AI's story is a cycle of soaring hype, crushing "winters," and an eventual breakthrough that changed everything.
The timeline
| Era | What happened |
|---|---|
| 1950 | Turing's "Computing Machinery and Intelligence" — the Turing Test |
| 1956 | Dartmouth workshop coins "artificial intelligence" |
| 1960s–70s | Symbolic AI / expert systems; big promises |
| 1974–80 | First AI winter — funding collapses as promises fail |
| 1980s | Expert-systems boom… then bust |
| 1987–93 | Second AI winter |
| 1997 | Deep Blue beats Kasparov at chess |
| 2012 | AlexNet — deep CNNs crush image recognition; the deep-learning era begins |
| 2017 | "Attention Is All You Need" — the transformer |
| 2020s | LLMs (GPT, Claude, …) and reasoning models |
Why it kept stalling
Early AI relied on hand-written rules — brittle and unscalable. Two winters came when reality fell short of hype and the money dried up. The thaw came from a different bet: learn from data, plus enough compute and data to make it work.
The lesson the winters teach: separate genuine capability from hype. Today's progress is real — but the field has been "about to solve intelligence" before. A useful caution for claims about the singularity.
Related: What Is Machine Learning · The Turing Test · The Technological Singularity