Have been in a superposition since Nov talking to people about the future of Electrobun.
This included:
- over 140 meetings
- 2 companies I really like with acquihire interest but ultimately not finding full alignment after months of conversations each.
- almost moving back to SF
- Talking to investor and founder friends about raising
- a 3 month embedded research arrangement helping shape a unicorn’s internal AI use while also studying them for my own product research.
In that time I shipped:
- 18,000+ words on “How to be an agent first” direction and culture docs for the unicorn
- Electrobun v1 (stabilized architecture and cross platform)
- Electrobun passed 12k github stars and growing with hundreds of apps already shipped
- Electrobun WGPU (games, shaders, and ML workloads)
- co(lab) completely rebuilt as Dash
- hundreds of people signed up to try Dash at https://t.co/JdbrFCeT5T
What’s next:
- Ship Dash to alpha users
- Stabilize Electrobun’s new zig core and wrappers so you can build even tinier even faster desktop apps with zig, rust, go, bun, deno, or cottontail
- Build cottontail, a new advanced: tiny, fast, and batteries-included JS runtime
- Double down on the Toronto startup ecosystem as it enters a new era
Electrobun has a strong future, I can keep going for at least the next 20 years, and I’ve never been more excited, focused, and clear about what I’m building.
Back to shipping 🚀
Inspiring interview.
LLVM is to Zig, what Bun is to Electrobun.
Looking forward to stabilizing the new zig core and decoupling (or if anthropic won’t do human code reviews then deprecating) Bun.
Andrew Kelley’s stewardship of zig is one of the best things software right now.
In Toronto for Tech Week and open to a small number of investor/operator conversations.
Electrobun just crossed 12k GitHub stars. It started as a better way to build desktop apps with Bun + native webviews, and it’s becoming the foundation for a bigger system: Dash, an AI-native workspace where humans and agents can safely share context, tools, and workflows across machines.
I’ve been bootstrapping, have already passed on one strategic path, and am now deciding whether to keep going solo, raise, or partner more strategically.
If you understand devtools, AI-native workspaces, local-first infra, or the future of software engineering, DM me.
I have written another 8,000 word document on AI culture.
It’s under NDA and for the unicorn I’ve spent the last 3 months researching to use internally.
Do you want them to publish it?
In other news Electrobun just passed 12k stars
Looking forward to 2.0 that will add the same modular capabilities it has for views (system webview, pinned CEF, or raw GPU) to the main process (Bun, Deno, Node, Rust, Zig, Go), and a new runtime cottontail
https://t.co/WugFPKi3qB
Electrobun 2.0 will also be decoupled from Bun due to the rust rewrite.
It’s a combination of anthropic’s stance of not doing human reviews or any kind of rational roll out and stabilization.
Rust is great though, Electrobun 2.0 will have first class support for rust, zig, go, and more.
Credibly there is a gap in the JS runtime space for something that is both tiny, secure, and performant.
What Bun (even before it became a marketing gimmick) failed to deliver on.
I will build cottontail. It will be better than what has come before.
In a different tweet you mentioned the Overton window.
If you have access to advanced models that don’t hallucinate as often as the ones we all have access to it would have cost your team nothing to prove that by reading the code and letting it marinate for longer until people catch up.
But model cards and experience say it’s unlikely.
Choosing to roll it out this way means people can’t even pin to the last zig version while it marinates because any zero day discovered would force any company to choose between getting a security patch at the expense of undefined behaviour.
Any rational person has to stop using Bun in production now.
It doesn’t have to be this way. It’s not too late to take a more rational path.
The concern isn’t that it will break things, the unit tests pass.
The concern is how many hallucinations or misaligned additional code was added.
What was previously believed to be a battle tested thing that’s had lots of eyes on it running on my server handling my customer’s data.
Is now a black box that presumably no one has really looked at, that’s going to maintain a velocity and practice of no one looking at.
What’s in the blackbox handling my customer’s data?
How can Bun continue to be core infrastructure going forward?
I wish it was at least kept as canary for long enough for the community to plausibly look at what was generated. Instead of the next version of bun being this new unseen thing.
If Claude was so good, they could have had a rust port as a separate repo, and had claude port every PR.
Or just have it fix the segfaults it’s been introducing the last few months since acquisition.
There is no reason to YOLO +1M rust, -700k zig as a minor version bump in 6 days and announce no human will ever look at the code
That just kills bun for any serious person.
Sorry I just realized there was a small typo:
- Electrobun is currently C++, ObjC, TS, Zig
- I’m moving the Bun TS into a new Zig core
- The Bun TS is shrinking to be a thin ergonomic TS SDK
- SDKs for Zig, Rust, and Go so you can bundleBun:false and use the language you like for your main process
After Electrobun's packaging magic the Zig kitchen sink (if I disable CEF, WGPU, Bun) is 1.9MB down from 19.8MB when bundleBun:true.
That's 10x smaller.
And since Bun (really JSCore) bundles ICU data on Win/Linux that sub 2MB will be more consistent cross OS.
Alright got a lot working with zig as a main process. Here are the two most complex integration tests showing
- Webview with OOPIF compositing
- Webview with native WGPU layers rendered from zig
- CEF
No BunJS.
It's on main, will pick this up again https://t.co/xTXKolIVyz