Viewing profile — scolvin

This guy has nice hair.

Please report any panics, we'll fix them!

No one is going to let an LLM get prompted by end users to write python code I just run on my server, there's no real debate on that.

As discussed on twitter, v8 shows that's not true. But to be clear, we're not even targeting the same "computer use" use case I think e2b, daytona, cloudflare, modal, fly.io, deno,…

Can't be sure where this might end, but the primary goal is to enable codemode/programmatic tool calling, using the external function call mechanism for anything more complicated. …

this is pretty performant for short scripts if you measure time "from code to rust" which can be as low as 1us. Of course it's slow for complex numerical calculations, but that's t…

ROFL. I clearly didn't build pydantic "to take VC money" - I maintained Pydantic for 5 years before deciding to take VC money. I wanted to build logfire and I didn't want to compet…

We have a SaaS platform (Pydantic Logfire - General and AI observability), and we raised our Series A from Sequoia. For good or bad, I think we're pretty "SaaS/vc/etc." already.

Correction, logging used to suck - now it's fixed https://pydantic.dev/logfire :-)

Pydantic creator here - I kind of agree with the article. I (obviously very biased) use Pydantic a fair bit, but there are places where it's the wrong tool and I use dataclasses, t…

Pydantic author here. We have plans for an improvement to pydantic where JSON is parsed iteratively, which will make way for reading a file as we parse it. Details in https://githu…

See https://github.com/pydantic/pydantic-ai/issues/112 people have tried quite a few models. Llama3.2 worked well and used <2gb ram

Interesting, but this doesn't support multiple tools with different arguments, or an arbitrary number of steps in the conversation. If you add that, you basically get PydanticAI.

Thank you so much!