Herbert Simon argues that information systems should consider the attention required from their recipients.

    Read context and sources

    In Designing Organizations for an Information-Rich World, Simon examines organisations and information processing. More available information creates a need to allocate attention. His organisational argument predates the Web; applying it to today’s feeds is a contemporary interpretation.

    unscatter relevance: digital overload, attention economy. This is an editorial connection.

    Related concepts: digital overload · attention economy

    Why this matters A system may help by selecting relevant information rather than simply delivering more of it.

    Sources
  2. Research publication

    Language, meaning and motivation

    Intrinsic motivation

    Deci’s experiments on externally mediated rewards

    Edward Deci studies how different external rewards affect people’s interest in activities they already find engaging.

    Read context and sources

    Effects of Externally Mediated Rewards on Intrinsic Motivation reports two laboratory experiments and one field experiment. Monetary rewards and verbal reinforcement showed different patterns. These findings helped develop later motivation theory; they do not mean that all rewards undermine motivation or that intrinsic motivation was invented in 1971.

    unscatter relevance: intrinsic motivation. This is an editorial connection.

    Related concepts: motivation · interest · rewards · autonomy · intrinsic motivation

    Why this matters What makes a task worth doing for the person, beyond an external reward?

    Sources
  3. Research model

    Attention and agency

    Working memory

    Baddeley and Hitch’s model; later episodic-buffer refinement

    Alan Baddeley and Graham Hitch develop an account of working memory: temporarily holding and using information.

    Read context and sources

    Working Memory proposes a multi-component account of temporarily holding and using information. Baddeley’s November 2000 paper adds an episodic buffer to the model. The model concerns relationships between storage, processing and integration; it is neither a universal item limit nor a complete account of all cognition.

    Dated stages
    1. · Research modelBaddeley and Hitch publish Working Memory.
    2. · Model refinementBaddeley proposes the episodic buffer as a component for integrating information.

    unscatter relevance: working memory, cognitive load. This is an editorial connection.

    Related concepts: working memory · cognitive load

    Why this matters Holding instructions in mind while carrying out another step creates different demands from merely recognising something on a screen.

    Sources
  4. Research synthesis

    Attention and agency

    Judgement and biases

    Tversky and Kahneman’s heuristics-and-biases synthesis

    Amos Tversky and Daniel Kahneman describe how useful judgement shortcuts can also produce systematic errors.

    Read context and sources

    Judgment under Uncertainty: Heuristics and Biases synthesises research on representativeness, availability, and adjustment from an anchor. Its claim concerns particular judgements under uncertainty. The 1974 article is an earlier research anchor than the popular 2011 book Thinking, Fast and Slow; it does not establish that all thinking fits two simple systems.

    unscatter relevance: self-awareness, decision making. This is an editorial connection.

    Related concepts: decision making · self-awareness · reasoning · metacognition

    Why this matters Which assumptions, memorable examples or starting values might be shaping a choice?

    Sources
  5. Research framework

    Attention and agency

    Metacognition

    John Flavell’s synthesis of metacognition and cognitive monitoring

    John Flavell sets out a framework for metacognition: knowledge and monitoring of one’s own thinking.

    Read context and sources

    In Metacognition and Cognitive Monitoring, Flavell discusses children’s developing awareness of memory and comprehension, connecting knowledge, experiences, goals and strategies. The account distinguishes believing that you understand something from being able to demonstrate that understanding.

    unscatter relevance: metacognition, self-awareness. This is an editorial connection.

    Related concepts: metacognition · self-awareness

    Why this matters Pausing to check what you know, what is uncertain and whether a strategy is working can be part of self-directed activity.

    Sources
  6. Research application / book

    Technology and AI

    HCI and interface design

    Card, Moran and Newell, The Psychology of Human-Computer Interaction

    Card, Moran and Newell apply information-processing psychology to people using interactive computer systems.

    Read context and sources

    The Psychology of Human-Computer Interaction offers models for analysing tasks such as text editing. These are deliberately simplified engineering tools for particular kinds of interaction, not complete descriptions of a person’s experience or values.

    unscatter relevance: cognitive ergonomics, interface design. This is an editorial connection.

    Related concepts: cognitive ergonomics · interface design

    Why this matters Interface design can draw on evidence about human processing while recognising what a model leaves out.

    Sources
  7. Where the histories meet Research application

    From cognitive models to interface design

    Information-processing psychology becomes a practical tool for analysing interaction.

    Read the connection

    Cognitive psychology describes how people take in information, remember it and choose actions. Card, Moran and Newell turn that approach into tools for interface design: they break a task such as editing text into goals and operations, then use simplified models to estimate how someone will carry it out. The connection is the move from explaining cognition to using a model to make practical design decisions. Neisser’s earlier survey represents the wider research tradition; the design models remain approximations.

    Explore the linked milestones · links show all histories

    Sources
  8. Human-factors publication

    Technology and AI

    Automation and human control

    Lisanne Bainbridge, Ironies of Automation

    Lisanne Bainbridge examines how automating routine work can leave people with demanding oversight and exception handling.

    Read context and sources

    Her Automatica paper, Ironies of Automation, focuses on industrial process control, with examples from flight-deck automation. People may still need to monitor systems and take over in unusual conditions, even when routine practice has been reduced. It develops an earlier conference paper, rather than announcing a new product.

    unscatter relevance: human oversight, cognitive control. This is an editorial connection.

    Related concepts: human oversight · cognitive control

    Why this matters Reducing routine effort does not remove the need for skill, feedback and preparation for situations the system cannot handle.

    Sources
  9. Theoretical book publication

    Language, meaning and motivation

    Self-determination

    Deci and Ryan’s Intrinsic Motivation and Self-Determination in Human Behavior

    Edward Deci and Richard Ryan develop self-determination theory, linking motivation with psychological needs and social conditions.

    Read context and sources

    Intrinsic Motivation and Self-Determination in Human Behavior develops a theory of self-determination and competence in 1985, while identifying relatedness as an area needing further integration. Ryan and Deci’s 2000 synthesis explicitly brings together autonomy, competence and relatedness. In this account, autonomy means volition rather than independence; needs and context matter alongside individual effort.

    Dated stages
    1. · Theoretical book publicationSelf-determination theory is articulated in a book
    2. · Research synthesisAutonomy, competence and relatedness are brought together in a synthesis

    unscatter relevance: psychological needs, autonomy. This is an editorial connection.

    Related concepts: psychological needs · autonomy · competence · relatedness · intrinsic motivation

    Why this matters Does a tool support choice, a sense of capability and connection with others?

    Sources
  10. Research publication

    Technology and AI

    Learning representations

    Rumelhart, Hinton and Williams’ backpropagation paper

    David Rumelhart, Geoffrey Hinton and Ronald Williams show how neural networks can learn internal representations by propagating errors backwards.

    Read context and sources

    Their Nature paper, Learning representations by back-propagating errors, describes adjusting the weights of connections in a network to reduce the difference between its output and a desired output. Error signals pass backwards through the layers, allowing hidden units to learn useful features from examples.

    Related concepts: neural networks · representation learning

    Why this matters Learning representations expands what software can infer from examples. This computational process is different from the psychological attention and human agency explored in the other track.

    Sources
  11. Research publication

    Attention and agency

    Cognitive load

    John Sweller’s problem-solving and learning research

    John Sweller investigates how some problem-solving demands can compete with the processes needed to learn.

    Read context and sources

    Cognitive Load During Problem Solving: Effects on Learning examines cognitive load during problem solving and schema acquisition: developing organised knowledge for a domain. Its model and experiments concern learning tasks. Cognitive load should not be used as a synonym for every emotional, social or practical demand.

    unscatter relevance: cognitive load, learning. This is an editorial connection.

    Related concepts: cognitive load · learning

    Why this matters Making a task possible and making it a good learning experience are related but different design questions.

    Sources
  12. Proposal, implementation and software release

    Technology and AI

    World Wide Web

    World Wide Web proposal, implementation and open release

    ★ Key milestone

    The Web develops at CERN from a 1989 information-management proposal into working software and a 1993 public-domain release.

    Read context and sources

    The Web links documents through hypertext over the Internet; it is not the Internet itself. Berners-Lee’s first proposal dates to March 1989, implementation was running by the end of 1990, and CERN released basic Web software into the public domain on 30 April 1993. This development crosses the chapter boundary and is anchored to its first proposal, not to a claim about adoption or attention.

    Dated stages
    1. · Information-management proposalFirst Web proposal at CERN
    2. · Formal project proposalBerners-Lee and Cailliau formalise WorldWideWeb
    3. · ImplementationBrowser/editor and server running by the end of 1990
    4. · Public-domain software releaseCERN releases basic client, server and common-code library

    Why it’s a key milestone The Web brought hypertext and Internet networking together so that documents could link across computers. CERN’s 1993 public-domain release allowed others to use and develop the software freely, helping it grow from a research tool into a global information system. Historical reference 1 ↗Historical reference 2 ↗

    Related concepts: hypertext · open access · networked computing

    Why this matters Links create new routes to knowledge while also creating choices about what to follow next.

    Sources
04

1990–2010

Connected everyday life

The Web, platforms and smartphones expand connected everyday life while research examines agency, interruptions and goal pursuit. Flow, psychological needs and meaning in life offer complementary perspectives on experience.

  1. Book publication

    Language, meaning and motivation

    Flow

    Csikszentmihalyi’s synthesis of research on optimal experience

    Mihaly Csikszentmihalyi describes flow as deep involvement in an activity that is rewarding in itself.

    Read context and sources

    Flow: The Psychology of Optimal Experience describes absorption in activities people pursue for their own sake. Csikszentmihalyi relates these experiences to the fit between perceived challenges and skills, drawing on studies of work, play and creative activity.

    unscatter relevance: flow, intrinsic motivation. This is an editorial connection.

    Related concepts: intrinsic motivation · attention · engagement

    Why this matters Engagement also depends on the relationship between a person’s skills and what an activity demands.

    Sources
  2. Research synthesis

    Attention and agency

    Self-regulation and task initiation

    Gollwitzer’s implementation-intentions synthesis

    Peter Gollwitzer describes plans that link a specified situation to a goal-directed action.

    Read context and sources

    Implementation Intentions: Strong Effects of Simple Plans reviews research on linking anticipated situations with actions, often expressed as if–then plans. It distinguishes intending a goal from planning when and how to begin. Effects depend on goals and circumstances; a plan is not a guarantee of action or a universal remedy for task-initiation difficulties.

    unscatter relevance: task initiation, self-regulation. This is an editorial connection.

    Related concepts: self-regulation · task initiation · goal pursuit · executive function

    Why this matters Could naming a specific situation and one next action make an intention easier to start?

    Sources
  3. Psychological theory / review

    Attention and agency

    Human agency

    Albert Bandura’s agentic perspective in social cognitive theory

    Albert Bandura describes agency through intentions, forethought, self-regulation and reflection, within social conditions.

    Read context and sources

    Social Cognitive Theory: An Agentic Perspective develops a long-running research tradition; agency research did not begin in 2001. Bandura distinguishes personal, proxy and collective agency, recognising that acting effectively also depends on other people and social systems.

    unscatter relevance: human agency. This is an editorial connection.

    Related concepts: human agency

    Why this matters Human agency concerns opportunities to shape action and circumstances, not simply individual willpower or making more choices.

    Sources
  4. Where the histories meet unscatter interpretation

    A partnership needs a human purpose

    Augmentation and agency offer complementary questions about what a tool helps a person do.

    Read the connection

    Licklider and Engelbart describe computers as partners in thinking and problem-solving. Bandura’s account of agency adds the person’s intentions and capacity to influence what happens. Unscatter connects these ideas by asking what makes increased capability useful to the person: a purpose they recognise as their own and room to guide the work. Human purpose is the link between augmentation and agency in this editorial comparison.

    Explore the linked milestones · links show all histories

    Sources
  5. Platform feature release

    Technology and AI

    Social media feeds

    Facebook’s News Feed feature launch

    Facebook introduces News Feed, bringing friends’ activity into a stream on the home page.

    Read context and sources

    News Feed launches on 5 September 2006, as recorded in Meta’s company history. The feed brings updates from people in a social network into one place, changing how users encounter activity without visiting each profile separately.

    Related concepts: platforms · information selection

    Why this matters The way updates are collected and presented shapes what is easy to notice next.

    Sources
  6. Where the histories meet unscatter interpretation

    Whose goal is the design serving?

    A social feed and an account of agency bring the intentions behind an interaction into view.

    Read the connection

    A social feed organises updates into a stream that a person can browse. Fogg’s persuasive-computing framework considers how technology can be designed to influence attitudes or behaviour; Bandura places intentional action at the centre of agency. Unscatter connects these ideas by asking whose intended outcome guides an interaction: the person’s chosen purpose, the platform’s objectives, or both? This is an editorial comparison of a documented design feature and research frameworks.

    Explore the linked milestones · links show all histories

    Sources
  7. Announcement, sale, app ecosystem and adoption

    Technology and AI

    Smartphones and apps

    Smartphone interfaces, apps and measured US adoption

    Smartphone touch interfaces, an app marketplace and growing ownership form a changing mobile ecosystem.

    Read context and sources

    Apple announced iPhone on 9 January 2007, began US sales on 29 June, and opened the App Store on 10 July 2008. Pew’s June 2013 report estimated smartphone ownership at 56% of US adults from interviews in April and May. These are distinct product and adoption stages. iPhone did not invent touchscreens or smartphones, and ownership is not a measure of attention, benefit or harm.

    Dated stages
    1. · Product announcementApple announces iPhone
    2. · Initial US saleiPhone sales begin in the US
    3. · App marketplace launchApple opens App Store with 500 apps
    4. · Adoption report / US adultsPew reports 56% ownership from 17 April–19 May 2013 interviews with 2,252 adults aged 18+

    Related concepts: touch interfaces · apps · US adoption

    Why this matters Carrying communication and information tools can create useful possibilities as well as decisions about when to engage with them.

    Sources
  8. Controlled study / conference paper

    Attention and agency

    Interruptions and task switching

    Workplace observation and a later controlled interruption study

    Gloria Mark, Daniela Gudith and Ulrich Klocke study how interruptions affect work speed and experienced demands.

    Read context and sources

    González and Mark’s 2004 field study observes fragmented work across projects. Mark, Gudith and Klocke’s 2008 controlled study reports faster work under interruptions alongside greater stress, frustration, time pressure and effort. These studies address particular work settings and tasks; they do not supply a universal recovery time for every interruption.

    Dated stages
    1. · Workplace field studyObserved information work moves between multiple projects and interruptions.
    2. · Controlled studyInterrupted participants work faster but report greater subjective demands.

    unscatter relevance: task switching, digital overload. This is an editorial connection.

    Related concepts: task switching · digital overload

    Why this matters Assess a workflow by its demands on the person as well as the time taken to finish.

    Sources
  9. Experimental publication

    Attention and agency

    Attention residue

    Sophie Leroy’s experiments on switching between work tasks

    Sophie Leroy studies attention residue: thoughts about one task persisting after a switch to another.

    Read context and sources

    Why is it so hard to do my work? examines transitions between tasks in two experiments. Unfinished prior tasks can leave attention residue and impair performance on the next task; completion and time pressure also affect the transition.

    unscatter relevance: focus, task switching. This is an editorial connection.

    Related concepts: focus · task switching

    Why this matters Changing the open window is only part of making a transition between tasks.

    Sources
05

2010–2026

Contemporary AI

Deep learning, generative systems and agents renew questions about cognitive effort, human control, language understanding, meaning and motivation. Emerging research remains distinct from established historical landmarks.

  1. Research publication

    Technology and AI

    Deep learning

    Krizhevsky, Sutskever and Hinton’s ImageNet research

    Alex Krizhevsky, Ilya Sutskever and Geoffrey Hinton train a deep network for large-scale ImageNet classification.

    Read context and sources

    ImageNet Classification with Deep Convolutional Neural Networks describes a network trained on labelled images using GPUs. It combines learned visual representations with efficient training and reports a substantial improvement in the ImageNet competition.

    Related concepts: deep learning · visual recognition

    Why this matters Visual recognition can be learned from examples rather than specified entirely through hand-written features.

    Sources
  2. System research publication

    Technology and AI

    Recommendation systems

    Covington, Adams and Sargin’s YouTube system paper

    YouTube researchers describe deep neural networks for selecting candidate videos and ranking recommendations.

    Read context and sources

    In Deep Neural Networks for YouTube Recommendations, Covington, Adams and Sargin document a large deployed recommendation system and its two-stage design. The paper establishes how the described system works; it does not by itself demonstrate long-term psychological effects or the behaviour of every recommendation platform.

    Related concepts: recommendation systems · selection algorithms

    Why this matters Recommendations can help people find material, while the system’s ranking goals influence which possibilities are most visible.

    Sources
  3. Where the histories meet unscatter interpretation

    More information creates a selection problem

    Opening access and ranking what is available address different sides of information abundance.

    Read the connection

    Simon explains why a growing supply of information creates a demand for limited human attention. The Web opens more material to publishing and access; recommendation systems rank large collections so people encounter a selection. The connection is the need to decide what reaches a person as the available information grows. For unscatter, comparing these developments raises a design question: does that selection help someone pursue their own purpose? This is an editorial comparison across the sources.

    Explore the linked milestones · links show all histories

    Sources
  4. Theoretical review

    Language, meaning and motivation

    Meaning in life

    Martela and Steger distinguish coherence, purpose and significance

    Frank Martela and Michael Steger distinguish making sense of life, having direction and experiencing life as valuable.

    Read context and sources

    The three meanings of meaning in life offers a theoretical overview of coherence, purpose and significance, including their differences and connections. It builds on earlier research rather than claiming to invent these concerns. Psychological meaning in life concerns people’s experience of meaningfulness; it is not the same scientific construct as linguistic meaning.

    unscatter relevance: meaning, purpose. This is an editorial connection.

    Related concepts: purpose · coherence · significance · motivation · meaning

    Why this matters Is the question about understanding what is happening, choosing a direction, or recognising what matters?

    Sources
  5. Research review

    Attention and agency

    Cognitive offloading

    Risko and Gilbert’s integrative review

    Evan Risko and Sam Gilbert review how actions and external tools can change a task’s cognitive demands.

    Read context and sources

    Cognitive Offloading reviews behaviours such as using reminders, external records and bodily actions to reduce internal processing demands. It proposes a metacognitive framework for decisions about offloading and examines consequences. The behaviour has ancient roots: 2016 dates this review, not the invention of using tools to think.

    unscatter relevance: cognitive offloading, external memory. This is an editorial connection.

    Related concepts: working memory · metacognition · cognitive load · external memory · cognitive offloading

    Why this matters What is useful to keep in mind, and what can a dependable external aid hold instead?

    Sources
  6. Research preprint / architecture

    Technology and AI

    Transformers

    Vaswani and colleagues, Attention Is All You Need

    Attention Is All You Need introduces a neural-network architecture built around computational attention mechanisms.

    Read context and sources

    The paper first appears on arXiv in June 2017 and is published at NeurIPS that year. Computational attention weights relationships between representations. It is not the same construct as human attention, and the shared name does not show that an AI thinks like a person.

    Related concepts: computational attention · neural architectures

    Why this matters Clear language helps distinguish a useful computational method from claims about human understanding or consciousness.

    Sources
  7. Operating-system tool announcements

    Technology and AI

    Digital wellbeing tools

    Google Digital Wellbeing and Apple Screen Time introductions

    Google and Apple announce controls for viewing device use, managing notifications and setting application limits.

    Read context and sources

    Google announced Digital Wellbeing features on 8 May 2018; Apple announced Screen Time on 4 June 2018. These interface controls make aspects of use visible and adjustable. Their introduction is a technology event, not itself evidence that they improve self-regulation or wellbeing.

    Dated stages
    1. · Tool announcementGoogle announces Digital Wellbeing features at I/O.
    2. · Tool announcementApple announces Screen Time for iOS 12.

    unscatter relevance: self-regulation, digital wellbeing. This is an editorial connection.

    Related concepts: self-regulation · digital wellbeing

    Why this matters Usage information and adjustable defaults give people more ways to shape the conditions around attention.

    Sources
  8. HCI research / design guidelines

    Technology and AI

    Human–AI interaction

    Amershi and colleagues’ published interaction guidelines

    Amershi and colleagues propose 18 guidelines for human-AI interaction, including feedback, correction and user control.

    Read context and sources

    The CHI paper Guidelines for Human-AI Interaction evaluates design guidelines across multiple rounds, including a study with design practitioners. It addresses what AI systems should communicate and how people can respond when they are wrong. Guidelines are design guidance, not a guarantee that a particular AI product preserves agency.

    unscatter relevance: human–AI interaction, human agency. This is an editorial connection.

    Related concepts: human–AI interaction · human agency

    Why this matters Being able to understand, dismiss or correct a system can matter as much as its ability to generate an answer.

    Sources
  9. Where the histories meet Research application

    From usable interfaces to correctable AI

    Human-AI interaction extends the design question from completing a task to handling uncertainty and mistakes.

    Read the connection

    Card, Moran and Newell give designers models for analysing the steps involved in a person’s task. Amershi and colleagues address a further difficulty in AI interfaces: the system may make mistakes or behave unexpectedly. Their guidelines make that uncertainty part of interface design, including ways to dismiss suggestions, correct errors and give feedback. The connection is a shared focus on helping people use a system effectively, extended to helping them recover when it is wrong. This comparison follows a design problem across two different research applications.

    Explore the linked milestones · links show all histories

    Sources
  10. Position paper and later review

    Language, meaning and motivation

    Language understanding in AI

    Bender and Koller’s position paper; Mahowald and colleagues’ later review

    Researchers distinguish success with linguistic form from broader questions about meaning, grounding and language use.

    Read context and sources

    In Climbing towards NLU: On Meaning, Form, and Understanding in the Age of Data (July 2020), Bender and Koller argue that training only on linguistic form cannot establish meaning. Mahowald and colleagues’ review, published online on 19 March 2024, distinguishes formal linguistic competence from functional abilities such as using language in the world. These contributions frame an active debate; neither is settled proof about every AI system or future architecture.

    Dated stages
    1. · Position paperForm and meaning are distinguished in an ACL position paper
    2. · Research review; online publicationA review distinguishes formal and functional linguistic competence

    unscatter relevance: language and meaning, ethical AI. This is an editorial connection.

    Related concepts: meaning · grounding · language models · understanding · human–AI interaction · language and meaning · ethical AI

    Why this matters What does a fluent answer demonstrate, and what evidence is still needed about understanding or reliability?

    Sources
  11. Research-preview launch and staged model releases

    Technology and AI

    Conversational and multimodal AI

    Conversational and multimodal generative AI

    ★ Key milestone

    OpenAI’s ChatGPT research preview makes a conversational language-model interface public; later releases add multimodal capabilities.

    Read context and sources

    ChatGPT launched as a public research preview on 30 November 2022. GPT-4’s March 2023 report describes image and text input with text output; those research capabilities did not all become generally available at once. GPT-4o’s May 2024 announcement began a staged rollout. OpenAI’s organisational history begins with a December 2015 announcement, decades after early AI; neither founding nor company announcements independently establish psychological benefit or human-like understanding.

    Dated stages
    1. · Public research-preview launchOpenAI releases ChatGPT
    2. · Model release / research reportGPT-4 announced; technical report describes image/text inputs and text outputs
    3. · Model announcement / staged rolloutGPT-4o announced; text/image capabilities start rollout, new audio/video modalities were not all publicly available that day

    Why it’s a key milestone ChatGPT made a generative language model publicly accessible through ordinary questions and follow-up conversation. Its spread helped bring generative AI into everyday work, education and public debate; later multimodal interfaces broadened the kinds of information people could use with these systems. Historical reference ↗

    Related concepts: generative AI · multimodal interfaces · language models

    Why this matters Conversation creates a new route to assistance, alongside the need to decide what to delegate and what to check.

    Sources
  12. Where the histories meet unscatter interpretation

    When work changes, responsibility changes too

    Delegating routine work can leave people with the difficult task of checking exceptions.

    Read the connection

    Bainbridge examines how industrial automation can move a person’s work from performing a task to monitoring a system and handling unusual failures. ChatGPT’s launch acknowledges that its answers can be incorrect. Unscatter connects these accounts through the responsibility that remains after work is delegated: someone must judge the result and decide when to intervene. Applying the older human-factors question to language-model use, we ask whether the person has the knowledge, time and feedback needed to check the output.

    Explore the linked milestones · links show all histories

    Sources
  13. Agent product announcement

    Technology and AI

    AI agents

    ChatGPT agent’s announced integration of conversation and tool use

    OpenAI introduces ChatGPT agent, combining research and action through a virtual computer and connected tools.

    Read context and sources

    OpenAI announced ChatGPT agent on 17 July 2025, combining a conversational interface with browsing, code execution and other tools for multi-step tasks. Launch access depended on plan and region. This is one product development within a much older field of autonomous agents, not their invention. Product demonstrations and announcements establish offered functions, not reliable performance in every setting.

    Related concepts: tool use · task delegation · human oversight

    Why this matters Delegation makes the boundaries between a suggestion, an action and a user’s approval more consequential.

    Sources

A broader historical overview

This timeline brings together three perspectives on attention and human agency. The overview maps broader historical developments; the graph explores selected research, ideas and technologies.

PeriodAttention and agencyTechnology and AILanguage, meaning and motivation
Early foundations1870–1950Experimental psychology becomes established, with structuralist and functional approaches investigating experience, attention, memory and learning. Behaviourism, Gestalt psychology and developmental approaches later investigate learning, perception and the organisation of mental activity through differing methods. [8]Formal logic, electromechanical tabulation and electrical communication lay groundwork for information processing. Formal computation, early electronic computers, neural models, information theory and cybernetics establish foundations for computing and AI. [11][3][4]Historical and comparative linguistics develop alongside semantics, semiotics and philosophical accounts of meaning. Structural linguistics and analytic philosophy gain influence, while cultural-historical, needs-based and existential approaches explore human experience. [6][7]
The cognitive turn1950–1970The cognitive revolution advances information-processing models of attention, memory and reasoning, alongside continuing behavioural and developmental research traditions. [2]Electronic computing expands as symbolic and neural AI, time-sharing, interactive systems and early touch interfaces develop in research and industry. [4][3]Generative grammar, psycholinguistics, language-in-use and speech-act approaches expand; humanistic psychology and research on intrinsic motivation gain influence. [6][7]
Cognition and interactive tools1970–1990Working-memory models, metacognition, cognitive load and social cognitive theory develop, while cognitive neuroscience and alternative accounts of cognition expand. [1]Personal computing, graphical interfaces and networks spread; expert systems and renewed connectionist learning become influential in computing and AI. [3][4]Pragmatics, discourse and cognitive linguistics expand; self-efficacy, goal pursuit and self-determination theories develop alongside other approaches to motivation. [6][7]
Connected everyday life1990–2010Cognitive neuroscience and attention-control research expand; situated and distributed cognition, self-regulation, interruptions and cognitive offloading receive increasing attention. [1]The Web and open-source software expand networked computing; search, social platforms, smartphones and statistical machine learning gain influence. [5][4]Usage-based and social approaches complement other linguistic theories; narrative meaning, psychological needs, belonging and meaning-in-life research continue to develop. [6][7]
Contemporary AI2010–2026Research examines cognition in connected environments and human–AI collaboration, including attention, learning, cognitive effort, reliance and the exercise of control. [1][13]Deep learning, Transformers and generative AI expand capabilities; recommendation systems, multimodal interfaces and AI agents become prominent in research and deployment. [4][10][12][14]Language models renew debates about meaning, grounding and understanding, while research on motivation, purpose and social context continues. [9][7]

More to explore

The graph is a selection of 51 main milestones. Additional developments, technical releases and recent research remain available here.

Expanded history and context 33 additional records
Summer 1956 · Researchers gather at DartmouthResearch meeting

· Research meeting

Researchers gather at Dartmouth

The Dartmouth Summer Research Project brings researchers together to investigate machine intelligence.

Context and sources

Dartmouth’s institutional history, The Founding and Future of AI, describes the 1956 gathering as a landmark in the formation of AI as a research field. It follows the 1955 proposal, which had already named artificial intelligence. A research meeting is not a product release or the origin of every idea about intelligent machines.

Why this matters The history of AI includes research programmes and changing aspirations, not just a sequence of consumer products.

Sources
Image credits
“Dartmouth Hall” — Unknown photographer (Commons contributor Urban) · Wikimedia Commons · Public domain
“JOHNNIAC computer” — Adam Schuster from Orlando, FL, USA · Wikimedia Commons · CC BY 2.0
“Herbert A. Simon and Allen Newell chess match” — Unknown original photographer; scan via Paolo Massa · Wikimedia Commons · CC BY-SA 2.0
October 1960 · Attention can attenuate a messageResearch publication

· Research publication

Attention can attenuate a message

Anne Treisman’s listening experiments suggest that unattended information is weakened rather than completely blocked.

Context and sources

In Contextual Cues in Selective Listening, participants followed speech presented to one ear while passages switched between ears. Their occasional responses to the other message led Treisman to propose a graded account of selection, in which context and signal strength affect what gets through.

Why this matters Selection can vary with a task and its meaning. Attention is more flexible than a simple all-or-nothing switch.

Sources
Image credits
“Anne Treisman” — National Science Foundation · Wikimedia Commons · Public domain
“The switched-ear experiment” — unscatter original illustration
“Attenuation rather than blocking” — unscatter original illustration
December 1968 · A display responds to the movement of the headResearch prototype

· Research prototype

A display responds to the movement of the head

Ivan Sutherland describes a head-mounted display that changes a three-dimensional scene as the observer moves.

Context and sources

A Head-mounted Three Dimensional Display describes a system that combines stereoscopic views with head-position sensing to make a computer-drawn scene respond to the observer. This research establishes an important route toward interactive immersive displays.

Why this matters An interface can shape what surrounds our attention as well as what appears on a flat screen.

Sources
Image credits
“Ivan Sutherland at the Computer History Museum” — Dick Lyon · Wikimedia Commons · CC BY-SA 3.0 Cropped.
“NASA virtual-reality equipment demonstration (AC89-0437-9)” — NASA · Wikimedia Commons · Public domain
“VPL Eyephone and Dataglove” — Dave Pape · Wikimedia Commons · Public domain
1973 · Attention and mental effortBook / research model

· Book / research model

Attention and mental effort

Daniel Kahneman relates attention to the allocation of limited mental effort across tasks.

Context and sources

Attention and Effort models attention as processing capacity that can be allocated among activities. Kahneman examines task demands, arousal and interference when activities compete for effort.

Why this matters Consider what a task asks someone to hold, monitor and do at the same time.

Sources
Image credits
“Daniel Kahneman” — nrkbeta · Wikimedia Commons · CC BY-SA 2.0 Cropped.
“Allocating available effort” — unscatter original illustration
“Coordinating two tasks” — unscatter original illustration
1980 · Attention binds visual featuresResearch publication

· Research publication

Attention binds visual features

Anne Treisman and Garry Gelade propose that focused attention helps combine visual features into perceived objects.

Context and sources

A Feature-Integration Theory of Attention proposes a role for attention in combining separately processed features. Experiments compare searches for a single feature with searches for a combination, such as a particular colour and shape.

Why this matters Distinct visual features and their combinations can affect how someone locates a target.

Sources
Image credits
“Anne Treisman” — National Science Foundation · Wikimedia Commons · Public domain
“Feature and conjunction search” — unscatter original illustration
“Binding features into objects” — unscatter original illustration
1984 · Susan Kare gives computer actions a visual languageInterface design / product release

· Interface design / product release

Susan Kare gives computer actions a visual language

Kare’s icons and typefaces help make the Macintosh’s actions and system states recognizable.

Context and sources

Working on the Macintosh from 1983, Susan Kare translated commands into small bitmap drawings and designed readable screen typefaces. Her work reached users with the Macintosh in 1984 and helped establish familiar visual ways to interact with computers. Her later account appears in Interview with Susan Kare.

Why this matters Small visual choices can help people recognize an action, understand feedback and maintain their place in a task.

Sources
Image credits
“Susan Kare, 2019” — Cooper Hewitt, Smithsonian Design Museum · Wikimedia Commons · CC BY 3.0 Cropped.
“Macintosh 128K (transparent background)” — All About Apple Museum; derivative by TDS / Grm wnr · Wikimedia Commons · CC BY-SA 3.0
“Building a bitmap form” — unscatter original illustration
1988 · Everyday design makes actions understandableBook publication

· Book publication

Everyday design makes actions understandable

Donald Norman connects the usability of everyday objects with visibility, mappings and feedback.

Context and sources

The Psychology of Everyday Things examines how visible controls, mappings and feedback help people understand everyday objects. Norman uses familiar failures, such as confusing doors and controls, to show how design shapes the ease of an action. Later editions use the title The Design of Everyday Things.

Why this matters Make the next action and its result easy to recognize.

Sources
Image credits
“Donald Norman at AWF05” — jordanfischer · Wikimedia Commons · CC BY 2.0 Cropped.
“Bell System rotary telephone dial, 1940” — Kbrose · Wikimedia Commons · CC BY-SA 4.0 Cropped.
“Spatial mapping” — unscatter original illustration
30 April 1993 · Web software enters the public domainSoftware release / access

· Software release / access

Web software enters the public domain

CERN makes its World Wide Web software available in the public domain, supporting wider use and development.

Context and sources

CERN’s release removes a barrier to using and developing its Web software. An open release is an enabling milestone, distinct from the earlier proposal and from later measures of public adoption. CERN documents the release in The birth of the Web.

Why this matters Access and interoperability can expand people’s opportunities to publish, connect and learn.

Sources
Image credits
“The 1993 release” — unscatter original illustration
“First Web Server” — User:Coolcaesar at en.wikipedia · Wikimedia Commons · CC BY-SA 3.0 Cropped.
“World Wide Web pins” — Rama & Musée Bolo · Wikimedia Commons · CC BY-SA 2.0 fr Cropped.
April 1998 · When computers are designed to persuadeResearch framework

· Research framework

When computers are designed to persuade

B. J. Fogg examines interactive technologies intended to influence attitudes or behaviour, including their ethical questions.

Context and sources

Persuasive Computers builds on work presented at CHI in 1997. Fogg considers computers as tools, media and social actors, and asks whose intentions a system carries. The framework does not establish that every persuasive feature is effective, harmful or deceptive.

Why this matters A design can help someone pursue a chosen goal, or serve goals they have not chosen. The distinction deserves attention.

Sources
Image credits
“B. J. Fogg” — kris krüg · Wikimedia Commons · CC BY 2.0 Cropped.
“A computer monitor displaying a design system interface with plants on a desk” — Balázs Kétyi · Unsplash
“A woman holding a smart watch in her hand” — Káplár Bálint Áron · Unsplash
November 2000 · Working memory integrates informationResearch model

· Research model

Working memory integrates information

Alan Baddeley adds an episodic buffer to account for integrating different kinds of information in working memory.

Context and sources

The episodic buffer: a new component of working memory? proposes a limited-capacity temporary store that combines information from verbal and visual systems and long-term memory. The revision gives integration a more explicit place in the working-memory model.

Why this matters Following a task can involve bringing words, images and prior knowledge together.

Sources
Image credits
“Binding across kinds of information” — unscatter original illustration
“Temporary integrated storage” — unscatter original illustration
“Connecting with long-term memory” — unscatter original illustration
2001 · Attention becomes a business resourceFurther reading / business synthesis

· Further reading / business synthesis

Attention becomes a business resource

Thomas Davenport and John Beck examine attention as a scarce resource that organisations seek and allocate.

Context and sources

The Attention Economy: Understanding the New Currency of Business examines scarce attention amid abundant information. The authors discuss both capturing other people’s attention and directing attention toward important work.

Why this matters Attention is something organisations compete for and people also need to allocate.

Sources
Image credits
“a row of newspapers sitting on top of a shelf” — Jonathan Gong · Unsplash
“Computer screen showing photo editing software interface” — JJROCHA · Unsplash
“A scarce human resource” — unscatter original illustration
April 2004 · Work unfolds across interrupted projectsField study / conference paper

· Field study / conference paper

Work unfolds across interrupted projects

Victor González and Gloria Mark observe information workers switching repeatedly among related projects and activities.

Context and sources

“Constant, Constant, Multi-tasking Craziness”: Managing Multiple Working Spheres reports field observations of analysts, developers and managers. The authors describe working spheres: larger, thematically connected units of work fragmented across tools and shorter activities.

Why this matters A useful workspace can help someone retain the project context while moving between its tasks.

Sources
Image credits
“Person is working on a laptop and writing in a notebook” — MJ Duford · Unsplash
“Fragmented activity” — unscatter original illustration
“Multiple working spheres” — unscatter original illustration
28 February 2006 · Mindset brings beliefs about learning to a wider audienceFurther reading / popular synthesis

· Further reading / popular synthesis

Mindset brings beliefs about learning to a wider audience

Carol Dweck introduces fixed and growth mindsets to a general readership.

Context and sources

Mindset translates Dweck’s research on beliefs about ability and responses to setbacks into examples from education, work and relationships. It is a popular account of an existing research programme, with applications that depend on the person and setting.

Why this matters Beliefs about improvement can shape how people approach difficult tasks, alongside instruction, resources and social conditions.

Sources
Image credits
“Carol Dweck for Innovation” — Satheesh Gopalan · Wikimedia Commons · CC BY-SA 3.0 Cropped.
“person writing on notepad” — Calum MacAulay · Unsplash
“A young green seedling growing from dark soil” — Pranav Gavali · Unsplash
10 July 2008 · Apps become a distribution platformSoftware marketplace launch

· Software marketplace launch

Apps become a distribution platform

Apple opens the App Store with 500 apps, giving iPhone users a central place to obtain third-party software.

Context and sources

The App Store turns 10, Apple’s later history, dates the launch to 10 July 2008. A shared storefront connects developers with phone users and makes installing additional tools part of everyday device use.

Why this matters A phone becomes a changing collection of tools, each offering new possibilities and demands.

Sources
Image credits
“iPhone and iPhone 3G” — Dan Taylor from London, UK · Wikimedia Commons · CC BY 2.0
“A mobile software marketplace” — unscatter original illustration
“Many uses on one device” — unscatter original illustration
2011 · Fast impressions and effortful thoughtFurther reading / popular synthesis

· Further reading / popular synthesis

Fast impressions and effortful thought

Daniel Kahneman brings research on judgment and choice to a broad audience through a distinction between rapid and effortful thinking.

Context and sources

Thinking, Fast and Slow uses System 1 and System 2 as explanatory shorthand for fast, automatic responses and slower, effortful operations. It examines heuristics and recurring biases in judgment, drawing on Kahneman’s work with Amos Tversky and other researchers.

Why this matters The distinction invites a pause to check assumptions in a difficult decision.

Sources
Image credits
“Daniel Kahneman” — nrkbeta · Wikimedia Commons · CC BY-SA 2.0 Cropped.
“An immediate visual impression” — unscatter original illustration
“Familiar and effortful tasks” — unscatter original illustration
5 June 2013 · A majority of US adults own a smartphoneAdoption research / US adults

· Adoption research / US adults

A majority of US adults own a smartphone

Pew reports smartphone ownership among 56% of US adults, based on interviews in April and May 2013.

Context and sources

The survey behind Smartphone Ownership 2013 covered 2,252 adults aged 18 and over, interviewed from 17 April to 19 May 2013. The measure combines self-identification and reported smartphone platform. It is US adult ownership, not worldwide adoption, time spent using phones or a measure of attention.

Why this matters A launch date tells us when a product appeared. Adoption research tells a different story about who has access.

Sources
Image credits
“Samsung Galaxy S4” — Kārlis Dambrāns · Wikimedia Commons · CC BY 2.0 Cropped.
“A pencil rests on a map” — unavailable parts · Unsplash
“Cell tower” — Yq2Yq · Wikimedia Commons · CC0
1 September 2014 · Translation learns where to lookResearch preprint / later published model

· Research preprint / later published model

Translation learns where to look

Dzmitry Bahdanau, Kyunghyun Cho and Yoshua Bengio introduce a translation model that learns to align input and output words.

Context and sources

Neural Machine Translation by Jointly Learning to Align and Translate lets a model weight relevant parts of an input sentence while producing a translation. This computational attention mechanism reduces reliance on a single fixed-length representation of the whole input.

Why this matters An attention mechanism selects information within a computation; it does not by itself establish human-like attention.

Sources
Image credits
“Colorful letters scattered on a textured surface.” — NHN · Unsplash
“Soft alignment between languages” — unscatter original illustration
“Context for each target word” — unscatter original illustration
4 November 2014 · Product design aims to build habitsCommercial design framework / popular book

· Commercial design framework / popular book

Product design aims to build habits

Nir Eyal presents a framework for designing products that encourage repeated use.

Context and sources

Hooked: How to Build Habit-Forming Products organizes its model around a trigger, an action, a variable reward and an investment. The book translates ideas about habit and engagement into advice for product designers.

Why this matters Understanding a product’s engagement goals can help people recognize what encourages them to return.

Sources
Image credits
“person holding black phone” — ROBIN WORRALL · Unsplash
“The proposed Hook cycle” — unscatter original illustration
“Varying outcomes” — unscatter original illustration
5 January 2016 · Deep Work brings sustained concentration into everyday discussionFurther reading / popular synthesis

· Further reading / popular synthesis

Deep Work brings sustained concentration into everyday discussion

Cal Newport argues for protecting stretches of concentration for cognitively demanding work.

Context and sources

Deep Work distinguishes focused, demanding tasks from fragmented routine work and proposes schedules and rituals for protecting concentration. The book brings attention and interruption into a public discussion about how knowledge work is organised.

Why this matters Attention management includes the conditions around work: time, boundaries, interruptions and the concentration a task needs.

Sources
Image credits
“Protecting a work block” — unscatter original illustration
“Desk with lamp, sketchbook, pencils, and eraser” — Calvin Wise · Unsplash
“Do not disturb toggle switch on black background” — Morgan Housel · Unsplash
27 January 2016 · Learning and search combine in GoResearch publication

· Research publication

Learning and search combine in Go

DeepMind’s AlphaGo combines neural networks and tree search to reach professional-level play in Go.

Context and sources

Mastering the game of Go with deep neural networks and tree search describes policy networks for choosing moves and value networks for evaluating positions. The paper reports a 5–0 result against European Go champion Fan Hui, using networks trained from expert games and self-play.

Why this matters A system can combine learned estimates with an explicit search through alternatives.

Sources
Image credits
“Go stones on a go board with containers.” — Kyle Hinkson · Unsplash
“Learning from games” — unscatter original illustration
“Search guided by neural networks” — unscatter original illustration
11 October 2018 · Language representations use both contextsResearch preprint

· Research preprint

Language representations use both contexts

BERT pre-trains language representations using context on both sides of a word.

Context and sources

BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding learns representations from unlabelled text using left and right context. The pretrained model can then be fine-tuned for tasks such as question answering and language inference.

Why this matters The context around a word contributes to how a language model represents it.

Sources
Image credits
“Masked-language pretraining” — unscatter original illustration
“Bidirectional context” — unscatter original illustration
“Pretraining and fine-tuning” — unscatter original illustration
28 May 2020 · Examples become part of the promptResearch preprint

· Research preprint

Examples become part of the prompt

GPT-3 shows how a large language model can respond to tasks described through instructions and examples in its input.

Context and sources

Language Models are Few-Shot Learners evaluates GPT-3 across tasks using text prompts with demonstrations. In this setting, examples change the input context without updating the model’s weights.

Why this matters Examples and instructions become a practical part of how people steer a model.

Sources
Image credits
“Ilya Sutskever at Democratizing Deep Learning” — Steve Jurvetson · Wikimedia Commons · CC BY 2.0 Cropped.
“Examples in the prompt” — unscatter original illustration
“Generating a continuation” — unscatter original illustration
June 2022 · Image generation works in a compressed spaceConference publication

· Conference publication

Image generation works in a compressed space

Latent diffusion models generate images through a learned compressed representation instead of working only in pixel space.

Context and sources

High-Resolution Image Synthesis With Latent Diffusion Models moves the diffusion process into a learned representation of images. Cross-attention lets generation be conditioned on text and other inputs. The CVPR 2022 paper builds on a preprint first posted in December 2021.

Why this matters A written prompt becomes a control over image synthesis, alongside a model’s learned representations.

Sources
Image credits
“Close-up of a professional camera lens” — Jaimy de Hon · Unsplash
“A learned compressed space” — unscatter original illustration
“Conditioning on language” — unscatter original illustration
10 January 2023 · Attention is studied in everyday digital workFurther reading / popular synthesis

· Further reading / popular synthesis

Attention is studied in everyday digital work

Gloria Mark brings research on technology use, attention switching and stress to a general readership.

Context and sources

Attention Span synthesizes Mark’s field research on people using computers and phones in everyday settings. The book examines interruptions, changing patterns of screen engagement and the rhythms of focused work and breaks.

Why this matters Attention management includes the work setting and its rhythms, alongside an individual’s intentions.

Sources
Image credits
“Library of Congress Main Reading Room” — Carol M. Highsmith · Wikimedia Commons · Public domain
“Awareness and intentional attention” — unscatter original illustration
“Attention and recovery” — unscatter original illustration
14 March 2023 · A multimodal model broadens the interfaceModel announcement

· Model announcement

A multimodal model broadens the interface

OpenAI introduces GPT-4 and publishes evaluations of a model that can process text and images.

Context and sources

GPT-4 is introduced as a model accepting text and image inputs and producing text. The accompanying technical report describes pretraining, post-training and benchmark evaluations, alongside limitations.

Why this matters An image can become part of a request, making the interpretation of shared context more important.

Sources
Image credits
“photo of computer cables” — Kvistholt Photography · Unsplash
“Text and image context” — unscatter original illustration
“Evaluating a response” — unscatter original illustration
13 May 2024 · Voice, vision and text share a modelModel announcement

· Model announcement

Voice, vision and text share a model

OpenAI introduces GPT-4o as a model trained across text, vision and audio for more immediate interaction.

Context and sources

Hello GPT-4o describes one model trained across text, vision and audio. It presents voice and visual interaction with shorter response times. The May 2024 announcement describes a staged rollout beginning with text and image input and text output.

Why this matters Voice and visual input change how people make requests and share context.

Sources
Image credits
“HTC One on an Apple keyboard” — Kārlis Dambrāns · Wikimedia Commons · CC BY 2.0 Cropped.
“Text, vision and audio” — unscatter original illustration
“Continuing the conversation” — unscatter original illustration
1 August 2024 · AI oversight becomes a regulatory questionRegulatory context

· Regulatory context

AI oversight becomes a regulatory question

The European Union’s AI Act enters into force, establishing a risk-based framework for AI governance.

Context and sources

AI Act enters into force records the law’s entry into force on 1 August 2024. The framework addresses AI systems and general-purpose AI models, with requirements organized around risks and uses. Application of its provisions is phased rather than simultaneous.

Why this matters Transparency and human oversight become questions for governance as well as interface design.

Sources
Image credits
“Entry into force” — unscatter original illustration
“A risk-based framework” — unscatter original illustration
“Human oversight” — unscatter original illustration
12 September 2024 · Models use more computation before answeringModel preview

· Model preview

Models use more computation before answering

OpenAI introduces o1-preview, a model trained to use additional computation on difficult reasoning problems.

Context and sources

Introducing OpenAI o1-preview describes reinforcement learning that encourages intermediate reasoning before a final answer. The release focuses on tasks in mathematics, science and coding.

Why this matters A response can involve a trade-off between waiting time and the computation used to produce it.

Sources
Image credits
“Chessboard with pieces and a digital clock” — Michal Vrba · Unsplash
“Additional inference work” — unscatter original illustration
“The computation trade-off” — unscatter original illustration
25 November 2024 · Assistants connect through a shared protocolProtocol announcement

· Protocol announcement

Assistants connect through a shared protocol

Anthropic introduces an open protocol for connecting AI applications with data sources and tools.

Context and sources

Introducing the Model Context Protocol describes a shared interface between AI applications and servers that expose data or tools. Its client–server design lets compatible connections be reused instead of rebuilding every integration separately.

Why this matters Connections and permissions shape what an assistant can see and do.

Sources
Image credits
“Clients and servers” — unscatter original illustration
“Reading and acting” — unscatter original illustration
“Scoped access” — unscatter original illustration
20 January 2025 · Reasoning models become available as weightsModel release

· Model release

Reasoning models become available as weights

DeepSeek releases R1 and smaller distilled models with publicly available weights.

Context and sources

DeepSeek-R1 Release makes model weights available and links a technical report on reinforcement learning for reasoning. The report distinguishes R1, which combines reinforcement learning with supervised training stages, from the R1-Zero experiment trained without initial supervised fine-tuning.

Why this matters Access to model weights broadens the ways researchers can compare, adapt and deploy systems.

Sources
Image credits
“multicolored abacus” — Crissy Jarvis · Unsplash
“R1 and the R1-Zero experiment” — unscatter original illustration
“Training smaller models” — unscatter original illustration
April 2025 · AI confidence and critical-thinking effortEmerging research / survey / conference paper

· Emerging research / survey / conference paper

AI confidence and critical-thinking effort

A Microsoft and Carnegie Mellon study surveys how knowledge workers describe critical thinking when using generative AI.

Context and sources

The Impact of Generative AI on Critical Thinking surveys 319 knowledge workers about their own work practices. Higher confidence in AI is associated with less reported critical-thinking effort, while the work of checking and integrating responses remains important. The findings describe self-reported practices and associations.

Why this matters Confidence in an answer and the work of evaluating it deserve separate attention.

Sources
Image credits
“A self-report survey” — unscatter original illustration
“Confidence and effort” — unscatter original illustration
“Verification and integration” — unscatter original illustration
28 January 2026 · Assistance and learning can divergeEmerging research / randomized study / preprint

· Emerging research / randomized study / preprint

Assistance and learning can diverge

Judy Hanwen Shen and Alex Tamkin test how AI assistance affects learning an unfamiliar programming library.

Context and sources

How AI Impacts Skill Formation randomly assigns developers to learn with or without AI assistance. In this task, the assisted group scores lower on a subsequent comprehension quiz, while the average time difference is not statistically significant. The assessment measures learning shortly after the task.

Why this matters Using assistance while learning requires attention to what the learner actually practises.

Sources
Image credits
“Random assignment” — unscatter original illustration
“What was assessed” — unscatter original illustration
“Observed interaction patterns” — unscatter original illustration
3 February 2026 · Evidence about AI risks is brought togetherContext / international evidence assessment

· Context / international evidence assessment

Evidence about AI risks is brought together

An international expert group synthesizes research on general-purpose AI capabilities, risks and risk management.

Context and sources

International AI Safety Report 2026 reviews advances in general-purpose models and evidence about misuse, malfunctions and loss of control. It compares approaches to risk management and identifies where evidence remains limited. The report provides a shared assessment to support policy decisions.

Why this matters Deployment decisions need evidence about capabilities and the practical limits of oversight.

Sources
Image credits
“International AI Safety Report 2026” — Department for Science, Innovation and Technology, © Crown 2026 · International AI Safety Report · Open Government Licence v3.0
“Three assessment questions” — unscatter original illustration
“Layers of risk management” — unscatter original illustration
Further reading 54 publications

Foundational works and later syntheses, including popular books. Their publication dates do not establish when an underlying idea was discovered.

  1. Tractatus Logico-Philosophicus ↗

    Ludwig Wittgenstein

    English–German edition
  2. Man’s Search for Meaning ↗

    Viktor E. Frankl

    Original German publication; English title, 1962.
  3. Ironies of Automation ↗

    Lisanne Bainbridge

  4. Deep Neural Networks for YouTube Recommendations ↗

    Paul Covington, Jay Adams and Emre Sargin

  5. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding ↗

    Jacob Devlin, Ming-Wei Chang, Kenton Lee and Kristina Toutanova

  6. Guidelines for Human-AI Interaction ↗

    Saleema Amershi and colleagues

  7. High-Resolution Image Synthesis With Latent Diffusion Models ↗

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser and Björn Ommer

  8. How AI Impacts Skill Formation ↗

    Judy Hanwen Shen and Alex Tamkin

  9. International AI Safety Report 2026 ↗

    Yoshua Bengio and the international writing group

Historical reference sources

Choose a numbered account to read a short summary and selected further reading. Account numbers match the citations in the historical overview. Further-reading numbers refer to the full bibliography.

  1. [1]Cognitive science: history and approaches

    Encyclopedia overview

    Thagard surveys cognitive science as a meeting point for psychology, neuroscience, linguistics, philosophy and AI. He explains how experiments and computational models investigate the mind, compares approaches including rules, neural networks and Bayesian models, and discusses challenges to treating cognition mainly as representation and computation.

    Selected further reading

  2. [2]The cognitive revolution: a historical perspective

    Personal historical review

    Miller looks back on the emergence of cognitive science in the 1950s, when psychology, linguistics and computing increasingly addressed shared questions about the mind. His account describes the move beyond behaviourism and the collaborations that followed; it is a participant’s historical perspective rather than a single experimental finding.

    Selected further reading

  3. [3]Computer History Museum timeline

    Museum chronology

    The Computer History Museum’s illustrated chronology follows computing through relay machines, electronic computers, commercial systems and personal devices. Individual entries explain what particular machines made possible and who developed them, offering historical context for changes in information processing without implying that technical progress alone explains changes in human attention.

    Selected further reading

  4. [4]A short history of AI

    Historical synthesis

    Stanford’s 2016 AI100 report traces AI from early ideas about computation and machine intelligence through symbolic reasoning, neural networks, expert systems and data-driven learning. It explains how different approaches rose, encountered limitations and combined, providing historical context for AI’s development rather than evidence that machine and human intelligence are equivalent.

    Selected further reading

  5. [5]CERN history of the Web

    Institutional history

    CERN recounts how Tim Berners-Lee developed the Web to help scientists share information, beginning with his 1989 proposal and the first browser and server. The account follows early software releases, wider browser adoption and open standards, showing how a research communication project became a widely shared information system.

    Selected further reading

  6. [6]Philosophy of linguistics

    Philosophy and methodology overview

    This encyclopedia entry compares approaches that study linguistic patterns, language in social and cognitive activity, and abstract grammatical principles. It examines disagreements about evidence, explanation and language acquisition, helping readers understand why linguistic theories differ. Its focus is the philosophy of linguistic science rather than a complete history of language research.

    Selected further reading

  7. [7]Self-determination, motivation and social context

    Theory and research review

    Ryan and Deci review how social conditions can support or undermine motivation, self-regulation and well-being. Their self-determination framework centres on autonomy, competence and relatedness, distinguishing activity driven by interest or personally endorsed values from externally controlled behaviour. It presents a research-based account of motivation rather than a universal recipe for productivity.

    Selected further reading

  8. [8]Major paradigms and approaches in psychology

    Historical book chapter

    Benjafield traces psychology’s emergence as an independent discipline through competing approaches to experience, behaviour and mental activity. He contrasts introspectionism, functionalism, behaviourism, Gestalt psychology, psychoanalysis and cognitive psychology, emphasizing that disagreements were not resolved by one final winning school and that researchers often adapt methods to the problem they study.

    Selected further reading

  9. [9]Language and thought in large language models

    Research review

    Mahowald and colleagues distinguish knowledge of linguistic patterns from the ability to understand and use language in the world. Drawing on language-model research and human neuroscience, they argue that strong performance on linguistic form does not by itself establish wider thinking abilities, and discuss systems that combine language models with other mechanisms.

    Selected further reading

  10. [10]The Transformer architecture

    Architecture and benchmark experiments

    Vaswani and colleagues introduce the Transformer, a neural architecture that uses attention mechanisms without the recurrent or convolutional layers common in earlier sequence models. Their translation experiments report strong results with greater parallelization and lower training costs. Here, “attention” names a computational mechanism, not evidence of human-like awareness or attention control.

    Selected further reading

  11. [11]Earlier computing and communication foundations

    Museum exhibition

    The museum’s Revolution exhibition places electronic computers within a much longer history of tools for calculation and information processing. Its objects and first-hand accounts connect devices such as the abacus, mechanical calculators and punched-card tabulators with later computers, showing that modern computing emerged through many inventions, practical needs and contributors.

    Selected further reading

  12. [12]AI agents: a product announcement

    Company product announcement

    OpenAI’s July 2025 announcement describes a ChatGPT system that combines reasoning with browser, terminal and other tools to carry out multi-step tasks. It also describes user controls and risks when AI can act through software. This is a company’s account of a product launch, not independent evidence of its effectiveness.

    Selected further reading

  13. [13]AI and cognitive effort: an emerging survey study

    Empirical self-report survey

    Lee and colleagues surveyed 319 knowledge workers about generative AI use and critical thinking. Greater confidence in AI was associated with less reported critical-thinking effort, while greater confidence in one’s own ability was associated with more. The findings concern reported experiences and associations; they do not establish that AI causes lasting declines in thinking ability.

    Selected further reading

  14. [14]AI capabilities and adoption: the 2025 AI Index

    Annual evidence synthesis

    Stanford’s 2025 AI Index brings together data on AI research, technical performance, investment, adoption, policy and public opinion. It helps place individual breakthroughs within broader trends, while showing that progress differs across tasks and measures. As an annual synthesis, it describes its reporting period rather than providing a live measure of AI capability.

    Selected further reading

The question running through it all

A blue robotic arm reaching with an open hand

What room is left for human choice?

Human agency is our capacity to act intentionally and influence what happens. It depends on our circumstances and the support available to us, as well as our skills.

For unscatter, this raises a practical question: how can technology support people’s own goals and choices?

  • Whose goals does this tool serve?
  • Can you understand, question or reject its suggestions?
  • What support or change would give you more room to choose?
Explore human agency
Image credits 230 pictures
  1. 1879 · Experimental psychology
    Graph image · “Corsi block-tapping test, Figure 1A” — Åsa Bartonek, Cecilia Guariglia and Laura Piccardi · Frontiers in Psychiatry · CC BY 4.0 Cropped.
    “Measuring a response” — unscatter original illustration
    “Comparing conditions” — unscatter original illustration
    “Observing across trials” — unscatter original illustration
  2. 1890 · Attention as selection
    Graph image · “William James in the 1890s, photographed by Sarah Choate Sears.” — Sarah Choate Sears (1858-1935) · Wikimedia Commons · Public domain
    “person holding black round camera lens during sunset” — Danny Lines · Unsplash
    “Filtering information” — unscatter original illustration
    “Choosing a focus” — unscatter original illustration
  3. 1921 · Rorschach inkblot test
    Graph image · “Rorschach blot 01” — Hermann Rorschach · Wikimedia Commons · Public domain
    “Hermann Rorschach c.1910” — Unknown photographer · Wikimedia Commons · Public domain (Commons)
    “Face or vase ata 01” — Nevit Dilmen · Wikimedia Commons · CC BY-SA 3.0 Artwork unchanged; labelled layout by unscatter.
    “Description and interpretation” — unscatter original illustration
  4. 1922 · Language and meaning
    “Tractatus Logico-Philosophicus, 1922 title page” — Harcourt, Brace (1922 title-page layout) · Wikimedia Commons · Public domain
    “Ludwig Wittgenstein, 1929” — Moritz Nähr · Wikimedia Commons · Public domain
    “Language and a possible world” — unscatter original illustration
  5. December 1935 · Cognitive control
    Graph image · “Cognitive control” — unscatter original illustration
    “Matching and conflicting examples” — unscatter original illustration
    “The intended ink-colour response” — unscatter original illustration
    “Competing responses” — unscatter original illustration
  6. 1936–1937 · Formal computation
    Graph image · “getty-images-NIak9LqUP9c-unsplash” — Getty Images · Unsplash+ · Unsplash+ License
    “Symbols, tape and a rule” — unscatter original illustration
    “An illustrative procedure and checking” — unscatter original illustration
    “JOHNNIAC computer” — Adam Schuster from Orlando, FL, USA · Wikimedia Commons · CC BY 2.0
  7. 1943 · Motivation and needs
    Graph image · “Motivation and needs” — unscatter original illustration
    “Five proposed need categories” — unscatter original illustration
    “green and orange vegetables on black plastic container” — Beth Macdonald · Unsplash
    “Multiple motives at once” — unscatter original illustration
  8. December 1943 · Early neural modelling
    Graph image · “a group of computers” — Rose Pilkington / Google DeepMind · Unsplash
    “Mark I Perceptron (operator’s manual, Figure 2)” — John C. Hay, Albert E. Murray · Wikimedia Commons · Public domain
    “An idealised threshold unit” — unscatter original illustration
    “Logical relationships in a network” — unscatter original illustration
  9. 1946 · Meaning and purpose
    Graph image · “benaja-germann-En8F8zc4LDQ-unsplash.jpg” — Benaja Germann · Unsplash
    “Viktor Frankl” — Prof. Dr. Franz Vesely · Wikimedia Commons · CC BY-SA 3.0 de Cropped.
    “Routes to meaning” — unscatter original illustration
    “Meaning in a concrete situation” — unscatter original illustration
  10. 1948 · Cybernetics
    Graph image · “Cybernetics: Or Control and Communication in the Animal and the Machine” — The MIT Press · The MIT Press
    “Norbert Wiener” — Konrad Jacobs · Wikimedia Commons · CC BY-SA 2.0 de
    “Boulton and Watt centrifugal governor” — Andy Dingley · Wikimedia Commons · CC BY 3.0
    “Feedback and control” — unscatter original illustration
  11. 1948 · Information theory
    Graph image · “white and blue light on dark room” — Denny Müller · Unsplash
    “Claude Elwood Shannon (Tekniska museet 43069)” — Unknown photographer / Tekniska museet · Wikimedia Commons · CC BY 2.0 Cropped.
    “A communication system” — unscatter original illustration
    “Information in bits” — unscatter original illustration
  12. 1953 · Language in use
    Graph image · “Colorful letters scattered on a textured surface.” — NHN · Unsplash
    “a row of newspapers sitting on top of a shelf” — Jonathan Gong · Unsplash
    “Meaning in everyday use” — unscatter original illustration
    “person writing on notepad” — Calum MacAulay · Unsplash
  13. 1955–1956 · Early artificial intelligence
    Graph image · “Uma tela de computador com a palavra AI” — Philip Oroni · Unsplash+ · Unsplash+ License
    “John McCarthy at Stanford” — null0 · Wikimedia Commons · CC BY-SA 2.0 Cropped.
    “Marvin Minsky at OLPC, 2008” — Bcjordan / Sethwoodworth · Wikimedia Commons · CC BY 3.0 Cropped.
    “Claude Elwood Shannon (Tekniska museet 43069)” — Unknown photographer / Tekniska museet · Wikimedia Commons · CC BY 2.0 Cropped.
  14. 1956 · Memory and capacity
    Graph image · “An abstract image of a cube in blue and purple” — Milad Fakurian · Unsplash
    “Absolute judgement” — unscatter original illustration
    “Immediate memory” — unscatter original illustration
    “Meaningful chunking” — unscatter original illustration
  15. Summer 1956 · Researchers gather at Dartmouth
    Graph image · “a group of cubes that are connected to each other” — Shubham Dhage · Unsplash
    “Dartmouth Hall” — Unknown photographer (Commons contributor Urban) · Wikimedia Commons · Public domain
    “JOHNNIAC computer” — Adam Schuster from Orlando, FL, USA · Wikimedia Commons · CC BY 2.0
    “Herbert A. Simon and Allen Newell chess match” — Unknown original photographer; scan via Paolo Massa · Wikimedia Commons · CC BY-SA 2.0
  16. 1957 · Generative grammar
    Graph image · “Syntactic Structures” — Mouton de Gruyter · Open Library
    “Noam Chomsky, 2015” — Augusto Starita / Ministerio de Cultura de la Nación; retouching by Wugapodes · Wikimedia Commons · CC BY-SA 2.0 Cropped.
    “Hierarchical sentence structure” — unscatter original illustration
    “Related sentence forms” — unscatter original illustration
  17. 1958 · Selective attention
    Graph image · “Perception and Communication” — Pergamon Press · Open Library
    “Donald Broadbent” — Albert kok · Wikimedia Commons · CC BY-SA 3.0 Cropped.
    “Sennheiser HD600 headphones” — Ulfbastel / Benutzer123 · Wikimedia Commons · CC BY-SA 3.0 Cropped.
    “An early selection model” — unscatter original illustration
  18. March 1960 · Interactive computing
    Graph image · “allison-saeng-rs-tSzJvpBs-unsplash.jpg” — Allison Saeng · Unsplash
    “Portrait of J. C. R. Licklider” — Unknown photographer / U.S. National Library of Medicine · Wikimedia Commons · Public domain
    “DEC PDP-1 at the Computer History Museum” — Marcin Wichary · Wikimedia Commons · CC BY 2.0 Cropped.
    “SAGE console and light gun” — Tomwsulcer · Wikimedia Commons · CC0
  19. October 1960 · Attention can attenuate a message
    Graph image · “a pair of headphones floating in the air” — Hazel Z · Unsplash
    “Anne Treisman” — National Science Foundation · Wikimedia Commons · Public domain
    “The switched-ear experiment” — unscatter original illustration
    “Attenuation rather than blocking” — unscatter original illustration
  20. 1962 · Speech acts
    Graph image · “Bell System rotary telephone dial, 1940” — Kbrose · Wikimedia Commons · CC BY-SA 4.0 Cropped.
    “A request coordinates action” — unscatter original illustration
    “Making a commitment” — unscatter original illustration
    “Do not disturb toggle switch on black background” — Morgan Housel · Unsplash
  21. 1962–1968 · Human augmentation
    Graph image · “a computer keyboard on an orange background” — Dextar Studio · Unsplash
    “Douglas Engelbart at SRI, 1968” — SRI International · Wikimedia Commons · CC BY-SA 3.0 Cropped.
    “SRI computer mouse” — SRI International · Wikimedia Commons · CC BY-SA 3.0
    “NLS chord keyboard at the Computer History Museum” — Michael Hicks · Wikimedia Commons · CC BY 2.0
  22. 1964 · Understanding Media
    Graph image · “Understanding Media: The Extensions of Man” — The MIT Press · The MIT Press
    “a row of newspapers sitting on top of a shelf” — Jonathan Gong · Unsplash
    “Electric transmission towers with blue lines” — Getty Images · Unsplash+ · Unsplash+ License
    “The form of a medium” — unscatter original illustration
  23. October 1965 · Touchscreens
    Graph image · “Multitouch demonstration” — rorycellan (Flickr) · Wikimedia Commons · CC BY-SA 2.0 Cropped.
    “SAGE console and light gun” — Tomwsulcer · Wikimedia Commons · CC0
    “First-generation iPhone” — Carl Berkeley · Wikimedia Commons · CC BY-SA 2.0 Cropped.
    “Mapping controls to actions” — unscatter original illustration
  24. January 1966 · Early conversational interfaces
    Graph image · “Early conversational interfaces” — unscatter original illustration
    “DEC PDP-1 at the Computer History Museum” — Marcin Wichary · Wikimedia Commons · CC BY 2.0 Cropped.
    “Colorful letters scattered on a textured surface.” — NHN · Unsplash
    “Four purple speech bubbles on a translucent screen with floating geometric shapes” — kuu akura · Unsplash
  25. 1967 · Cognitive psychology
    Graph image · “Visual-cliff experiment (1960)” — Gibson and Walk (1960) · Wikimedia Commons Copyright 1960 Nature Publishing Group; free reuse rights not verified.
    “Corsi block-tapping test, Figure 1A” — Åsa Bartonek, Cecilia Guariglia and Laura Piccardi · Frontiers in Psychiatry · CC BY 4.0 Cropped.
    “Pattern recognition” — unscatter original illustration
    “Perceiving, remembering and using” — unscatter original illustration
  26. December 1968 · A display responds to the movement of the head
    Graph image · “getty-images-8sMrcHfgYVw-unsplash.jpg” — Getty Images · Unsplash
    “Ivan Sutherland at the Computer History Museum” — Dick Lyon · Wikimedia Commons · CC BY-SA 3.0 Cropped.
    “NASA virtual-reality equipment demonstration (AC89-0437-9)” — NASA · Wikimedia Commons · Public domain
    “VPL Eyephone and Dataglove” — Dave Pape · Wikimedia Commons · Public domain
  27. 1971 · Information and scarce attention
    Graph image · “An antique hourglass with sand flowing” — Callum Mullin · Unsplash
    “Herbert Simon, 1981” — Rochester Institute of Technology · Wikimedia Commons · Public domain
    “The New York Times newsroom, 1942” — Marjory Collins · Wikimedia Commons · Public domain
    “Selecting relevant information” — unscatter original illustration
  28. April 1971 · Intrinsic motivation
    Graph image · “a person playing a piano with their hands” — Kateryna Hliznitsova · Unsplash
    “person writing on notepad” — Calum MacAulay · Unsplash
    “A woman holding a smart watch in her hand” — Káplár Bálint Áron · Unsplash
    “Different reasons to engage” — unscatter original illustration
  29. 1973 · Attention and mental effort
    Graph image · “Attention and Effort” — Prentice-Hall; photograph by Downtown Brown Books · AbeBooks
    “Daniel Kahneman” — nrkbeta · Wikimedia Commons · CC BY-SA 2.0 Cropped.
    “Allocating available effort” — unscatter original illustration
    “Coordinating two tasks” — unscatter original illustration
  30. 1974 · refinement in 2000 · Working memory
    Graph image · “iridescent brain render on blue purple background” — Milad Fakurian · Unsplash
    “Corsi block-tapping test, Figure 1A” — Åsa Bartonek, Cecilia Guariglia and Laura Piccardi · Frontiers in Psychiatry · CC BY 4.0 Cropped.
    “Meaningful groups” — unscatter original illustration
    “Competing input” — unscatter original illustration
  31. 27 September 1974 · Judgement and biases
    Graph image · “a magnifying glass over a newspaper with the words now we have another way” — Road Ahead · Unsplash
    “Daniel Kahneman” — nrkbeta · Wikimedia Commons · CC BY-SA 2.0 Cropped.
    “An immediate visual impression” — unscatter original illustration
    “Familiar and effortful tasks” — unscatter original illustration
  32. October 1979 · Metacognition
    Graph image · “a mirror sitting on top of a table next to a plant” — MIHAO · Unsplash
    “Goals and strategies” — unscatter original illustration
    “Checking understanding” — unscatter original illustration
    “person writing on notepad” — Calum MacAulay · Unsplash
  33. 1980 · Attention binds visual features
    Graph image · “assorted-color color pencils” — salvatore ventura · Unsplash
    “Anne Treisman” — National Science Foundation · Wikimedia Commons · Public domain
    “Feature and conjunction search” — unscatter original illustration
    “Binding features into objects” — unscatter original illustration
  34. 1983 · HCI and interface design
    Graph image · “The Psychology of Human-Computer Interaction” — CRC Press / Taylor & Francis · Routledge
    “Xerox Alto workstation at the Computer History Museum” — Carlo Nardone · Wikimedia Commons · CC BY-SA 2.0 Cropped.
    “Xerox Star 8010” — vonguard · Wikimedia Commons · CC BY-SA 2.0 Cropped.
    “Apple Lisa, 1983” — Alan Light · Wikimedia Commons · CC BY 2.0 Cropped.
  35. November 1983 · Automation and human control
    Graph image · “A robotic arm in a clean, minimalist laboratory setting” — Brecht Corbeel · Unsplash
    “Three Mile Island control room” — John G. Kemeny et al · Wikimedia Commons · Public domain
    “Margaret Hamilton with Apollo software (restored)” — Draper Laboratory; restoration by Adam Cuerden · Wikimedia Commons · Public domain
    “Feedback and oversight” — unscatter original illustration
  36. 1984 · Susan Kare gives computer actions a visual language
    Graph image · “a close up of a computer keyboard with a blurry background” — MW · Unsplash
    “Susan Kare, 2019” — Cooper Hewitt, Smithsonian Design Museum · Wikimedia Commons · CC BY 3.0 Cropped.
    “Macintosh 128K (transparent background)” — All About Apple Museum; derivative by TDS / Grm wnr · Wikimedia Commons · CC BY-SA 3.0
    “Building a bitmap form” — unscatter original illustration
  37. 1985 · Self-determination
    Graph image · “A young green seedling growing from dark soil” — Pranav Gavali · Unsplash
    “A black double toggle light switch on a white wall” — Lunyon · Unsplash
    “a person playing a piano with their hands” — Kateryna Hliznitsova · Unsplash
    “Relatedness and support” — unscatter original illustration
  38. 9 October 1986 · Learning representations
    Graph image · “Geoffrey E. Hinton, 2024 Nobel Prize Laureate in Physics (cropped1)” — Arthur Petron · Wikimedia Commons · CC BY-SA 4.0 Cropped.
    “Learning useful features” — unscatter original illustration
    “Learning from error” — unscatter original illustration
    “Mark I Perceptron (operator’s manual, Figure 2)” — John C. Hay, Albert E. Murray · Wikimedia Commons · Public domain
  39. 1988 · Everyday design makes actions understandable
    Graph image · “The Psychology of Everyday Things” — Basic Books · Open Library
    “Donald Norman at AWF05” — jordanfischer · Wikimedia Commons · CC BY 2.0 Cropped.
    “Bell System rotary telephone dial, 1940” — Kbrose · Wikimedia Commons · CC BY-SA 4.0 Cropped.
    “Spatial mapping” — unscatter original illustration
  40. April 1988 · Cognitive load
    Graph image · “a group of cubes that are on a table” — Milad Fakurian / FAKURIAN DESIGN · Unsplash
    “Worked examples” — unscatter original illustration
    “Demands during learning” — unscatter original illustration
    “Reusable patterns” — unscatter original illustration
  41. 1989–1993 · World Wide Web
    Graph image · “The earth globe is depicted against a blue backdrop” — Igor Omilaev · Unsplash
    “Tim Berners-Lee” — Uldis Bojārs · Wikimedia Commons · CC BY-SA 2.0 Cropped.
    “Hypertext links” — unscatter original illustration
    “First Web Server” — User:Coolcaesar at en.wikipedia · Wikimedia Commons · CC BY-SA 3.0 Cropped.
  42. 1990 · Flow
    Graph image · “Flow: The Psychology of Optimal Experience” — Harper & Row · Open Library
    “Mihály Csíkszentmihályi” — Ehirsh · Wikimedia Commons · Public domain
    “Challenge and skill” — unscatter original illustration
    “a person playing a piano with their hands” — Kateryna Hliznitsova · Unsplash
  43. 30 April 1993 · Web software enters the public domain
    Graph image · “lines of HTML codes” — Florian Olivo · Unsplash
    “The 1993 release” — unscatter original illustration
    “First Web Server” — User:Coolcaesar at en.wikipedia · Wikimedia Commons · CC BY-SA 3.0 Cropped.
    “World Wide Web pins” — Rama & Musée Bolo · Wikimedia Commons · CC BY-SA 2.0 fr Cropped.
  44. April 1998 · When computers are designed to persuade
    Graph image · “A close up of a red liquid or substance” — Logan Voss · Unsplash
    “B. J. Fogg” — kris krüg · Wikimedia Commons · CC BY 2.0 Cropped.
    “A computer monitor displaying a design system interface with plants on a desk” — Balázs Kétyi · Unsplash
    “A woman holding a smart watch in her hand” — Káplár Bálint Áron · Unsplash
  45. July 1999 · Self-regulation and task initiation
    Graph image · “Hand checking off items on a to-do list” — Jakub Żerdzicki · Unsplash
    “person writing on notepad” — Calum MacAulay · Unsplash
    “Linking a cue to an intended action” — unscatter original illustration
    “Noticing and choosing the next step” — unscatter original illustration
  46. November 2000 · Working memory integrates information
    Graph image · “Hands connecting colorful puzzle pieces outdoors” — Alan Aprilio · Unsplash
    “Binding across kinds of information” — unscatter original illustration
    “Temporary integrated storage” — unscatter original illustration
    “Connecting with long-term memory” — unscatter original illustration
  47. 2001 · Attention becomes a business resource
    Graph image · “The Attention Economy: Understanding the New Currency of Business” — Harvard Business School Press · Open Library
    “a row of newspapers sitting on top of a shelf” — Jonathan Gong · Unsplash
    “Computer screen showing photo editing software interface” — JJROCHA · Unsplash
    “A scarce human resource” — unscatter original illustration
  48. February 2001 · Human agency
    Graph image · “A symmetric, iridescent floral abstract shape” — Milad Fakurian · Unsplash
    “Albert Bandura” — bandura@stanford.edu · Wikimedia Commons · CC BY-SA 4.0 Cropped.
    “Hand checking off items on a to-do list” — Jakub Żerdzicki · Unsplash
    “A black double toggle light switch on a white wall” — Lunyon · Unsplash
  49. April 2004 · Work unfolds across interrupted projects
    Graph image · “A wall with sticky notes attached to it” — Jakub Żerdzicki · Unsplash
    “Person is working on a laptop and writing in a notebook” — MJ Duford · Unsplash
    “Fragmented activity” — unscatter original illustration
    “Multiple working spheres” — unscatter original illustration
  50. 28 February 2006 · Mindset brings beliefs about learning to a wider audience
    Graph image · “Mindset: The New Psychology of Success” — Random House · Open Library
    “Carol Dweck for Innovation” — Satheesh Gopalan · Wikimedia Commons · CC BY-SA 3.0 Cropped.
    “person writing on notepad” — Calum MacAulay · Unsplash
    “A young green seedling growing from dark soil” — Pranav Gavali · Unsplash
  51. 5 September 2006 · Social media feeds
    Graph image · “person holding black phone” — ROBIN WORRALL · Unsplash
    “Updates gathered into a feed” — unscatter original illustration
    “Updates from a social network” — unscatter original illustration
    “Visibility and user control” — unscatter original illustration
  52. 2007–2013 · Smartphones and apps
    Graph image · “a close up of an iphone on a table” — Brian J. Tromp · Unsplash
    “First-generation iPhone” — Carl Berkeley · Wikimedia Commons · CC BY-SA 2.0 Cropped.
    “Steve Jobs presents iPhone” — Blake Patterson · Wikimedia Commons · CC BY 2.0 Cropped.
    “Multitouch demonstration” — rorycellan (Flickr) · Wikimedia Commons · CC BY-SA 2.0 Cropped.
  53. April 2008 · Interruptions and task switching
    Graph image · “person holding red and beige twin bell analog alarm clock” — Malvestida · Unsplash
    “Nokia Lumia 920” — Indrek · Wikimedia Commons · CC BY-SA 4.0 Cropped.
    “Breaking and resuming a task” — unscatter original illustration
    “Costs beyond task time” — unscatter original illustration
  54. 10 July 2008 · Apps become a distribution platform
    Graph image · “A modern smartphone with a blank white screen” — Hudai Gayiran · Unsplash
    “iPhone and iPhone 3G” — Dan Taylor from London, UK · Wikimedia Commons · CC BY 2.0
    “A mobile software marketplace” — unscatter original illustration
    “Many uses on one device” — unscatter original illustration
  55. July 2009 · Attention residue
    Graph image · “a stack of colorful paper” — Kelsy Gagnebin · Unsplash
    “Person is working on a laptop and writing in a notebook” — MJ Duford · Unsplash
    “Attention carried from a previous task” — unscatter original illustration
    “Disengagement before a transition” — unscatter original illustration
  56. 2011 · Fast impressions and effortful thought
    Graph image · “Thinking, Fast and Slow” — Farrar, Straus and Giroux / Macmillan · Macmillan
    “Daniel Kahneman” — nrkbeta · Wikimedia Commons · CC BY-SA 2.0 Cropped.
    “An immediate visual impression” — unscatter original illustration
    “Familiar and effortful tasks” — unscatter original illustration
  57. 2012 · Deep learning
    Graph image · “A high-performance graphics card with three fans” — Back2Gaming · Unsplash
    “EVGA GeForce GTX580” — TheStriker · Wikimedia Commons · CC BY-SA 4.0
    “Learned convolutional features” — unscatter original illustration
    “Classifying visual content” — unscatter original illustration
  58. 5 June 2013 · A majority of US adults own a smartphone
    Graph image · “Two phones with chat bubbles and a red heart” — Omar:. Lopez-Rincon · Unsplash
    “Samsung Galaxy S4” — Kārlis Dambrāns · Wikimedia Commons · CC BY 2.0 Cropped.
    “A pencil rests on a map” — unavailable parts · Unsplash
    “Cell tower” — Yq2Yq · Wikimedia Commons · CC0
  59. 1 September 2014 · Translation learns where to look
    Graph image · “A wireframe cube containing swirling clusters of orange and black particles” — Aurora Mititelu / Google DeepMind · Unsplash
    “Colorful letters scattered on a textured surface.” — NHN · Unsplash
    “Soft alignment between languages” — unscatter original illustration
    “Context for each target word” — unscatter original illustration
  60. 4 November 2014 · Product design aims to build habits
    Graph image · “Hooked: How to Build Habit-Forming Products” — Portfolio/Penguin · Open Library
    “person holding black phone” — ROBIN WORRALL · Unsplash
    “The proposed Hook cycle” — unscatter original illustration
    “Varying outcomes” — unscatter original illustration
  61. 2016 · Recommendation systems
    Graph image · “A 3D red play button icon on a dark blue background” — Alexander Shatov · Unsplash
    “Tensor Processing Unit 3.0” — Zinskauf · Wikimedia Commons · CC BY-SA 4.0 Cropped.
    “a group of computers” — Rose Pilkington / Google DeepMind · Unsplash
    “Orange and blue 3D bar chart columns on a white grid surface” — Nick Brunner · Unsplash
  62. 5 January 2016 · Deep Work brings sustained concentration into everyday discussion
    Graph image · “Deep Work: Rules for Focused Success in a Distracted World” — Grand Central Publishing · Open Library
    “Protecting a work block” — unscatter original illustration
    “Desk with lamp, sketchbook, pencils, and eraser” — Calvin Wise · Unsplash
    “Do not disturb toggle switch on black background” — Morgan Housel · Unsplash
  63. 27 January 2016 · Meaning in life
    Graph image · “A pencil rests on a map” — unavailable parts · Unsplash
    “a magnifying glass over a newspaper with the words now we have another way” — Road Ahead · Unsplash
    “Hand checking off items on a to-do list” — Jakub Żerdzicki · Unsplash
    “a person playing a piano with their hands” — Kateryna Hliznitsova · Unsplash
  64. 27 January 2016 · Learning and search combine in Go
    Graph image · “Robot watching people play go board game” — Amos K · Unsplash
    “Go stones on a go board with containers.” — Kyle Hinkson · Unsplash
    “Learning from games” — unscatter original illustration
    “Search guided by neural networks” — unscatter original illustration
  65. September 2016 · Cognitive offloading
    Graph image · “A wall with sticky notes attached to it” — Jakub Żerdzicki · Unsplash
    “person writing on notepad” — Calum MacAulay · Unsplash
    “Person is working on a laptop and writing in a notebook” — MJ Duford · Unsplash
    “A pencil rests on a map” — unavailable parts · Unsplash
  66. 12 June 2017 · Transformers
    Graph image · “abstract blue and pink 3D cubes” — Rose Pilkington / Google DeepMind · Unsplash
    “Aidan Gomez at Collision, 2023” — Piaras Ó Mídheach · Wikimedia Commons · CC BY 2.0 Cropped.
    “Weighting relationships” — unscatter original illustration
    “NVIDIA Tesla P100 processor die” — Fritzchens Fritz · Wikimedia Commons · CC0
  67. May–June 2018 · Digital wellbeing tools
    Graph image · “a cell phone with a picture of a plant on the screen” — Dawit · Unsplash
    “Wooden bench bathed in sunlight amidst lush green park” — Ashique Anan Abir · Unsplash
    “Making use visible” — unscatter original illustration
    “Managing interruptions and time” — unscatter original illustration
  68. 11 October 2018 · Language representations use both contexts
    Graph image · “Dark 3D letters spelling infinite on a white surface” — Logan Voss · Unsplash
    “Masked-language pretraining” — unscatter original illustration
    “Bidirectional context” — unscatter original illustration
    “Pretraining and fine-tuning” — unscatter original illustration
  69. May 2019 · Human–AI interaction
    Graph image · “a hand reaching out towards a robot hand” — Cash Macanaya · Unsplash+ · Unsplash+ License
    “A computer monitor displaying a design system interface with plants on a desk” — Balázs Kétyi · Unsplash
    “A pencil eraser rubs away pencil marks” — Adam Custer · Unsplash
  70. 28 May 2020 · Examples become part of the prompt
    Graph image · “Person typing on a laptop keyboard” — Back2Gaming · Unsplash
    “Ilya Sutskever at Democratizing Deep Learning” — Steve Jurvetson · Wikimedia Commons · CC BY 2.0 Cropped.
    “Examples in the prompt” — unscatter original illustration
    “Generating a continuation” — unscatter original illustration
  71. July 2020 · review in 2024 · Language understanding in AI
    Graph image · “Person typing on a laptop keyboard” — Back2Gaming · Unsplash
    “Predicting a word continuation” — unscatter original illustration
    “Language with visual context” — unscatter original illustration
    “Linguistic form and situated use” — unscatter original illustration
  72. June 2022 · Image generation works in a compressed space
    Graph image · “pink and blue ink in water on a white background” — engin akyurt · Unsplash
    “Close-up of a professional camera lens” — Jaimy de Hon · Unsplash
    “A learned compressed space” — unscatter original illustration
    “Conditioning on language” — unscatter original illustration
  73. 2022–2024 · Conversational and multimodal AI
    Graph image · “the open ai logo is displayed on a computer screen” — Andrew Neel · Unsplash
    “Sam Altman, November 2022” — Village Global · Wikimedia Commons · CC BY 2.0 Cropped.
    “Four purple speech bubbles on a translucent screen with floating geometric shapes” — kuu akura · Unsplash
    “Rows of black server racks with white logos in a data center” — imgix · Unsplash
  74. 10 January 2023 · Attention is studied in everyday digital work
    Graph image · “Attention Span: A Groundbreaking Way to Restore Balance, Happiness and Productivity” — HarperCollins / Hanover Square Press · HarperCollins
    “Library of Congress Main Reading Room” — Carol M. Highsmith · Wikimedia Commons · Public domain
    “Awareness and intentional attention” — unscatter original illustration
    “Attention and recovery” — unscatter original illustration
  75. 14 March 2023 · A multimodal model broadens the interface
    Graph image · “NVIDIA GA102 processor die” — Fritzchens Fritz · Wikimedia Commons · CC0
    “photo of computer cables” — Kvistholt Photography · Unsplash
    “Text and image context” — unscatter original illustration
    “Evaluating a response” — unscatter original illustration
  76. 13 May 2024 · Voice, vision and text share a model
    Graph image · “black condenser microphone close-up photography” — Obi · Unsplash
    “HTC One on an Apple keyboard” — Kārlis Dambrāns · Wikimedia Commons · CC BY 2.0 Cropped.
    “Text, vision and audio” — unscatter original illustration
    “Continuing the conversation” — unscatter original illustration
  77. 1 August 2024 · AI oversight becomes a regulatory question
    Graph image · “blue and yellow star flag” — Christian Lue · Unsplash
    “Entry into force” — unscatter original illustration
    “A risk-based framework” — unscatter original illustration
    “Human oversight” — unscatter original illustration
  78. 12 September 2024 · Models use more computation before answering
    Graph image · “Silver balls rolling down a colorful maze structure” — Luke Jones · Unsplash
    “Chessboard with pieces and a digital clock” — Michal Vrba · Unsplash
    “Additional inference work” — unscatter original illustration
    “The computation trade-off” — unscatter original illustration
  79. 25 November 2024 · Assistants connect through a shared protocol
    Graph image · “blue UTP cord” — Jordan Harrison · Unsplash
    “Clients and servers” — unscatter original illustration
    “Reading and acting” — unscatter original illustration
    “Scoped access” — unscatter original illustration
  80. 20 January 2025 · Reasoning models become available as weights
    Graph image · “a blue and pink abstract background with a variety of shapes” — Milad Fakurian · Unsplash
    “multicolored abacus” — Crissy Jarvis · Unsplash
    “R1 and the R1-Zero experiment” — unscatter original illustration
    “Training smaller models” — unscatter original illustration
  81. April 2025 · AI confidence and critical-thinking effort
    Graph image · “a magnifying glass over a newspaper with the words now we have another way” — Road Ahead · Unsplash
    “A self-report survey” — unscatter original illustration
    “Confidence and effort” — unscatter original illustration
    “Verification and integration” — unscatter original illustration
  82. 17 July 2025 · AI agents
    Graph image · “Laptop, keyboard, and mouse on a desk with sunlight” — ReyLabs Studio · Unsplash
    “From goals to tools” — unscatter original illustration
    “A toolbox for work” — unscatter original illustration
    “User control” — unscatter original illustration
  83. 28 January 2026 · Assistance and learning can diverge
    Graph image · “Person typing on a laptop with code displayed” — Daniil Komov · Unsplash
    “Random assignment” — unscatter original illustration
    “What was assessed” — unscatter original illustration
    “Observed interaction patterns” — unscatter original illustration
  84. 3 February 2026 · Evidence about AI risks is brought together
    Graph image · “a long row of flags in front of a building” — Meizhi Lang · Unsplash
    “International AI Safety Report 2026” — Department for Science, Innovation and Technology, © Crown 2026 · International AI Safety Report · Open Government Licence v3.0
    “Three assessment questions” — unscatter original illustration
    “Layers of risk management” — unscatter original illustration
Full bibliography 128 sources
  1. Anne M. Treisman (1960). Contextual Cues in Selective Listening
  2. Ivan E. Sutherland (1968). A Head-mounted Three Dimensional Display
  3. Susan Kare; interview by Alex Pang (2000). Interview with Susan Kare
  4. Carol S. Dweck / Random House (2006). Mindset: The New Psychology of Success
  5. Cal Newport / Grand Central Publishing (2016). Deep Work: Rules for Focused Success in a Distracted World
  6. David E. Rumelhart, Geoffrey E. Hinton and Ronald J. Williams (1986). Learning representations by back-propagating errors
  7. William James (1890). The Principles of Psychology, Chapter XI: Attention
  8. John McCarthy, Marvin Minsky, Nathaniel Rochester and Claude Shannon (1955). A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence
  9. George A. Miller (1956). The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information
  10. Dartmouth (Undated). The Founding and Future of AI (institutional history)
  11. Donald E. Broadbent (1958). Perception and Communication (Max Planck publication record)
  12. Donald E. Broadbent (1984). Citation Classic commentary on Perception and Communication
  13. J. C. R. Licklider (1960). Man-Computer Symbiosis
  14. Douglas C. Engelbart (1962). Augmenting Human Intellect: A Conceptual Framework
  15. Marshall McLuhan (1964; MIT Press reissue 1994). Understanding Media: The Extensions of Man
  16. Marshall McLuhan (1964). Understanding Media — Chapter 1, The Medium Is the Message (excerpt)
  17. Ulric Neisser (1967; reissued 2014). Cognitive Psychology (publisher description of the classic edition)
  18. Herbert A. Simon (1971). Designing Organizations for an Information-Rich World, in Computers, Communications, and the Public Interest
  19. Alan D. Baddeley and Graham Hitch (1974). Working Memory
  20. Alan Baddeley (2000). The episodic buffer: a new component of working memory?
  21. John H. Flavell (1979). Metacognition and Cognitive Monitoring: A New Area of Cognitive–Developmental Inquiry
  22. Stuart K. Card, Thomas P. Moran and Allen Newell (1983). The Psychology of Human-Computer Interaction
  23. Allen Newell and Stuart K. Card (1985). The Prospects for Psychological Science in Human-Computer Interaction
  24. Lisanne Bainbridge (1983). Ironies of Automation, Automatica 19(6), 775–779
  25. John Sweller (1988). Cognitive Load During Problem Solving: Effects on Learning
  26. CERN (Undated; covers 1989–1993). The birth of the Web (institutional history)
  27. Tim Berners-Lee (1989). Information Management: A Proposal
  28. B. J. Fogg (1998). Persuasive Computers: Perspectives and Research Directions
  29. Albert Bandura (2001). Social Cognitive Theory: An Agentic Perspective
  30. Apple (2007). Apple Reinvents the Phone with iPhone (product announcement)
  31. Apple (2007). iPhone Premieres This Friday Night at Apple Retail Stores
  32. Aaron Smith / Pew Research Center (2013). Smartphone Ownership 2013
  33. Aaron Smith / Pew Research Center (2013). Smartphone Ownership 2013: Methods
  34. Paul Covington, Jay Adams and Emre Sargin (2016). Deep Neural Networks for YouTube Recommendations
  35. Ashish Vaswani and colleagues (2017). Attention Is All You Need
  36. Saleema Amershi and colleagues (2019). Guidelines for Human-AI Interaction
  37. OpenAI (2022). Introducing ChatGPT (public research-preview announcement)
  38. International Ergonomics Association (Undated; definition adopted 2000). What Is Ergonomics (HFE)?
  39. Cognitive Science Society (Undated). About the Cognitive Science Society
  40. Abraham H. Maslow (1943). A Theory of Human Motivation
  41. Viktor E. Frankl (1946). Man’s Search for Meaning
  42. Norbert Wiener (1948). Cybernetics: Or Control and Communication in the Animal and the Machine
  43. Claude E. Shannon (1948). A Mathematical Theory of Communication
  44. Noam Chomsky (1957). Syntactic Structures
  45. Daniel Kahneman (1973). Attention and Effort
  46. Anne M. Treisman and Garry Gelade (1980). A Feature-Integration Theory of Attention
  47. Donald A. Norman (1988). The Psychology of Everyday Things
  48. Mihaly Csikszentmihalyi (1990). Flow: The Psychology of Optimal Experience
  49. Thomas H. Davenport and John C. Beck (2001). The Attention Economy: Understanding the New Currency of Business
  50. Daniel Kahneman (2011). Thinking, Fast and Slow
  51. Victor M. González and Gloria Mark (2004). “Constant, Constant, Multi-tasking Craziness”: Managing Multiple Working Spheres
  52. Facebook / Meta (2006). Company Info: Launch of News Feed
  53. Apple (2018). The App Store turns 10
  54. Gloria Mark, Daniela Gudith and Ulrich Klocke (2008). The Cost of Interrupted Work: More Speed and Stress
  55. Sophie Leroy (2009). Why is it so hard to do my work? The challenge of attention residue when switching between work tasks
  56. Alex Krizhevsky, Ilya Sutskever and Geoffrey E. Hinton (2012). ImageNet Classification with Deep Convolutional Neural Networks
  57. Dzmitry Bahdanau, Kyunghyun Cho and Yoshua Bengio (2014). Neural Machine Translation by Jointly Learning to Align and Translate
  58. Nir Eyal (2014). Hooked: How to Build Habit-Forming Products
  59. David Silver and colleagues (2016). Mastering the game of Go with deep neural networks and tree search
  60. Jacob Devlin, Ming-Wei Chang, Kenton Lee and Kristina Toutanova (2018). BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
  61. Apple (2018). iOS 12 introduces new features to reduce interruptions and manage Screen Time
  62. Sameer Samat / Google (2018). Android P: Packed with smarts and simpler than ever
  63. Tom B. Brown and colleagues (2020). Language Models are Few-Shot Learners
  64. Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser and Björn Ommer (2022). High-Resolution Image Synthesis With Latent Diffusion Models
  65. Gloria Mark (2023). Attention Span: A Groundbreaking Way to Restore Balance, Happiness and Productivity
  66. OpenAI (2023). GPT-4
  67. OpenAI (2023). GPT-4 Technical Report
  68. OpenAI (2024). Hello GPT-4o
  69. European Commission (2024). AI Act enters into force
  70. OpenAI (2024). Introducing OpenAI o1-preview
  71. Anthropic (2024). Introducing the Model Context Protocol
  72. DeepSeek (2025). DeepSeek-R1 Release
  73. DeepSeek-AI (2025). DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
  74. OpenAI (2025). Introducing ChatGPT agent: bridging research and action
  75. Hao-Ping (Hank) Lee and colleagues (2025). The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers
  76. Judy Hanwen Shen and Alex Tamkin (2026). How AI Impacts Skill Formation
  77. Yoshua Bengio and the international writing group (2026). International AI Safety Report 2026
  78. Ludwig Wittgenstein (1922). Tractatus Logico-Philosophicus
  79. Wilhelm Wundt (1910). Das Institut für experimentelle Psychologie zu Leipzig
  80. University of Leipzig (2021). Vom kleinen Auditorium zum ersten psychologischen Institut der Welt
  81. J. Ridley Stroop (1935). Studies of Interference in Serial Verbal Reactions
  82. Colin M. MacLeod (1991). Half a Century of Research on the Stroop Effect: An Integrative Review
  83. Amos Tversky and Daniel Kahneman (1974). Judgment under Uncertainty: Heuristics and Biases
  84. Peter M. Gollwitzer (1999). Implementation Intentions: Strong Effects of Simple Plans
  85. Evan F. Risko and Sam J. Gilbert (2016). Cognitive Offloading
  86. Ludwig Wittgenstein (1953; 4th ed. 2009). Philosophical Investigations (first published 1953; linked fourth edition)
  87. Anat Biletzki and Anat Matar (2026 revision). Ludwig Wittgenstein
  88. J. L. Austin (1962; 2nd ed. 1975). How To Do Things With Words: The William James Lectures delivered at Harvard University in 1955 (first published 1962; linked second edition)
  89. Edward L. Deci (1971). Effects of Externally Mediated Rewards on Intrinsic Motivation
  90. Edward L. Deci and Richard M. Ryan (1985). Intrinsic Motivation and Self-Determination in Human Behavior
  91. Richard M. Ryan and Edward L. Deci (2000). Self-Determination Theory and the Facilitation of Intrinsic Motivation, Social Development, and Well-Being
  92. Frank Martela and Michael F. Steger (2016). The three meanings of meaning in life: Distinguishing coherence, purpose, and significance
  93. Emily M. Bender and Alexander Koller (2020). Climbing towards NLU: On Meaning, Form, and Understanding in the Age of Data
  94. Kyle Mahowald, Anna A. Ivanova, Idan A. Blank, Nancy Kanwisher, Joshua B. Tenenbaum and Evelina Fedorenko (2024). Dissociating language and thought in large language models
  95. George A. Miller (2003). The cognitive revolution: a historical perspective
  96. Paul Thagard (2023 revision). Cognitive Science
  97. Barbara C. Scholz, Francis Jeffry Pelletier, Geoffrey K. Pullum and Ryan Nefdt (2024 revision). Philosophy of Linguistics
  98. Alan M. Turing (1936–1937). On Computable Numbers, with an Application to the Entscheidungsproblem
  99. Warren S. McCulloch and Walter Pitts (1943). A logical calculus of the ideas immanent in nervous activity
  100. E. A. Johnson (1965). Touch display—a novel input/output device for computers
  101. Bill Buxton (2023 update). Some Milestones in Computer Input Devices: An Informal Timeline
  102. Joseph Weizenbaum (1966). ELIZA—A Computer Program for the Study of Natural Language Communication Between Man and Machine
  103. Doug Engelbart Institute / Stanford University archival collection (1968; archival presentation). Doug Engelbart 1968 Demo
  104. CERN (1993). Software release of WWW into public domain
  105. CERN (Institutional history; covers 1989–1993). A short history of the Web
  106. OpenAI (2015). Introducing OpenAI
  107. Noah Shinn, Federico Cassano, Edward Berman, Ashwin Gopinath, Karthik Narasimhan and Shunyu Yao (2023). Reflexion: Language Agents with Verbal Reinforcement Learning
  108. Stanford University AI100 Study Panel (2016). Appendix I: A Short History of AI
  109. Computer History Museum (Institutional history). Timeline of Computer History: Computers
  110. Edward N. Zalta (Fall 2026 edition). Gottlob Frege, Stanford Encyclopedia of Philosophy
  111. United States Census Bureau (Institutional history). The Hollerith Machine
  112. International Telecommunication Union (Institutional history). Overview of ITU’s History: From telegraph to telephone
  113. Larry Hardesty / MIT News (2010). Explained: The Shannon limit
  114. Stanford Institute for Human-Centered Artificial Intelligence (2025). The 2025 AI Index Report
  115. Paul Thagard (2023). Cognitive Science — Stanford Encyclopedia of Philosophy (history and approaches; revised 2023)
  116. George A. Miller (2003). The cognitive revolution: a historical perspective
  117. Computer History Museum (Undated; accessed 28 October 2026). Timeline of Computer History: Computers
  118. Stanford University AI100 Study Panel (2016). Appendix I: A Short History of AI — Artificial Intelligence and Life in 2030
  119. CERN (Undated; accessed 28 October 2026). A short history of the Web
  120. Barbara C. Scholz, Francis Jeffry Pelletier, Geoffrey K. Pullum and Ryan Nefdt (2024). Philosophy of Linguistics — Stanford Encyclopedia of Philosophy (revised 2024)
  121. Richard M. Ryan and Edward L. Deci (2000). Self-Determination Theory and the Facilitation of Intrinsic Motivation, Social Development, and Well-Being
  122. John G. Benjafield (2019). Major Paradigms and Approaches in Psychology — The Cambridge Handbook of the Intellectual History of Psychology
  123. Kyle Mahowald, Anna A. Ivanova, Idan A. Blank, Nancy Kanwisher, Joshua B. Tenenbaum and Evelina Fedorenko (2024). Dissociating language and thought in large language models (review)
  124. Computer History Museum (Undated; accessed 28 October 2026). Revolution: The First 2000 Years of Computing (historical galleries)
  125. Todd Bridgman, Stephen Cummings and John Ballard (2019). Who Built Maslow’s Pyramid? A History of the Creation of Management Studies’ Most Famous Symbol and Its Implications for Management Education
  126. Hermann Rorschach (1921). Psychodiagnostik: Methodik und Ergebnisse eines wahrnehmungsdiagnostischen Experiments (Deutenlassen von Zufallsformen)
  127. University of Bern, Institute for the History of Medicine (Institutional history; accessed 28 October 2026). Rorschach Archives and Collection

Clear Mind