AI web research reconstructs lost history of a demolished building from scattered archives
flowersslop · x · 2026-09-15
The author used Astra's web research capabilities to investigate a long-demolished building in his city, once semi-famous as a filming location. Reddit, forums and old fan pages offered little; even archive nerds couldn't piece it together. Astra dug through 200-year-old city archives, obscure documents, and even a private Facebook post containing a decades-old aerial photo where the building, though tiny, answered questions no other known photo could.
After cross-referencing and filtering noise, Astra produced genuinely new knowledge about something lost to history—not because evidence never existed, but because no one had connected the fragments. The author argues models can reconstruct huge amounts of lost historical knowledge this way, but most results vanish when chats end instead of entering what future models know. He hopes OpenAI is pursuing this kind of reconstruction at scale.
More from AGI Musings
- Columbia economist daveholtz takes leave to lead AGI economics research at OpenAI — daveholtz · 2026-09-15
- KP: Personal AI agents will build 1000x more apps than humans; AX beats UI — thisiskp_ · 2026-09-15
- Chollet: by experience-to-competence efficiency, current AI is ~6 OOMs behind humans — GregKamradt · 2026-09-15
- Article argues intelligence has a speed limit, curbing recursive self-improvement hype — docmilanfar · 2026-09-15
- Banning data centers to save the world? A 500-year history lesson says otherwise — thursdai_pod · 2026-09-15
- Why Is LeCun So Unconcerned About AI Loss-of-Control Risks? — IndependentFresh628 · 2026-09-15