kamal

The Kamal stack is K - you don't have A - product market fit yet M - why you fucking around A - with kubernetes L - just use heroku

@brynary @ryancarson @linear ah i thought this would be kosher if you used the claude code sdk

So it's up to the providers whether they allow seat subscriptions to be used with open source harnesses like Fabro... OpenAI has blessed it, and we support that. If `qlty install` detects Codex it will offer an OAuth-based path. With Anthropic, it's unclear what they allow and until they clarify there is a risk they consider it a TOS violation. Hopefully they clarify soon.

The company that actually builds the agent-first code factory is going to be worth hundreds of billions of dollars. No one has cracked it yet. It can't be the model labs because then you're tied to one model. I'm hoping a company like @linear will do this. I'd happily pay thousands of dollars a month for that (+ the token cost). Basically, we need SDLC 2.0 for the agent age. (Also, the right solution can't rely on gh - we need whoever does this to completely replace it as well.)

@brynary @ryancarson @linear i looked at the docs and it looks like you're using API keys. Any plans to support subscriptions to Claude/Codex?

@ryancarson @linear This is exactly what we are doing with Fabro (fabro.sh) except free and open source. Anything you need that’s not there, let me know and I will build it.

how codex desktop deals with branches and worktrees is confusing af

Can @graphite just work with worktrees natively?

@alexalbert__ i've found remote control to be extremely unreliable. posting from claude ios goes into the void

Remote control for Claude Code

It was easy to miss all the stuff we shipped at Anthropic this past week given... everything else that happened. A few of my favorites across Claude Code, claude dot ai, and Cowork:

@DevinAI Review’s ability to find concurrency bugs is unmatched

I would like a Devin review CLI so that it can be a skill. Really excellent results but copy pasting the findings each time is arduous.

Claude Code doing clear context and execute plan is clever. If they embed context in each task, and tasks survive /new, they can essentially run Ralph Wiggum natively with the same approach

@sbstnjmnzv Devin Review

What’s the coolest product you’ve seen this week? 👀

I’ve been using a Project in Claude as I refine a spec for a big-ish work project. I’m asking all sorts of random things related to the area, and the project memory auto update is great. The next step I’d like to do is to fire off Claude code. If only they can link to Projects.

@gaddafirusli maybe can remove this option in settings for the full year view since it has no effect

New on Bila Cuti 🗓️ — Added "full" year view to see everything together — Filtering by state and other settings works as usual Haven't link from the home screen yet, still figuring out how to do it nicely. But you can try it now: bilacuti.my/full

Hats off to the person who coined MQTT’s Last Will and Testament feature. That naming goes hard.

It doesn't look like MCP Apps (SEP-1865) supports toggling between different display modes. Most of the launch apps on ChatGPT Apps make use of the fullscreen display mode to provide richer interactions. Hope this changes.

@lg *ChatGPT Apps

received in slack today: "fyi: workshop for ChatGPT is moving to tomorrow since ChatGPT is down" -- yeahhh seems reasonable

@AmpCode new error today

@AmpCode gemini API is returning 400

@sqs Amphetamine

Not naming names, but literally the whole rest of the Amp team thinks Clicky is a terrible name. So I’m asking you, internet friends: is it? What’s a better name?

@AmpCode can you update the Librarian to work with local repos?

This is the form factor @sandbar should be aiming for

It’s infuriating how @OpenAI manages to mistreat its external developer community and face no real consequences. many such cases community.openai.com/t/apps-sdk-all…