UCLA Professor: LLM complexity is now beyond understanding
Tolopono · reddit · 2026-08-25
A UCLA Professor expressed on social media that Large Language Models have become so advanced that their internal mechanisms are no longer understandable, even to researchers. He noted this is occurring before models like Astra have even been seen, and has published a paper detailing his perspective.
More from AGI Musings
- Training the 'Other Mind' to disappear won't solve alignment — repligate · 2026-08-25
- Don't let math models become slop cannons like image/video models — aiamblichus · 2026-08-25
- How AI Agents Will Transform Mathematical Proof Workflows by 2026 — RexDouglass · 2026-08-25
- AI Hallucinations Effectively Solved in Most Domains, Says Expert — davidmanheim · 2026-08-25
- 1964 prophecy: Machines will eventually surpass humans as biological evolution ends — RichardSSutton · 2026-08-25
- What business models fail when AI makes checking very cheap? — kach_janani · 2026-08-25