Ibrahim retweeted
It’s professional courtesy in the industry to not call out your competitors in public when they have an outage fwiw because you will have one as well
In private it’s different but in public always been a big no no
These ones learn now the hard way why
Ibrahim retweeted
One interesting data point from @rootlyhq is how devs are getting paged ~2x as much outside working hours vs a year ago.
The reason?
Most likely ~2x PR frequency, thanks to a lot more PRs and code pushed that are AI-generated / written
@GergelyOrosz wrote about "AI SRE"s, a problem-space i have been working in for the past 1 year. In usual pragmatic eng style, it is well written with a lot of depth.
- This is a use-case of AI helping Engineers instead of just replacing them 😉
substack.com/inbox/post/2223…
Ibrahim retweeted
the codebase is the memory
if you allow agents too much discretion with memory, slop accretes
You must control how memories are created in an agentic system.
Ibrahim retweeted
Gdmi whenever I want to try a Gemini model it’s not publicly available. Don’t bother announcing your scores if we can’t use it.
Ibrahim retweeted
In my toy apps, vibe code works.
But in client code with millions of lines, there is no patch that is right on the get go. I steer by reading the code.
Those not reading... if you were in this project, you'd be shipping shit. Incorrect code and incorrect tests.
Part of what makes Opus writing so grating to me is how it anthropomorphizes inanimate objects (e.g. code/functions/etc). Wonder if it stems from Claude's "constitution" where they tend to anthropomorphize it. It makes me feel like mustafa-suleyman.ai/a-warnin… is spot-on.
When I ask Codex to review Claude or Claude to review Codex on @Superconductor, they tend to be more critical to implementations the other model did than implementations from the same provider. Makes sense as they probably weight different considerations differently
Ibrahim retweeted
You can now launch Superconductor agents directly from @linear!
Assign any Linear issue to Superconductor, run coding agents in the cloud, see a live preview once the agents are done.
superconductor.com/blog/line…
Ibrahim retweeted
ban and delete most if not all code comments!
agents love copying what already exists. a seemingly harmless comment explaining some temporary workaround is one of the worst viruses you can have in your codebase. like a weed it'll spread quickly as a tool for agents to quickly get their job done. before you know it, your codebase is overrun with hacks and stale comments. just one more workaround bro
you can use pstack's /no-comments skill as a weedkiller. it'll seek out useless comments while keeping the ones that actually should exist (like explaining a browser quirk), and figure out how to solve them on a first principles basis
here's a prompt you can try asking @bot:
create a project agent that uses /poteto-mode to spawn a /swarm to use /no-comments and identify useless code comments we can prune in my repo. come back with a high level plan of which of these workarounds would be best solved by refactoring and which we can straight up delete
x.ai/bot/plugin/9717366
Ibrahim retweeted
It's remarkable how all of the ai products converge over time at each step
1. autocomplete in ides / chat interface
2. use bash for everythin with claude code <> codex <> <> open code <> grok build
3. more capable desktop apps
4. remote sandboxes
5. ai coworker with claude tag <> grok bot <> dots
Announcing dots. Dots work 24/7 for you, learn from your feedback, have their own computer, browser and can be connected to over 4k apps in our ecosystem.
They’re powered by Astra our best model yet. Included in your Pro plan, without drawing down on any of your usage. You can even call a dot while it’s working.
We’re getting you started with your primary dot today and soon you’ll be able to create entire teams of them.
that feeling when you can't use your best buddy @Superconductor to talk through the maintenance operation for Superconductor.com 😭
