High Signal Podcasts Evidence ledger
Method
Browse
← Back to evidence

Evidence receipt / commitment

Published · transcript-backed

Nathan Lambert: commitment

1 Feb 2026 Lex Fridman Podcast #490 – State of AI in 2026: LLMs, Coding, Scaling Laws, China, Agents, GPUs, AGI

“I will regularly have five pro queries going simultaneously, each looking for one specific paper or feedback on an equation.”

— Nathan Lambert

Source trail

Everything needed to verify it.

Speaker
Nathan Lambert
Attribution
Verified speaker
Claim type
commitment
Recorded
1 Feb 2026
Publisher
Lex Fridman Podcast

Transcript context

…I think it’s a nice variety actually, or the option to have a toggle there. For my personal usage, most of the time when I look something up, I use ChatGPT to ask a quick question and get the information I wanted fast. For most daily tasks, I use the quick model. Nowadays, I think the auto mode is pretty good where you don’t have to specifically say “thinking” or “non-thinking.” Then again, I also sometimes want the pro mode. Very often, when I have something written, I put it into ChatGPT and say, “Hey, do a very thorough check. Are all my references correct? Are all my thoughts correct? Did I make any formatting mistakes? Are the figure numbers wrong?” or something like that. And I don’t need that right away. I can finish my stuff, maybe have dinner, let it run, come back and go through it. This is where I think it’s important to have this option. I would go crazy if for each query I had to wait 30 minutes, or even 10 minutes. That’s me. I’m sitting over here losing my mind that you use the router and the non-thinking model. I’m like, “How do you live with that?” That’s like my reaction. I’ve been heavily on ChatGPT for a while. I never touched GPT-5 non-thinking. I find it just… its tone and then its propensity for errors. It just has a higher likelihood of errors. Some of this is from back when OpenAI released o3, which was the first model to do this Deep Research and find many sources and integrate them for you. So I became habituated with that. I will only use GPT-5.2 thinking or pro when I’m finding any sort of information query for work, whether that’s a paper or some code reference. I will regularly have five pro queries going simultaneously, each looking for one specific paper or feedback on an equation. I have a fun example where I just needed the answer as fast as possible for this podcast before I was going on the trip. I have a local GPU running at home and I wanted to run a long RL experiment. Usually I unplug things because if you’re not at home, you don’t want to have things plugged in, and I accidentally unplugged the GPU. My wife was already in the car and it was like, “Oh dang.” Basically, I wanted a Bash script as fast as possible that runs my different experiments and the evaluation. I know how to use the Bash terminal, but in that moment I just needed the command in 10 seconds.…

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

Search evidence