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

It might be faster.

And then now you have this dimension of the denoising process.

The more steps you do, the better the text becomes.

And people, you know, I mean, you can scale in different ways.

They try to see if that is maybe a valid alternative to the autoregressive model in terms of giving you...

the same quality for less compute.

Right now, I think it's, you know, there are papers that suggest, okay, if you want to get the same quality, you have to crank up the denoising steps and then you end up spending the same compute you would spend on an autoregressive model.

The other downside is, well, it's parallel, which sounds appealing, but some tasks are not parallel.

Like, you know, like reasoning tasks, tool use, maybe where you have to ask a code interpreter to give you an intermediate result.

And that is kind of tricky with diffusion models.

So there are some hybrids.

But the main idea is, can we parallelize it?

And so interesting avenue.

I think right now there are mostly research, let's say, models out there like LADA and some other ones.

I saw some by startups, some deployed models.

There is no big diffusion model at scale yet, like, you know, like Gemini Chachapiti scale in that level.

But there was an announcement by Google or like a site where they said they are launching Gemini diffusion and they put it into context of their, I think, Nano 2 model.

And they said basically for the same quality on most benchmarks, we can generate things much faster there.

So you mentioned what's next.

I don't think the text diffusion model is going to replace autoregressive LLMs, but it will be something maybe for quick, cheap at scale tasks.

Showing 661–680 of 1,024 · page 34 of 52