Bun got bought by Anthropic

[truongan07]

Q: Is Bun still open-source & MIT-licensed ?
A: Yes

The big questions : Will Bun be complete more features ?

1. Native HTTP2 support for Bun.serve() [#14672]
2. Make IPC fast [#8641]
3. Bun SQL native iterator/stream select [#25307][#8365]

This is going to be really fun. I hope the team still working on Bun.

[cjnoname]

Bun’s team has become more well-funded, but I’m not sure whether they can prioritise fixing issues. We want to adopt Bun, but we’re concerned about the frequent “segfault” “Bun crashes,” and similar issues, given the size of Bun’s user base.

[truongan07]

Bun’s performance is about half that of Node.js when processing large datasets on AWS Lambda.

Show benchmark right now please.

[cjnoname]

It feels strange to ask an end user to provide benchmarks. Isn’t this something Bun’s team should be responsible for? I’m not earning even a cent from this project, yet I’ve already run some tests for free and shared the results. It seems odd to be asked to perform additional testing.

My questions are:

1. Does Bun have official support for AWS Lambda?
   If so, have any benchmarks or tests been conducted specifically for serverless environments?

2. Bun has published many benchmarks, but I haven’t seen any that evaluate sorting or aggregating large in-memory lists. Is this type of workload missing from the benchmark suite?

3. Who is expected to perform this testing—the paid team, or the community?

[cjnoname]

@truongan07

It feels strange to ask an end user to provide benchmarks. Isn’t this something Bun’s team should be responsible for? I’m not earning even a cent from this project, yet I’ve already run some tests for free and shared the results. It seems odd to be asked to perform additional testing.

My questions are:

1. Does Bun have official support for AWS Lambda?
   If so, have any benchmarks or tests been conducted specifically for serverless environments?

2. Bun has published many benchmarks, but I haven’t seen any that evaluate sorting or aggregating large in-memory lists. Is this type of workload missing from the benchmark suite?

3. Who is expected to perform this testing—the paid team, or the community?

[truongan07]

@cjnoname , Hi friend

I am an end user and I’m not earning even a cent from this project. I have seen many benchmarks from devs who published it on their youtube and i haven't never seen "Bun’s performance is about half that of Node.js when processing large datasets on AWS Lambda."

You did talk about it so you need to show your benchmark to confirm. I don't mean to hurt you but I want to know what caused it to slow down by half for prevention again on me

[cjnoname]

Have you seen any benchmark testing for this? If not, then this part of the work is either missing or has never been a concern for the Bun team. As I am not involved in this project, I’m not going to spend my time testing it or providing test data unless I’m paid for that work. However, I am free to share my observations, and I’d be glad if someone could post something useful here to show that I’m wrong. Is that clear?

[kodonaut]

So now we can finally start measuring your five years before this project ends. It’s sad to see, Bun was really a great project. It’s ironic that a project born thanks to Zig gets destroyed by AI trash, the exact opposite of what Zig itself does. Well, it’s time to move on.

[truongan07]

Bun brought me to Zig. I think I should learn Zig. Zig vs C3 ?

[kodonaut]

Hi, that’s awesome that Bun brought you to Zig. I started with Zig years ago, and during my studies I decided I wanted to get to know C better :) so I spent a long time with C, and then went back to Zig. Have you already gone through Ziglings?