Eno Reyes
๐ค SpeakerAppearances Over Time
Podcast Appearances
Our view is that we're definitely not opposed to building models.
I think that the challenge is you want to build, you basically want to fine tune or build models as basically as late as possible before it becomes important to, right?
That's sort of our view.
And what I mean by that is,
You know, if you had fine-tuned GLM2 or some model that's earlier, you know, it gets blown away by the model providers at the next level, right?
Which is three weeks away, usually.
Exactly, exactly.
You probably have three weeks before there's a better model out, and you have a month to two months before it's completely forgotten about, right?
Yeah.
So if your bet is we're going to beat the people with $50 billion data centers, I think that you have to have a very clear strategy.
And I think that there are reasons to do this, by the way, cost, right?
If you know that you can execute on a specific task at the max that a human cares about, and you just need to bring down cost, that's a great reason to fine tune a model, right?
Yeah.
Yeah.
but at the end of the day, uh, I think that we've already seen this, a bunch of, you know, software development organizations, uh, or, you know, software development, AI tooling, uh, they've all trained their own models and, you know, within two weeks to Gemini flash came out and it's like better, uh, or within three weeks later, 4.5 Opus is now like, not just better, but it's like two orders of magnitude better.
Uh, and so I think that it's going to be a really challenging battle to fine tune or build your own models, uh, in the near future, but we're definitely not ruling it out.
Um,
On the earlier point about being an agent research lab and sort of like how we tune the agent's behaviors, and in particular for seeking verification or validation, there's a lot that you can do at the harness level to enable this, right?
Some of it might be like context engineering.
So adjusting not just like the base system prompt, but reminders that you can insert automatically or, you know, prebuilt environmental reactions to tool calls.