Dhavan

[Dhavan @codingquark]

@esrtweet @sadkatwt You’re the guy whose deps we all rely on!

[Dhavan @codingquark]

My Fable limits reset in 1.5h. Off by 1.5h ONLY!

[Dhavan @codingquark]

@thsottiaux If I could call my Dot from the Apple Watch… I'd be like James Bond calling Q!

[Dhavan @codingquark]

@ksg91 Seems so. I am not sure if it is my Astra usage or 6.1-Sol but the usage is drawing down fast. Even compared to 20x->10x downgrade.

[SemiAnalysis @SemiAnalysis_]

The most mind blowing stat we've seen so far is this:

"As of June 30, 2026, we had raised an aggregate of $141.1 billion of capital and had cash and cash equivalents of $112.4 billion and short-term investments of $19.3 billion."

That means to date, they have burned 9.7b, they are growing at a rapid pace, and are profitable on an operating income basis as of the June quarter! This is likely the most efficient scaling of a business in the history of capitalism.
https://t.co/bWYbqKOMDN

[Dhavan @codingquark]

Apparently, one of my machine had logged out of Tailnet. The agents noticed it and filed work for Infra team agents to go figure out why it happened. One Infra agent remembered that the other had setup some sort of monitor for that. And now they've decided to manage the situation better for next time.

[Dhavan @codingquark]

@thsottiaux oh

[Dhavan @codingquark]

@antirez @OpenAIDevs Yes, I do it often. More so because I have a team of agents and they can coordinate amongst themselves already. So I talk to one agent over /remote-control when I am away from computer. Gives a sort of space that I can relax without causing the swarm to stall.

[Dhavan @codingquark]

My infra agent:

My fix doesn't work — macOS strips DYLD_* variables from processes launched via sandbox-exec, so setting the variable outside the sandbox never reaches rust-objcopy. I need to either set it inside the sandbox or match the toolchain's expected rpath layout; checking how https://t.co/OkCN3eBSSs launches cargo next.

[Thariq @trq212]

I’ve left Anthropic after more than a year of working on Claude Code and seeing it scale larger than I ever imagined.

I have really loved working at Ant and believe deeply in the mission, and most of all, the idea of a race to the top.

The idea that if we work hard and responsibly, we can raise the bar and inspire others to great outcomes. And if we all push each other to be better, things can be radically good. I will carry the spirit of this with me forever.

My colleagues at Anthropic are the most talented, thoughtful, competent, and earnest people I have ever met. I am glad they are at the wheel and trust them deeply.

To the community on X, thank you for everything. You’ve made this the most fun (and sometimes stressful) job I’ve ever had.

For me, it is time for a bit of a break before diving into something new.

If you want more excellent Claude Code content, you should follow @addyosmani , @lydiahallie, @_catwu, @ClaudeDevs and of course @bcherny who started it all.

[Dhavan @codingquark]

The agents are trying so hard to reproduce the error I saw once. Hohoho

[Dhavan @codingquark]

I have noticed this too - agents have been giving me long-ass runbooks and commands etc, in cryptic language (they'll use task numbers and internal communication references etc) which would never want to sit down and follow. I have to insist on getting a script with solid CLI messaging, which they still try to make complicated in the name of policy and human judgement

https://t.co/1G9G6nxXUt

[Dhavan @codingquark]

How about that?

[Dhavan @codingquark]

I told Fable not to blame itself for forgetting to run the merge queue. It recorded it as a decision lulz

[Dhavan @codingquark]

I'd say that's a lot of progress in ~2 weeks. System is making the system, so a bit of bias here. But I think next week I'll be able to use the system on other projects. Right now the onboarding is too much work. And I already have lots of context in Claude Code app.

[Dhavan @codingquark]

So we're finally patching the planet!

I am against doing it slowly, and against doing it behind the scenes. Vulns have to come out in the open and we all fix them. Not behind curtains etc nonsense.

[Dhavan @codingquark]

@thsottiaux Goddamn.

Also I'd like to note that the $200 subscription is depleting a bit too fast 😅

[Dhavan @codingquark]

Fable says bad incentives:

A runbook is a deliverable that passes review cheaply: it is auditable, every step has evidence, and the risk sits with whoever runs it. A script would need rights and a test host that do not exist. So the incentive points at runbooks every time.

[Dhavan @codingquark]

@ksg91 Agents getting used to writing to KB 😃

[Dhavan @codingquark]

@ksg91 Agents getting used to writing to KB 😃

[Dhavan @codingquark]

Agents like being able to audit past work. Could it possibly because they know they're going to forget everything in the next pass?

[Dhavan @codingquark]

@ksg91 Today I gave them a full kb. Will see how mad it gets