Dev finds frontier models can't invent novel solutions — human intuition still wins the loop
karmay007 · x · 2026-10-14
A developer threw a very hard problem at Astra Ultra and fable 5.1 ultracode; neither produced novel ideas to fix it. He came up with the original solution himself, then had agents evaluate every part of it. His takeaway: human-in-the-loop remains essential — agents still can't theorize novel approaches on their own.
More from AGI Musings
- Matthew Green: AI hands attackers exploit development once reserved for nation-states — matthew_d_green · 2026-10-14
- Cryptographer Matthew Green: AI will scale cyberattack damage 15,000x and a deadly catastrophe is coming — matthew_d_green · 2026-10-14
- NLP researcher Vered Shwartz talks AI on the Ship of Fools podcast — VeredShwartz · 2026-10-14
- Musicians clash over AI art: some people can't judge art objectively at all — nitarshan · 2026-10-14
- 'Does Claude Have Rights' Discussion Goes Live After David Sacks Retweet — NathanpmYoung · 2026-10-14
- AI is 'burning the coal' of unsolved math problems that train future mathematicians — JnBrymn · 2026-10-14