Hacker News Reader

Comment by vmsp | Hacker News Reader

I wonder how Deno's faring.

Node's the stable solution and will be with us forever. You can now use TypeScript with it and, soon enough, you'll be able to build your app to a single executable -- including native deps.

Bun's chaotic but, nonetheless, it's _fast_ and it's taking an interesting approach by including everything in the stdlib. Plus, bought by Anthropic.

Deno had an awesome story with the sandbox and ease of import for third-party dependencies. Sandboxes feel pretty commoditized now and I'm not sure the import mechanism ended up being that much nicer than a `npm add`.

View on HN

They did lay off ~half the company ~2 months ago: https://news.ycombinator.com/item?id=47467746

> Plus, bought by Anthropic.

Who thinks this is a positive?!

That wasn't a value judgment on the acquisition. I was just pointing out that it made the project more sustainable.

It really doesn't. You think Anthropic will still be in business in 10 years? If they are, it's not likely they'll be in the same shape.

That's a yes and no. Venture funded companies like Anthropic have a history of low follow through with peripheral projects (like Bun is for them). Of course they do - their responsibility is ultimately primarily to their investors - not to Bun. So the risk now is that Anthropic will can Bun whenever they just lose interest or feel it's just a drain that's not contributing directly to their bottom line.

Node.js itself did have trouble finding a corporate home that was interested in providing good support for the project, and that's how we got the oi.js fork of Node, which luckily led to Node being transitioned to a foundation and the projects merged. This whole history is what made me so surprised that Ryan of all people would attempt another js runtime (Deno) project as a corporate project.

And it's the reason I'm staying away from both Bun and Node. I can't afford platform risk like this. I need my startup to be built on a project that has a more reliable future trajectory, which is what you get with a proper open source project (emphasis on project) that you get with Node. Node is stable and still getting features, but most importantly it's not going away.

Bun is not peripheral, it's what Claude Code runs on.

I see how this might seem like it would make Bun important to Anthropic, but there are just so many examples of companies losing interest in the technology that their product "runs on". It's just infrastructure to them. To the C-suite, infrastructure is fungible.

Only if you think Anthropic is a sustainable business per se which remains to be seen.

It means they're a whole lot less likely to run out of money, which makes them a safer bet as a dependency.

> safer bet as a dependency.

The recent 1 million line vibe coded PR suggests it is not so reliable as a dependency.

That was Bun at Anthropic, not uv at OpenAI. (UPDATE: My mistake, this thread is about Bun, not uv.)

Is this a joke I'm not getting or are your wires crossed? Bun is the topic of this subthread.

Huh, you're right. I'd just been reading this other thread about uv and got my wires crossed: https://news.ycombinator.com/item?id=48228788

Running out of money is never the issue with a big company buying an open source project. There are countless examples of projects dying or changing significantly for the worse after acquisition.

Also “no human wrote any of this code” is not my personal benchmark for a reliable dependency.

> Running out of money is never the issue with a big company buying an open source project.

I'm going to dare to say that running out of money was often an issue before a big company bought an open source project.

running out of money, for an open source project of almost any kind, is safer than "running into money" with the wrong strings attached

(still reserving judgement on Bun, though — I mean, we'll soon see, one way or the other!)

Afaik there is no proof Anthropic is profitable. This, and uv buyout by OpenAI only adds a risk to supply chains. In few years these companies can be overrun by open source models or startups delivering new hardware/software breakthrough in LLM. It is not like uv and bun are acquired by IBMs or Alphabets of today.

Wasn't it announced that Anthropic is having their first profitable quarter right now in Q2? From what I've personally seen it's all driven by enterprise adoption.

Open source/foreign models are already way cheaper and will work just fine for most use cases but a lot of businesses are already pretty locked in to Claude, and with enterprise costing $240 a year at a 20 seat minimum it's a pretty big investment to make and won't be worth migrating unless the gains are significant.

Commentary upon Anthropic's profitability propoganda: https://www.wheresyoured.at/anthropics-profitability-swindle...

What happens if Anthropic and OpenAI shut down?

Is it different from the status quo prior?

No, the same thing always happens. The community latches onto a new project where all the effort is spent chasing previous gains, then they get bought out, and over time the same pattern appears.

At this point our industry would benefit from public investment in open source in the form of grants rather than relying on corporate benefactors to not rat fuck the commons.

For those who care about their dependencies being "safe bets", Bun should already be out of the question after the recent "vibe code the entire thing into a different language in a week with zero human intervention" fiasco.

Exactly.

https://github.com/yt-dlp/yt-dlp/issues/16766

> which makes them a safer bet as a dependency

Wouldn't node be the safest bet as a dependency?

> It means they're a whole lot less likely to run out of money, which makes them a safer bet as a dependency.

I don't think this logically follows. That is, yes being acquired makes one less likely to run out of money, but doesn't necessarily make something safer as a dependency.

Plenty of open source projects have little to no funding and continue on for years with no problems. But being acquired suddenly creates a requirement of return-on-investment. A corporation will happily shut the whole thing down if and when it's decided that they're just not gaining enough value from it.

(There's also the general fact that, a corporate-acquired project is going to first and forement serve the needs of the corporation vs. the community at large - if your use case or edge case doesn't align with the needs of Anthropic then you should probably not hold your breath waiting for the Bun project to address it.)

Yes, famously, acquisitions only make services more reliable, as in the case of Microsoft and Github, or Apple and DarkSky.

There are so many it's a meme: https://ourincrediblejourney.tumblr.com

It's good to have some options, to prevent the ecosystem from stalling

I agree philosophically, but the JavaScript ecosystem has never been languishing for lack of options. If anything, excessive fragmentation is a real concern.

Interesting, I'm not aware of that many nodejs alternatives...

I agree that frameworks are abundant but there are not that many JS engine alternatives.

It has been raining JS engines since Netscape Application Server, alone on Wikipedia there are 22 listed.

Okay that's great

You can already ship single executables, my product's CLI is a Node single executable application

> and, soon enough, you'll be able to build your app to a single executable -- including native deps.

Whoa, did not know that. That's a killer feature!

Code, as a single executable you say?! Whatever will they think of next?

Reply on news.ycombinator.com