Kyle Corbitt
speaker
658 appearances
1 recordings
1 series
first heard Oct 2025
last heard 16 Oct
Kyle Corbitt’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsRecordings per month over the last 12 months — 1 in all, peaking in Oct 2025 with 1.
Appearances
And if you have, and if you have like a very complex system and you're doing RL and you need to set up a whole environment, it could be a lot longer.
It could be a couple of months of time.
So that's just like a fixed cost you have to pay.
There's also like an ongoing carrying cost where once you've committed to doing fine-tuning, it does make other parts of your stack less flexible, less nimble, because whenever you're updating your prompt or like you're adding new content,
context or whatever, like now you have to like, you know, spend a few hours training a model and that's just gonna like slow down your your iteration cycle, which is a real cost.
And in many cases that's the larger cost.
So you only want to do that if like the benefits are large enough.
The dollar cost I would say is
basically never a factor.
Um it's just so much less than the time the amount you're spending this engineer to to do the work that it's not I mean it's you know, each of these runs is uh between five and a couple hundred dollars.
Um and it's just you you don't have to do that many of them.
Yeah, so that was a big moment for us, um, with you know, there's all the leaks before that about strawberry and all this, and like, you know, a lot of people talking about, okay, are they doing it?
Um we realized through that that like, okay, someone's figured out how to make RL actually work with LLMs, which was not a thing.
I mean, it was a thing that like some people had played around with before that, but it wasn't like a thing many people were thinking about.
And so our bet at that point was, yes, let's figure out whether this works for a
Specifically.
And the space we we just I think it's important to kind of like tease out different parts of the market.
I think with the release of 01, and this has been like proved out many times with releases since then, I think like there's now like a very strong consensus that, like, okay, on the frontier model, like general purpose model side, investments in RL are paying off.
I think that I I don't think most people would argue.
with that.
Showing 141–160 of 658 · page 8 of 33
← Previous
Next →