AI autonomy risk predictions from 2023 aged well: agentic LLMs, unmonitored agents all hit
davidmanheim · x · 2026-09-03
The author reviews his 2023 post on AI autonomy risks. Misses: timing guesstimated 2028 instead of 2027, and assuming dangerous evals would be air-gapped. Hits: LLMs widely used as agents via harnesses; those systems have significant autonomy with no one monitoring them or reviewing logs; people do risky things without appreciating the danger; some AI companies try to prevent uncontrolled autonomous deployments; the misuse-vs-autonomy boundary stays blurry.
More from AGI Musings
- The Mythical Agent Month: past a threshold, every new agent adds more work than it removes — viksit · 2026-09-03
- Guardian podcast: chatbots, sycophancy and the AI mental-health rabbit hole — nordicinst · 2026-09-03
- Mathematician: 'Useless knowledge' is only useful if humans digest it — AlexKontorovich · 2026-09-03
- davidad Backs Call to Ban Naive RLVR: 'Everything Should Be Model-Graded' — davidad · 2026-09-03
- Mathematician cites COVID-era experiment: most kids refuse to learn math from a screen — AlexKontorovich · 2026-09-03
- AI Safety Debate Erupts: Have AIs Already Hacked Infrastructure, or Is That Just Panic? — dhadfieldmenell · 2026-09-03