Oh yeah I want to cache too. But caching implies you know somebody else owns the data. I'm also not very clear on how idempotent load is. Resume is a foot gun because if the agent has another interface to it, like all CLIs do, how do you even figure out if you need to re-load versus resume? There's an updated date somewhere, I believe. Highly skeptical that anyone manages it correctly. I love the idea of ACP but it makes me very uncomfortable about how difficult will it be to actually support it. I already tried ACP once about three months ago and it was pretty difficult but that was three months ago, which is a lifetime.
If you make a protocol and everything's optional, it's so annoying. I'm looking at you ACP. I find the semantics around session load and session resume to be annoying. The frustrating thing is it's not extremely clear who owns the data. I want the agent to own the data not the client. But if session/load doesn't exist that's basically impossible.
@ibuildthecloud I should clarify, the client doesn't own it. But many cache it themselves and wanted a resume that was faster without reloading the whole thread, which I felt was fair
@ibuildthecloud Indeed the agent does own the data. But lots of clients like owning it too, which is why resume exists
Expect v2 to require all of these session methods btw. We just needed to evolve and take the ecosystem along with us in a future compatible way.
@rainbow_0219 Hey Google just shipped several updates in v0.20 that made it so it worked in Zed again. But it's still not working for you? Can you explain more?
Lately I've been disappointed with Gemini CLI's ACP implementation. Slash commands and model selection still don't work, and now it's completely broken in Zed for me.
Integrating the agent you want into the IDE you love should be effortless. That’s why we've partnered with @zeddotdev to create the Agent Client Protocol (ACP). The first implementation is included in the 2025.3 JetBrains IDEs release. Read more in our blog: blog.jetbrains.com/ai/2025/12/bri…
Folks at @zeddotdev: the CLAUDECODE env var intentionally left unset? Or is there another env var Nx can use to detect that it's running in an agent in Zed?
@dsp_@nathansobo@as__cii I think it could. My main motivation is just to ensure the transport and auth stay consistent. It's too annoying to have to reinvent them.
@dsp_ do you have a connection with somebody at Zed or somebody who's authoritative on ACP. Or maybe you would just be willing to talk to me about it. It's so darn close to MCP yet unfortunately incompatible. Anything I could do to help bring these two things together?
@0xubermenshh@zeddotdev@jetbrains@MyDeathMachine is right. We use Rust to generate the schema, but the Rust SDK has a bunch of other stuff on top to help you get started with your client/agent implementation
@zeddotdev@jetbrains hi, i had some questions. the main acp repo is it just a specification? and what’s the use of the sdks we also have for rust? why do we have one for rust when the original repo is in rust?
ACP now has its own repo: github.com/agentclientpro… ✨
This makes it easier for editors and agents to collaborate on the protocol itself.
We're excited to evolve ACP in partnership with @jetbrains. They've already contributed a Kotlin SDK to get things started.
Zed has raised a $32M Series B, led by @Sequoia.
We’re full speed ahead making a great editor, but there’s a bigger picture: we're building DeltaDB, a new kind of database for collaborative coding.
Early days, but the vision is clear: zed.dev/blog/sequoia-b…
Ben Brandt joins Rust in Paris 2025 !
Excited to announce that Ben Brandt, Staff Engineer at Aleph Alpha, will be speaking at Rust in Paris, on March 14th, 2025 in Paris.
📣 Don’t miss the chance to learn from Ben’s expertise
🎟️ Get your ticket now
👉 ti.to/xperhub/rust-i…
On a whim I looked something up and...
github.com/Azure/iotedge/…
microsoft uses cargo-clippy
I CREATED A RUST LINTER NAMED AFTER CLIPPY ON A WHIM AND IT ACCIDENTALLY BECAME POPULAR AND NOW MICROSOFT USES IT
welp.