Post

LLMs can pass the Turing Test. That, not their successes on math and code, is why we can expect their general abilities to continue to make progress. The Turing Test by definition requires general ability, and Turing was very right to focus on that.

#BruceSterling@bruces.bsky.social

*I don't get why any "general" intelligence automatically appears in the process here. Maybe you just get huge oversupplies of math and code. Like, to the unpleasant levels of endless traffic jams and ear-splitting noise.

22/10/26, 10:02 pm0 reposts0 likes

A. H. Zakai

@kripken.com

Also, progress is inevitable once you get to a certain level on the Turing Test. If a computer can provide a useful, relevant answer to a general prompt, it will do that 2x faster when hardware is 2x faster. And if you run 2 agents, it is yet another general problem to pick the best between them.

A. H. Zakai

@kripken.com

Math and code are two tasks that are easy to verify, but that we can't objectively measure LLM ability on, say, summarization of the main points in a novel, doesn't mean it isn't doing very well there.

A. H. Zakai

@kripken.com

The main way this could be false is if LLMs were not actually passing the Turing Test. That we are tricked by superficial fluency into accepting incoherent output (anthropomorphization, like with ELIZA). This is a major skeptical argument, but the evidence for it is mixed at best.

Sean Trott

@seantrott.bsky.social

Yeah, while it's certainly true that we're prone to anthropomorphization, I think the fact that the TT paper explicitly compares LLMs to ELIZA is pretty compelling evidence that this isn't all that's going on. ELIZA here feels like a good baseline for that!