AI Solving Decades-Old Research Problems — and Eroding Human Skills
Ex-OpenAI researcher Scott Aaronson reports AI solved open problems from his 25-year-old papers rapidly, but warns the same tool is atrophying his own mathematical abilities.
What we know
An account on X (Sabine Hossenfelder) amplified a viral thread where Aaronson describes feeding old open problems into an AI and getting rapid solutions — but highlights the double-edged nature: the tool may be making human researchers cognitively dependent. A separate viral RT by another account on X — 'CLAUDE FOUND THE SEEDS' with 977 RTs — also points to AI producing surprising, concrete research results.
By the Global Eye 24 desk, with AI assistance. Updated as the story develops. Spot an error? Write to corrections@globaleye24.com.