pre-training

5 statements across 5 episodes · 3 bullish · 0 bearish · 5 people on the record · first statement Apr 25, 2023 by Dr. Percy Liang · across every show →

Everything said about pre-training, oldest first

Apr 25, 2023 positive
Insight
Liang: Controlling AI hallucination is easier once models understand the concept
“So I think there's pre-training, which is predicting the next word and developing a world model, so to speak. And with those capabilities, then you can, you still have to say don't hallucinate, but it will be much easier to control that model if it has a notio…”
Dr. Percy Liang Apr 25, 2023 ▶ 26:56 No Priors Ep. 7 | With Stanford Professor Dr. Percy Liang
Aug 1, 2024 bullish
Prediction Not checkable as stated
Vinyals predicts pre-training compute will drop to ~50% as RL expands
“So to me, that balance feels correct, like some on pre-training, and here we, we're trying to learn every task. So certainly that's going to be, you know, let's say it can be as high as 50%, not as high as over 90 like today. And then the rest mostly on reinfo…”
Oriol Vinyals Aug 1, 2024 ▶ 25:29 No Priors Ep. 74 | With Google DeepMind VP of Research Oriol Vinyals
Sep 19, 2024 neutral
Prediction Not checkable as stated
Taylor: AI pre-training will consolidate into a small number of frontier model builders
“I think that will probably play out with the frontier models. We'll end up with a relatively small number of companies doing pre-training you know, which is the really capital intensive part of model building not because, you know, there's not, they're the onl…”
Bret Taylor Sep 19, 2024 ▶ 15:41 No Priors Ep. 82 | With CEO of Sierra Bret Taylor
Nov 21, 2024
Insight
Fine-Tuning Cannot Effectively Add New Languages to Large Language Models
“There's just no way you can do that without intervening on pre-training. You can't like fine tune or post train Japanese into a model effectively. And so you have to start from scratch.”
Aidan Gomez Nov 21, 2024 ▶ 18:50 No Priors Ep. 91 | With Cohere Co-Founder and CEO Aidan Gomez
Jul 17, 2025 bullish
Insight
Laskin: RL requires far fewer FLOPs than pre-training for frontier models
“We're in this brief period in history right now where the RL flops are still manageable. Like you can really have a best in class product if you're focused. And yes, you'll need to put, you know, you still need a decent amount of GPUs, but from a flops perspec…”
Misha Laskin Jul 17, 2025 ▶ 30:41 No Priors Ep. 123 | With ReflectionAI Co-Founder and CEO Misha Laskin
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 100 episodes transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.