Tag: AI
-
GPT-4.5 passed a three-party Turing test in 2025 — four years ahead of Kurzweil’s 2029 bet. But the machine that crossed the line was not the one he predicted would.
-
Kurzweil predicted the Turing test would fall in 2029. GPT-4.5 passed it in 2025 — four years early, by a path he didn’t predict.
-
Kurzweil’s prediction that trained skills could move between machines in seconds arrived early, decoupled from biology. The nanobots never came.
-
Kurzweil bet that reverse-engineering the brain was the road to AGI. The intelligence arrived on a different road — and it runs on a learning rule he called biologically unrealistic.
-
Kurzweil predicted strong AI via brain reverse-engineering. It arrived via transformers and text. Every destination was right. Every route was wrong.
-
Kurzweil predicted human-level AI by the mid-2020s. He got the direction right and the mechanism completely wrong.