Sebastian Raschka

speaker

1,024 appearances 1 recordings 1 series first heard Feb 2026 last heard 1 Feb

Sebastian Raschka’s voice in public audio — every appearance, attributed to the second.

Trend

recordings per month · last 12 months

Recordings per month over the last 12 months — 1 in all, peaking in Feb 2026 with 1.

Appearances

newest first · ▶ plays the moment

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

So I think it's a huge unlock, which I think right now is not fully utilized yet by the open source, open-weight ecosystem.

A lot of people

don't use tool call modes because I think it's first, it's a trust thing.

You don't want to run this on your computer where it has access to tools, could wipe your hard drive or whatever.

So you want to maybe containerize that.

But I do think, you know, that is like a really important step for the upcoming years to have this ability.

Yeah.

For the user, I think also, I mean, there are users who just use the model locally, how they would use CGPD, but also for companies, I think it's a huge unlock to have these models because you can customize them, you can train them, you can add post-training, add more data, specialize them into, let's say, law, medical models, whatever you have.

And the appeal, you mentioned Lama, the appeal of the open weight models from China is that the open weight models are also, the licenses are even friendlier.

I think they are just unrestricted open source licenses where if you use something like Lama or Gemma, there are some strings attached.

I think it's like an upper limit in terms of how many users you have.

And then if you exceed, I don't know, so many million users, you have to report your finance situation to, let's say, Meta or something like that.

And I think, well, it is a free model, but there are strings attached and people do like things where strings are not attached.

So I think that's also one of the reasons, besides performance, why the open weight models from China are so popular, because you can just use them.

There's no catch in that sense.

Maybe we can go chronologically.

I mean, there was, of course, DeepSeek R1 that came out in January, if we just focus on 2025.

However, this was based on DeepSeek version 3, which came out the year before in December 2024.

There are multiple things on the architecture side.

What is fascinating is you can still, I mean, that's what I do in my from scratch coding projects.

Showing 101–120 of 1,024 · page 6 of 52