Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Astral to Join OpenAI

astral.sh

Post reply on HN
Print view View original

391–400 of 951 posts

Re: Astral to Join OpenAI

#391

Post by tgtweak »

zanie wrote: post #225
rvnx wrote: post #170

Earlier quoted context omitted.

They raised 4M USD, they have 26 full-time employees (they pay 120 200K / yr, cf https://pitchbook.com/profiles/company/523411-93 ). It means the company almost reached their runway, so all these employees would have to find a job. It's a very very good product, but it is open-source and Apache / MIT, so difficult to defend from anyone just clicking on fork. Especially a large company like OpenAI who has massive dist…

A brief note, your numbers are way off here — Astral subsequently raised a Series A and B (as mentioned in the blog post) but did not announce them. We were doing great financially. (I work at Astral)

I mean you pirouetted onto the AI hype train before running out of working capital - I guess that's doing great financially by some definitions.

Re: Astral to Join OpenAI

#392

Post by Syntaf »

throwaw12 wrote: post #322
throwaway63467 wrote: post #194

Earlier quoted context omitted.

It’s a small tool shop building a tiny part of the Python ecosystem, let’s not overstate their importance. They burned through their VC money and needed an exit and CLI tool chains are hyped now for LLMs, but this mostly sounds like an acquihire to me. Dev tools are among the hardest things to monetize with very few real winners, so good for them to get a good exit.

uv and ruff is not tiny part anymore, its growing fast

Not to mention their language server + type checker `ty` is incredible. We moved our extremely large python codebase over from MyPy and it's an absolute game changer.

It's so fast in fact that we just added `ty check` to our pre-commit hooks where MyPy previously had runtimes of 150+ seconds _and_ a mess of bugs around their caching.

Re: Astral to Join OpenAI

#393

Post by makapuf »

darkwater wrote: post #308
islandfox100 wrote: post #244

Earlier quoted context omitted.

Maybe I'm reading wrong here, but what's the implication of the clean room re-implementations? Someone else is cloning with a changed license, but if I'm still on the GPL licensed tool, how am I "not protected"?

1. Company A develops Project One as GPLv3 2. BigCo bus Company A 3a. usually here BigCo should continue to develop Project One as GPLv3, or stop working on it and the community would fork and it and continue working on it as GPLv3 3b. BigCo does a "clean-room" reimplementation of Project One and releases it under proprietary licence. Community can still fork the older version and work on it, but BigCo can continue t…

2. BigCo owns ProjectOne now 3a. Bigco is now free to release version N+1 as closed source only. 3b. Community can still fork the older version and work on it, but BigCo can continue to develop and sell their original version.

Re: Astral to Join OpenAI

#394

Post by lukeschlather »

fnordpiglet wrote: post #191
Maxion wrote: post #125

Earlier quoted context omitted.

If LLMs turn out to be such a force multiplier, the way to fight it is to ensure that there are open source LLMs.

The problem is even if an OSS had the resources (massive data centers the size of NYC packed with top end custom GPU kits) to produce the weights, you need enormous VRAM laden farms of GPUs to do inference on a model like Opus 4.6. Unless the very math of frontier LLMs changes, don’t expect frontier OSS on par to be practical.

I feel like you're overstating the resources required by a couple orders of magnitude. You do need a GPU farm to do training, but probably only $100M, maybe $1B of GPUs. And yes, that's a lot of GPUs, but they will fit in a single datacenter, and even in dollar terms, there are many individual buildings in NYC that are cheaper.

Re: Astral to Join OpenAI

#395

Post by applfanboysbgon »

wiseowise wrote: post #381
applfanboysbgon wrote: post #287

Earlier quoted context omitted.

If the product did what it was advertised to do, they could simply build their own ecosystem for producing software and train the model to use it.

And then what? History is laden with technically superior software that lost to popular one. They can create uw tomorrow, but who will use it when everyone uses uv and its good enough for them?

The "then what" is that their model uses it. Technically superior software loses to popular software on marketing. But LLM owners have the ultimate marketing tool, because they can make their model use the tool. Anyone who asks how to do X in Python gets recommended "OpenAI-Python-Tool-For-X". Anyone who asks Codex to do X, Codex automatically installs "OpenAI-Tool-For-X". It would be very easy for them to launch even technically inferior software into a prime position. On top of that, if software developers are being replaced altogether as we are bashed in the head with such tales again and again, the marketing of dev tools wouldn't even matter, only what models are trained to use.

Re: Astral to Join OpenAI

#396

Post by fortuitous-frog »

JoshTriplett wrote: post #119

Welp. I used to respect Astral. I hope someone responsible forks their Python tooling and maintains it. Ideally a foundation rather than a company.

I don't think it's a coincidence that a lot of the (stale) tooling that Astral replaced were managed by foundations instead of companies...

Re: Astral to Join OpenAI

#397

Post by OutOfHere »

pronik wrote: post #258
smallpipe wrote: post #118

Earlier quoted context omitted.

If Codex’s core quality is anything to go by, it’s time to create a community fork of UV

Maybe they are being acquired to improve the quality of Codex.

That's the thing. To me that says that as soon as cash becomes tight at OpenAI, the Astral staff will no longer get to work on Python tooling anymore, namely uv, etc.

Re: Astral to Join OpenAI

#398

Post by dcreager »

skydhash wrote: post #305
everforward wrote: post #172

Earlier quoted context omitted.

I think this is more about `ruff` than `uv`. Linting is all about parsing the code into something machines can analyze, which to me feels like something that could potentially be useful for AI in a similar way to JetBrains writing their own language parsers to make "find and replace" work sanely and what not. I'm sort of wondering if they're going to try to make a coding LLM that operates on an AST rather than text,…

Writing a parser is not that much of work to buy a company in order to do it. Piggybacking on LSP servers and treesitter would be more efficient.

The parser is not the hard part. The hard part is doing something useful with the parse trees. They even chose "oh is that all?" and a picture of a piece of cake as the teaser image for my Strange Loop talk on this subject!

https://www.youtube.com/watch?v=l2R1PTGcwrE

Re: Astral to Join OpenAI

#399

Post by wiseowise »

tgtweak wrote: post #149

Amusing that the best python tools are written entirely in rust.

…amusing how? CPython is written in C, JVM is written in mix of cpp and Java, Rust was written using OCaml initially. Don’t know why you’re snickering. Do you also find it amusing that by the time cpp/rust team scaffolds and compiles initial boilerplate, python team is already making money?

Re: Astral to Join OpenAI

#400

Post by woodruffw »

nullhole wrote: post #389
woodruffw wrote: post #331

Earlier quoted context omitted.

As a point of information: Astral did not, in fact, burn through its VC money. I agree that dev tools are difficult to monetize, though. (Source: I'm an Astral employee.)

> As a point of order: Astral did not, in fact, burn through its VC money. That's a point of information, not a point of order.

You're right, I've edited it.
Post reply on HN