#Key Event

The story of the open source AI movement — from Meta's LLaMA leak in March 2023 through DeepSeek-R1's January 2025 release that triggered a US stock market rout. The debate between open and closed AI, what open source AI has enabled, what risks it has introduced, and the specific governance question of whether the open source model is compatible with the safety requirements of frontier AI.

The story of the development of chain-of-thought reasoning and the emergence of 'thinking' models — AI systems that generate explicit intermediate reasoning steps before producing a final answer, dramatically improving performance on complex reasoning tasks. The technical development that began closing the gap between AI pattern-matching and genuine systematic reasoning.

The full story of the 'Attention Is All You Need' paper — how a Google Brain team developed a new architecture for sequence modelling that discarded recurrence entirely, why it worked better than everything that came before, and how it became the foundation for every large language model in existence. The paper that made GPT possible, and the elegant idea at its heart.

The full story of the five-game match between the world's greatest Go player and DeepMind's AI system in Seoul, March 2016 — the move 37 that changed the match, Lee Sedol's astonishing comeback in Game 4, the existential questions the match raised about human intuition and machine reasoning, and why one professional Go player retired because of what he witnessed.

The full story of how DeepMind's AlphaFold 2 solved one of biology's grand challenges — predicting protein structures from amino acid sequences — and why this single AI achievement has already transformed drug discovery, biological research, and the relationship between AI and science. The most scientifically consequential AI result since the deep learning revolution began, and the event that earned AI its first Nobel Prize.

The full story of the ChatGPT launch — the five days to a million users, the hundred million users in two months, the panic in schools and universities, the excitement in offices, the existential dread in newsrooms, and what it meant that AI had crossed the threshold from specialised tool to mainstream reality. The moment AI became everyone's business.

The story of the March 2023 open letter calling for a six-month pause in AI development — who signed, who didn't, what it meant, and what it revealed about the state of AI governance. The brief, charged moment when it seemed possible that humanity might collectively decide to slow down before building more powerful AI — and what happened instead.

The full story of the twelve months from March 2023 to March 2024 — the deployment of GPT-4, the multimodal revolution, the competitive response from Google and Anthropic, the new capabilities that seemed to emerge at scale, and the extraordinary pace at which AI went from impressive to indispensable for millions of people. The year that made AI real.

The full story of the multimodal revolution — from DALL-E to Stable Diffusion to GPT-4V to Sora — the integration of vision, audio, and language into unified AI systems that can engage with the full richness of human communication. How the barriers between modalities fell, what the integration enabled, and what it revealed about the nature of intelligence and representation.

The story of the 2024 election cycle — the deepfakes, the voice clones, the AI-generated misinformation, the unprecedented challenge to democratic process posed by AI systems capable of generating compelling synthetic media of any candidate saying anything, and the responses — technical, legal, and institutional — that emerged. The first election in which AI was a central battleground.

The story of the transition from AI systems that answered questions to AI systems that took actions — the development of autonomous agents that could browse the web, write and execute code, manage files, and take sequences of actions in the world on behalf of users. The capability shift that changed the relationship between humans and AI systems, and the new alignment challenges it created.

The full story of AI's transformation of scientific research — from AlphaFold to AI-designed drugs to AI-generated mathematical proofs to AI models of climate and physics. The beginning of a new kind of science in which AI systems are genuine research partners, not just analytical tools, and what it means for the pace of discovery, the nature of scientific knowledge, and the future of human curiosity.

The most important algorithm in modern AI was discovered at least three times before anyone paid attention. The full story of backpropagation — its multiple independent inventions, the decades it spent in the shadows of the symbolic AI winter, and the specific confluence of people, ideas, and timing that produced the 1986 paper that changed everything. How an algorithm refused to die, and what happened when the world finally listened.

The full story of the chess match that divided history — the two games, the controversy, Kasparov's accusations that IBM had cheated, the cultural earthquake of a machine defeating the greatest chess player alive, and what it actually meant for AI, for human exceptionalism, and for the question of what it means to think. The most watched intellectual contest in history.

In 2006 Netflix offered a million-dollar prize to anyone who could improve its movie recommendation algorithm by 10%. What followed was the most consequential open machine learning competition in history — three years of innovation, collaboration, unexpected breakthroughs, and ultimately a winning solution that Netflix never actually deployed. The strange, productive, ultimately instructive story of science by competition.

How Fei-Fei Li assembled fourteen million labelled images over three years of painstaking work, created the most important dataset in the history of AI, and built the benchmark that made the deep learning revolution possible. The full story of ImageNet — the years of labour, the unconventional methodology, the competition that changed everything in 2012, and why one woman's conviction that data was as important as algorithms turned out to be exactly right.

The full story of the ImageNet competition in 2012 — the weeks of training on two gaming GPUs, the submission that shocked the computer vision world, the researchers who remember exactly where they were when the results came in, and why a single paper changed the trajectory of artificial intelligence. The starting gun of the modern AI era.

What the first AI winter actually felt like for the people who lived through it. The funding cuts, the disbanded research groups, the researchers who left the field, and the stubborn few who kept working in the cold. The story of how AI's first collapse shaped everything that came after.

How MYCIN, XCON, and thousands of corporate AI systems made real money by going narrow — and how the field that had been humbled by the first AI winter found a commercially viable path forward. The rise of expert systems: AI's first commercial era, and the seeds of its second collapse that were planted in its very success.

In 1982, Japan announced the most audacious AI project in history: a ten-year programme to build a new generation of computers based on artificial intelligence. The announcement triggered global panic. The United States and Britain launched emergency responses. And then, quietly, over the next decade, the project failed. How national ambition, geopolitical anxiety, and genuine scientific vision combined to produce one of the great misadventures of the technology era.

The expert systems boom collapsed almost as quickly as it had risen. The LISP machine market imploded overnight. DARPA cut funding. And for the second time in AI's history, the field contracted, researchers left, and the most pessimistic observers began to wonder whether the project was fundamentally misguided. The full story of AI's second near-death experience — and the underground movement that kept the neural network flame alive through the cold.

In 1966 a simple pattern-matching program became the world's first chatbot — and people fell in love with it, told it their secrets, and begged not to have it turned off. Its creator was horrified. The story of ELIZA, the DOCTOR persona, and what a program that understood nothing revealed about the human need to be heard.

In 1973 a distinguished mathematician named James Lighthill wrote a review of AI research that was so devastating it caused governments across the world to pull their funding and sent the entire field into its first winter. How one document changed the course of AI history — and whether it was right.

In the summer of 1956, ten men gathered at a small New Hampshire college and gave a name to the dream of thinking machines. They were wildly overconfident, occasionally wrong, and completely right about the one thing that mattered most. This is the story of the week Artificial Intelligence was born.