Yi Tay

speaker

925 appearances 1 recordings 1 series first heard Jan 2026 last heard 23 Jan

Yi Tay’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 Jan 2026 with 1.

Appearances

newest first · ▶ plays the moment

Latent Space: The AI Engineer Podcast · Captaining IMO Gold, Deep Think, On-Policy RL, Feeling the AGI in Singapore — Yi Tay · 23 Jan 2026 podcast

point where you are very data bound but not compute bound or you just find algorithms that spend a lot of compute on every token on every token.

So I think

So the uh the overarching point is just that okay, that th that it's a learning algorithm thing for data efficiency and then if whether the correct way is actually just to apply more flops per token, just to squeeze out more from every

Also because humans actually don't like like they are exposed less.

When you say less or more data, it's also very ambiguous because they are technically like on twenty twenty-four seven and then you catch mostly visual.

Yeah.

And whether they actually spend more flops on everything that listen is also a question because maybe they are just beta efficient just because I mean I didn't somebody need to count like how much flops the brain used to process like how much maybe there's

just spending more compute on every token and also maybe the learning algorithm is different.

So and but I believe that data efficiency is very important given that that I think we're gonna like there's a limited amount of data in the world.

R environments that are like that have a lot of domain expertise are probably very valuable.

And actually I I don't know specifically about what RR environments people are actually specifically buying.

But what what's the thing that you're not satisfied by?

Like the It was

I don't know.

LMs are just used everywhere now.

But whether it's actually like a Like a big LM.

Like whether it's like a general retrieval type of model is like it's like another uh question.

Correct.

No, I don't care about IR, but I think DSI, okay, like DSI or genitive retrieval was like I think one of my favorite works in the old of like uh I I have some IR background in like when I was doing a PhD, I did some recommends work, I did some retrieval work with Rex and stuff like that.

So I have some IR and Rex background.

Showing 641–660 of 925 · page 33 of 47