Philosopher Revisits 2019 AI Predictions: Right on GPT-2, Wrong on LLM Understanding

Philosopher Henry Shevlin re-reads his 2019 essay on AI and grades himself B+, admitting he underestimated LLMs' capacity for genuine understanding despite correctly spotting GPT-2's potential. The original piece used everyday folk psychology, like an Uber driver reading tourists' intentions, to probe whether machines can truly understand people.

2026-10-06 ~ 2026-10-06 · 2 related posts

• Henry Shevlin grades his 2019 AI predictions: right on GPT-2, wrong on LLM understanding — dioscuri · 2026-10-06
• Shevlin's 2019 essay revisited: Xu Zhimo, fan fiction, and the Winograd Schema Challenge — dioscuri · 2026-10-06