Plinz pushes back on unsupported claims that LLMs can't do X

Plinz · x · 2026-09-27

AI researcher Plinz (Albert) clarifies his position in an ongoing debate: he isn't claiming LLMs can do everything, but stays agnostic about the unknown. His real gripe is full-throated, evidence-free declarations that LLMs can't do X — and he says his recent post was deliberately designed to bait such pronouncements, which it did. He also floats an open question: whether an end-to-end trained AI converging to a causal structure similar to an evolved nervous system is really more extraordinary than the opposite conjecture.

Related event: Researcher Plinz Argues Rigorous Definition Needed to Rule Out LLM Thinking(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →