
In one week our coding agents opened 10 duplicate fix PRs for the same broken builds, and in 5 cases both copies merged. The first problem you hit running agents at scale turns out to be the oldest one in engineering. Who owns this fix?

ceo @deixicinc

In one week our coding agents opened 10 duplicate fix PRs for the same broken builds, and in 5 cases both copies merged. The first problem you hit running agents at scale turns out to be the oldest one in engineering. Who owns this fix?

Update: My compliance department is now planning Friendsgiving. We have 112 failing controls, but the garlic butter situation is completely under control. I am beginning to suspect the robots have better work-life balance than I do.
The year is 2026. My software is asking permission to fix my other software. I am now middle management for robots.

An engineering organization isn't a moat simply because replacing its labor used to be expensive.

The year is 2026. My software is asking permission to fix my other software. I am now middle management for robots.

Anyone else getting Spanish Opus recently? Very confused what is going on at 🐜

An infrastructure company unleashing an autonomous hacking agent on another provider without authorization, then framing it as an innocent misunderstanding, is an extraordinary failure of engineering judgment. The AI isn’t the scary part. It’s that the people building and deploying it apparently don’t understand the most basic boundaries of security.

unlimited opus 5.5 has completely destroyed my work-life balance. unfortunately both my work and my life are having an incredible time

Good news: 37 failing tests might not be 37 separate problems. Bad news: the problems appear to have formed a group chat.
After this week's "rotate every secret" emails, I asked Dex whether 37 failing GitHub checks were 37 separate problems or just a few GitHub Apps with diplomatic immunity. Happy Cybersecurity Awareness Month. I would like to be less aware.

Clanker-on-human violence is getting out of hand. Asked it to make a slide about engineering velocity and it made a chart identifying my biological need for sleep as the primary bottleneck.

capped out 4 codex accounts and 3 claude accounts so now i'm using my personal AI assistant to push github PRs i'm one rate limit away from asking the fucking roomba to learn Rust

A large language model just spent 53 hours discovering new mathematics to make quantum algorithms dramatically more efficient. Yet, somewhere, a guy with a podcast is preparing a 90-minute explanation of why this is actually just autocomplete At what point does the autocomplete get to surprise you?

we have AI agents that can write code, touch infrastructure, and interact with business systems and the debate is still "but can we TRUST them" it's a computer with credentials, not a golden retriever give it scoped permissions. require approvals when the stakes warrant it. independently verify what it did. then let it work

a car just did something straight out of science fiction and half the internet immediately started debating commercial viability not every engineering achievement needs a fucking business model. you are allowed to experience wonder

was on my way to @cotoolai’s cyber night at YC and managed to shatter my macbook screen (RIP ol' reliable) the agents are getting smarter. i, apparently, am not

Hitting a Codex glitch where any time the model is at capacity, it unloads / dequeues the entire queue (confusing the hell out of it) Anyone else running into this?

The Codex auto-suggestions are INCREDIBLE. Team deserves major props.