# How AI works

Canonical page: https://dominiklukes.net/topics/how-ai-works

What large language models are, what they can and cannot do, and how the field has changed since ChatGPT.

86 items, newest first.

- [Four Years After ChatGPT](https://dominiklukes.net/speaking/four-years-after-chatgpt) (Talks, 2026)
- [What's new in AI this month?](https://dominiklukes.net/speaking/mondai-roundup-live) (Talks, 2026)
- [The future of data in the age of AI Agents: New opportunities for old methods in data management and collection](https://dominiklukes.net/speaking/oxford-data-meetup-november-2026) (Talks, 2026)
- [AI Thought Exchange, October meeting](https://dominiklukes.net/speaking/ai-thought-exchange-october-2026) (Talks, 2026)
- [Common misconceptions about Large Language Models and a few literal metaphors](https://dominiklukes.net/writing/what-models-want) (Writing, 2026)
- [York Festival of AI: nine sessions](https://dominiklukes.net/speaking/york-festival-of-ai-nine-sessions-2026) (Talks, 2026)
- [Teaching During the Cognitive Revolution](https://dominiklukes.net/speaking/teaching-during-the-cognitive-revolution-2026) (Talks, 2026)
- [AI and Expertise](https://dominiklukes.net/speaking/ai-and-expertise-2026) (Talks, 2026)
- [Language without structure? How can LLMs speak all the languages without knowing any of the rules?](https://dominiklukes.net/speaking/language-without-structure-how-can-llms-speak-all-the-langua-2026) (Talks, 2026)
- [Panel at the OECD](https://dominiklukes.net/speaking/panel-at-the-oecd-2026) (Talks, 2026)
- [Seminar on AI and programming at the Mathematical Institute](https://dominiklukes.net/speaking/seminar-on-ai-and-programming-at-the-mathematical-institute-2026) (Talks, 2026)
- [We've been using AI wrong: Agent Skills and the Semantic Powers of LLMs](https://dominiklukes.net/writing/weve-been-using-ai-wrong-skills-and-the-trouble-with-context) (Writing, 2026)
- [Teaching during the Cognitive Revolution: What will learning and teaching look like in the age of AI?](https://dominiklukes.net/speaking/teaching-cognitive-revolution) (Talks, 2026)
- [How Humans Hallucinate and How We Deal with It](https://dominiklukes.net/writing/how-humans-hallucinate-knowledge-memory-and-certainty) (Writing, 2025)
- [Three Years After ChatGPT: From chatbots to agents and other trends on the frontier of generative AI](https://dominiklukes.net/speaking/three-years-after-chatgpt) (Talks, 2025)
- [Three Years After ChatGPT: From Chatbots to Agents and the Shifting Frontier of Generative AI](https://dominiklukes.net/writing/three-years-after-chatgpt-from-chatbots-agents-shifting-dominik-lukes-3e2se) (Writing, 2025)
- [How LLMs Understand Text: Context and Meaning in the AI Age](https://dominiklukes.net/speaking/llms-understand-text) (Talks, 2025)
- [Symbols and Manipulation](https://dominiklukes.net/writing/symbols-and-manipulation) (Writing, 2025)
- [Karpathy Thinks in Code](https://dominiklukes.net/writing/karpathy-thinks-in-code) (Writing, 2025)
- [50 First Dates and the ChatGPT Memory Problem](https://dominiklukes.net/writing/50-first-dates-and-the-chatgpt-memory-problem) (Writing, 2025)
- [GPT-5 is here! Long time coming but blazingly fast!](https://dominiklukes.net/writing/gpt-5-here-long-time-coming-blazingly-fast-dominik-lukes-ofwge) (Writing, 2025)
- [What is ChatGPT Agent and what does it tell us about the future of AI?](https://dominiklukes.net/writing/what-chatgpt-agent-does-tell-us-future-ai-dominik-lukes-adewe) (Writing, 2025)
- [OPP Summer Lecture](https://dominiklukes.net/speaking/opp-summer-lecture-2025) (Talks, 2025)
- [Embarrassingly out of date and unreliable! Some thoughts on the new Stanford AI Index Report](https://dominiklukes.net/writing/embarrassingly-out-date-unreliable-quick-thoughts-new-dominik-lukes-qwlve) (Writing, 2025)
- [Creation is perception: On the role of skill in the age of AI](https://dominiklukes.net/writing/creation-perception-role-skill-age-ai-dominik-lukes-bryge) (Writing, 2025)
- [AI Fluency as a Job Requirement: Karpathy on Individual Advantage and Shopify's New Mandate](https://dominiklukes.net/writing/ai-fluency-job-requirement-karpathy-individual-advantage-lukes-4xrge) (Writing, 2025)
- [All art is conceptual art: Reflections on the aftershocks of an AI image-generation earthquake](https://dominiklukes.net/writing/all-art-conceptual-reflections-aftershocks-ai-earthquake-lukes-bpcfe) (Writing, 2025)
- ["You're doing AI wrong": Choosing an AI tool is more like hiring an employee than buying software](https://dominiklukes.net/writing/youre-doing-ai-wrong-choosing-tool-more-like-hiring-employee-lukes-c8sme) (Writing, 2025)
- [GPT-4o's native image generation is a true revolution in what is possible with AI and images](https://dominiklukes.net/writing/gpt-4os-native-image-generation-true-revolution-what-possible-lukes-2todc) (Writing, 2025)
- [Talk on AI at OxLearn](https://dominiklukes.net/speaking/talk-on-ai-at-oxlearn-2025) (Talks, 2025)
- [The future of tech is human: Redefining the role of educators in the wake of Industry 5.0](https://dominiklukes.net/speaking/the-future-of-tech-is-human-redefining-the-role-of-educators-2025) (Talks, 2025)
- [Two years after ChatGPT](https://dominiklukes.net/speaking/two-years-after-chatgpt-2025) (Talks, 2025)
- [What will 2025 be like for AI? 5 Predictions for the Future](https://dominiklukes.net/writing/what-2025-like-ai-5-predictions-future-dominik-lukes-0n0ee) (Writing, 2025)
- [A Year in AI: Doing more with less in 2024](https://dominiklukes.net/writing/year-ai-doing-more-less-2024-dominik-lukes-iqvue) (Writing, 2024)
- [Two Years After ChatGPT: Reflections on The State and Future Directions of Generative AI](https://dominiklukes.net/speaking/two-years-after-chatgpt) (Talks, 2024)
- [What have Large Language Models actually learned about language](https://dominiklukes.net/speaking/what-have-large-language-models-actually-learned-about-langu-2024) (Talks, 2024)
- [Magical Meaning Machines](https://dominiklukes.net/speaking/magical-meaning-machines-2024) (Talks, 2024)
- [Trend 2: Interface Innovations: (Re)defining the user experience of AI](https://dominiklukes.net/writing/trend-2-interface-innovations-redefining-user-experience-lukes-wuxxe) (Writing, 2024)
- [AI Trend 1: Multimodal Revolution: Generative AI at the birth of a new paradigm of computer interaction](https://dominiklukes.net/writing/ai-trend-1-multimodal-revolution-generative-birth-new-dominik-lukes-dmq2e) (Writing, 2024)
- [State of AI: 5 trends in generative AI that will shape academic practice in 2025](https://dominiklukes.net/writing/state-ai-5-trends-generative-shape-academic-practice-2025-lukes-5d1de) (Writing, 2024)
- [Semantic Machines](https://dominiklukes.net/speaking/semantic-machines-2024) (Talks, 2024)
- [Talk and panel on AI at Tsinghua University](https://dominiklukes.net/speaking/talk-and-panel-on-ai-at-tsinghua-university-2024) (Talks, 2024)
- [The future of AI in Academic Practice is one million tokens and why it matters](https://dominiklukes.net/writing/future-ai-academic-practice-one-million-tokens-why-matters-lukes-caaef) (Writing, 2024)
- [One year anniversary of ChatGPT this week - where are we,  what have we learned?](https://dominiklukes.net/writing/one-year-anniversary-chatgpt-week-where-we-what-have-dominik-lukes-zhhne) (Writing, 2023)
- [How (not) to learn about AI with metaphors: And how to use ChatGPT as a metaphor generation assistant](https://dominiklukes.net/writing/how-learn-ai-metaphors-use-chatgpt-metaphor-generation-dominik-lukes) (Writing, 2023)
- [Have you asked your bot to make you a table? ChatGPT as a cognitive scaffolding engine](https://dominiklukes.net/writing/have-you-asked-your-bot-make-table-chatgpt-cognitive-engine-lukes) (Writing, 2023)
- [The State of AI for August 2023: 6 trends from the last 3 months and 6 actions you can take today](https://dominiklukes.net/writing/state-ai-august-2023-6-trends-from-last-3-months-actions-lukes) (Writing, 2023)
- [Invited talk at the European Educational Publishers Group conference](https://dominiklukes.net/speaking/invited-talk-at-the-european-educational-publishers-group-co-2023) (Talks, 2023)
- [Learn to Be a Prompt Engineer: PREP and other guides to uncovering the secrets to ChatGPT mastery](https://dominiklukes.net/writing/learn-prompt-engineer-prep-other-guides-uncovering-secrets-lukes) (Writing, 2023)
- [More AI news: Incremental but meaningful progress - From models to products](https://dominiklukes.net/writing/more-ai-news-incremental-meaningful-progress-from-models-lukes) (Writing, 2023)
- [ChatGPT is a Weird Linguist: What Should Language Teachers Know?](https://dominiklukes.net/speaking/chatgpt-weird-linguist) (Talks, 2023)
- [Unintentional Pygmalions: 4 questions to ask when checking an artificial entity for sentience and how to think about the answers](https://dominiklukes.net/writing/2022-08-09-unintentional-pygmalions-4-questions-to-ask-when-checking-an-artifici) (Writing, 2022)
- [Turing tests in Chinese rooms: What does it mean for AI to outperform humans](https://dominiklukes.net/writing/2019-07-23-turing-tests-in-chinese-rooms-what-does-it-mean-for-ai-to-outperform-) (Writing, 2019)
- [3 “easy” things that are hard for both humans and AI](https://dominiklukes.net/writing/2018-10-19-3-easy-things-that-are-hard-for-both-humans-and-ai) (Writing, 2018)
- [Does machine learning produce mental representations?](https://dominiklukes.net/writing/2018-05-28-does-machine-learning-produce-mental-representations) (Writing, 2018)
- [Learning vs training in machines and organizations: Production of knowledge vs production of capability](https://dominiklukes.net/writing/learning-vs-training-in-machines-and-organizations-production-of-knowledge-vs-pr) (Writing, 2018)
- [Cognition, information, knowledge and the limits of serial computing](https://dominiklukes.net/writing/2007-09-11-cogntion-information-knowledge-and-the-limits-of-serial-computing) (Writing, 2007)
- [Algorithms and everything](https://dominiklukes.net/writing/2007-01-09-algorithms-and-everything) (Writing, 2007)
- [About Schemas and Propositions](https://dominiklukes.net/writing/about-schemas-and-propositions-substack) (Writing, undated)
- [AI for Research(ers): Principles, Context, and Tools](https://dominiklukes.net/teaching/ai-for-research-ers-principles-context-and-tools) (Teaching, undated)
- [AI Trends & Frontier Capabilities](https://dominiklukes.net/resources/ai-trends) (Guides, undated)
- [ai-model-benchmarks](https://dominiklukes.net/agents/skills/ai-model-benchmarks) (Agent skills, undated)
- [Cognitive Baseline: Calibrating your Mental Model](https://dominiklukes.net/teaching/cognitive-baseline-calibrating-your-mental-model) (Teaching, undated)
- [Comparing Model Knowledge](https://dominiklukes.net/teaching/comparing-model-knowledge) (Teaching, undated)
- [Context & Retrieval: Handling Information](https://dominiklukes.net/teaching/context-retrieval-handling-information) (Teaching, undated)
- [Data Analysis with Code Execution](https://dominiklukes.net/teaching/data-analysis-with-code-execution) (Teaching, undated)
- [Developing a vocabulary for thinking about AI](https://dominiklukes.net/teaching/developing-a-vocabulary-for-thinking-about-ai) (Teaching, undated)
- [Exploring image and multimedia generation](https://dominiklukes.net/teaching/exploring-image-and-multimedia-generation) (Teaching, undated)
- [Exploring Model Performance with MPI Explorer](https://dominiklukes.net/teaching/exploring-model-performance-with-mpi-explorer) (Teaching, undated)
- [Exploring Small/Local Models](https://dominiklukes.net/teaching/exploring-small-local-models) (Teaching, undated)
- [How LLMs Work](https://dominiklukes.net/apps/how-llms-work) (Apps, undated)
- [Long Context Window](https://dominiklukes.net/teaching/long-context-window) (Teaching, undated)
- [Model Context Explorer](https://dominiklukes.net/apps/model-context-explorer) (Apps, undated)
- [Model Vibe Check](https://dominiklukes.net/apps/model-vibe-check) (Apps, undated)
- [MondAI](https://dominiklukes.net/webs/mondai) (Websites, undated)
- [Multimodal Conversation](https://dominiklukes.net/teaching/multimodal-conversation) (Teaching, undated)
- [Practical AI](https://dominiklukes.net/resources/practical-ai) (Guides, undated)
- [Schemas & Propositions](https://dominiklukes.net/webs/schemas-and-propositions) (Websites, undated)
- [Search Grounding](https://dominiklukes.net/teaching/search-grounding) (Teaching, undated)
- [Semantic Machines (Notion)](https://dominiklukes.net/webs/semantic-machines-notion) (Websites, undated)
- [Semantic Machines workshops and activities](https://dominiklukes.net/webs/semantic-machines-workshops-and-activities) (Websites, undated)
- [Skills Learning Companion](https://dominiklukes.net/resources/skills-companion) (Guides, undated)
- [Testing Reasoning Capabilities](https://dominiklukes.net/teaching/testing-reasoning-capabilities) (Teaching, undated)
- [Testing the Semantic Baseline](https://dominiklukes.net/teaching/testing-the-semantic-baseline) (Teaching, undated)
- [Tool Use: Counting & Math](https://dominiklukes.net/teaching/tool-use-counting-math) (Teaching, undated)
- [Visualizing Context](https://dominiklukes.net/teaching/visualizing-context) (Teaching, undated)

All topics: [Topics](https://dominiklukes.net/topics)