High Signal Podcasts Evidence ledger
Method
Browse
← Back to evidence

Evidence receipt / belief

Published · transcript-backed

Trenton Bricken: belief

28 Mar 2024 Dwarkesh Podcast Sholto Douglas & Trenton Bricken — How LLMs actually think

“Can you define that? Because when I hear it, I think “if else” statements for symbolic logic.”

— Trenton Bricken

Source trail

Everything needed to verify it.

Speaker
Trenton Bricken
Attribution
Verified speaker
Claim type
belief
Recorded
28 Mar 2024
Publisher
Dwarkesh Podcast

Transcript context

…Will GOFAI be part of the intelligence explosion? You talked about synthetic data, but in fact it would be writing its own source code in some important way. There was an interesting paper that you can use diffusion to come up with model weights. I don't know how legit that was or whatever, but something like that. So GOFAI is good old-fashioned AI, right? Can you define that? Because when I hear it, I think “if else” statements for symbolic logic. I actually want to make sure we fully unpack the model improvement increments. I don't want people to come away with the perspective that this is super bearish and models aren't going to get much better. I want to emphasize that the jumps that we've seen so far are huge. Even if those continue on a smaller scale, we're still in for extremely smart, very reliable agents over the next couple of orders of magnitude. We didn't fully close the thread on the narrow window thing. Let's say GPT-4 cost a hundred million dollars or whatever. You have the 1B run, 10B run, 100B run. All seem very plausible by private company standards.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.

Search evidence