“Because of agent training data, it's very hard to deprecate things now. We see that agents often love using old versions of Modal syntax. And it's kind of hard to nudge them away from that. So we've actually had to be a little bit more conservative ... with how we deprecate and remove features.”

Latent Space · AI Builders · October 2026

Asked what Modal changed in its product now that so many users are agents, Bernhardsson said old syntax lives on in model training data. Modal now has to update every example online before removing a feature, and it runs internal tests to see whether agents recommend it.

Latent Space · 2026-10-22 Listen to the episode → More from Erik Bernhardsson → More AI Builders quotes →

Transcript

Latent Space

Speaker 2

…with it tear the sandbox down right like rather than sort of declarative using using you know some sort of dsl it also pushes like a little bit of a different interface that maybe previously didn't exist i think in either way like sdks are as relevant as they would as they always were if not more

Speaker 1

is there anything you changed in your product just because like you know most people are just running agents on on it

Speaker 2

we do uh i mean we also we actually do run a lot of internal tests where we have agents we tell agents To build a modal app and run it. And I mean, a very concrete thing that is kind of funny is that we've actually seen that because of agent training data, it's very hard to deprecate things now. We see that agents often love using old versions of modal syntax. And it's kind of hard to nudge them away from that. So we've actually had to be a little bit more conservative going forward. Over the last year or so with how we deprecate and remove features, we have to be very careful to update every single example that's out there online. New skills, right? Yeah, there's stuff like that. But there's also an insane amount of modal code that's now online too, that goes into the training data. I've seen that to be kind of an interesting example where another thing, by…

Speaker 1

do about that? Yeah, you say, hey, build a stable diffusion thing. Does it recommend modal or not? What do you do about that?

Speaker 2

Yeah, I think that's harder. I think, you know, it's like search engine optimization. I think it's primarily through content. It's like you just have to put out a lot of content. Interesting. You can also, to some extent, see the perception, I think, a little bit more. You can ask an agent, rank…

Speaker 1

Have you thought about just making a bunch of GitHub repos and having the agents write a bunch of modal code in there? Have you tried that? Slot

Speaker names from our own diarization · position estimated from where the line sits in the episode

More from Latent Space

“My view is code reviews in large part is people in many cases trying to just escape the accountability of merging things into production. And if you have that culture, then to me, that's a sign that you need to invest way more in CD and CI.” — Erik Bernhardsson, Latent Space

“Right now, if you took a B300 and you tried to run inference on it, you can make $40 an hour or so selling any model, basically. And so that's if you use it at 100%. And maybe if you only have like 50% utilization, so like $25 an hour. Meanwhile, you can buy B300s at like $5 to $6 an hour. One of these numbers has to change. It's a huge arbitrage.” — Evan Conrad, Latent Space