Kamradt: average general intelligence is easy to assert, outlier intelligence is sparse but testable
GregKamradt · x · 2026-09-01
Following his thread on asserting general intelligence, Greg Kamradt draws a distinction: the average human's general intelligence shows in smooth, low-surprise situations — in a brief conversation their output isn't far off a well-prompted modern AI, so it's hard to assert individually; outliers like Terrence Tao display intelligence in deep, narrow domains, which is sparser but easier to test. Augmenting knowledge work materially needs the former, though the latter is easier (if sparser) to evaluate. He notes general intelligence can only be asserted at two levels: humanity's collective tech progress over history, and a single person's output over a lifetime.
Related event: Kamradt: Asserting AGI in AI Is Getting Harder(2 posts)→
More from AGI Musings
- Governments Must Find Ways to Distribute AI Dividends — LuizaJarovsky · 2026-09-01
- UK's BOLD Lab is hiring a BOLD Fellow: massive compute plus full academic freedom — joelbot3000 · 2026-09-01
- Kids' Brains Already 'Fried' by Age 2, AI Just Continues Trend — kevinafischer · 2026-09-01
- Closing the Loop: The Key to Autonomous Scientific Discovery — marinkazitnik · 2026-09-01
- Nate Retracts Overconfident Claims on LLM Introspection — joshua_saxe · 2026-09-01
- Guardian: Can we stop AI from deceiving us? — tw1st3d_m3nt4t · 2026-09-01