Search for content...
A field guide
Seventy-five years of AI, from Turing's thought experiment to agents that use your computer. Built for developers who want the shape of the field, not just the latest headline.
ERA ONE
When humans first asked: can machines think?
1950–1979
Where the idea begins. Turing asks whether machines can think; Dartmouth gives the field its name.
ERA TWO
The first time AI was sold to businesses — and the first time it disappointed.
1980–1985
Symbolic AI finds its first commercial footing. Hand-crafted rules do narrow, useful work.
ERA THREE
A handful of researchers refused to give up on neural networks.
1986–2011
Backpropagation quietly matures. Big-company attention drifts; the groundwork gets laid.
ERA FOUR
GPUs + data + old ideas = a revolution nobody expected.
2012–2016
AlexNet wins ImageNet. Scale, GPUs, and data turn neural nets into a practical tool.
ERA FIVE
One paper changed everything: Attention Is All You Need.
2017–2021
Attention is all you need. The Transformer unlocks foundation models; GPT and BERT arrive.
ERA SIX
AI learned to speak — and 100 million people started listening.
2022–2023
ChatGPT hits a million users in five days. Claude and GPT-4 follow. AI becomes public infrastructure.
ERA SEVEN
AI stopped waiting for instructions and started taking action.
2024–2026
Models stop answering questions and start taking actions. MCP, tool use, and autonomous workflows.
A field guide
Six quick questions. No email, no sign-up. We'll suggest what to read next.
Question 1 of 6
Have you used an AI model in the last week?
ERA EIGHT
You're here. What happens next depends on who understands this technology.
ongoing
Unwritten.