Anthropic spun up 16 Opus 4.6 agents to build a Rust-based C compiler from scratch - and the compiler they produced was able to compile the Linux kernel 🤯
All 16 agents worked in parallel on a shared codebase without active human intervention. No compiler engineers were involved, and the entire effort cost just $20,000 in API usage. That's it!
The "agent team" produced a 100,000-line compiler capable of building Linux 6.9. Absolutely insane.
We are living in incredibly exciting times.
– We don't write tests.
– Why?
– Because we don’t have time for it.
– Why?
– Because there is too much work and pressure.
– Why?
– Because we don’t move fast enough.
– Why?
– Because changing software has become difficult and risky.
– Why?
– Because we don’t write tests.
@Bhuvneshdot@ankushdharkar@arpit_bhayani Can confirm — I’ve met him, and he’s as humble as he seems. Inspiring on every level. Despite everything he’s building and balancing, he stays grounded, curious, and super generous with his time and knowledge.
Sorry for the rant, but I had to share this.
@arpit_bhayani already a well-known name in the tech world — doesn’t need much of an introduction.
But over the last few months, I’ve been following his work more closely, and honestly, I’m blown away by everything he’s doing — whether solo or with a team.
- Runs his own website
- Publishes a newsletter regularly
- Maintains the open-source project DiceDB
- Creates deep-dive videos on YouTube
- Active on both X and LinkedIn
- Hosts and attends podcasts and events
- Works at one of the world’s top tech company
- Seems incredibly humble on camera (those who’ve met him — confirm this?)
- Spends time with his lovely daughter & family
- Possibly handles Bangalore traffic on top of all this 😄
- And just yesterday, I came across another product — Revine — run by him.
Not sure what else I’ve missed... but yep, he also has just 24 hours in a day!
Saying "this isn't my problem!" is a problem.
If you want to accelerate your career and make sure you are having a lasting impact at work,
1. do not build a habit of being spoon-fed
2. do not keep asking for help instead of checking code or docs
3. do not keep spending time on easy and low-impact work, fight for high-impact projects.
You must show a high bias for action at work, take ownership, and make an impact. Here are some actionable that you can apply right away at work to earn some brownie points
4. waiting for a code review? ping them
5. seeing high error rates? join the thread and help if required
6. outdated docs? update them instead of complaining
7. teammate stuck? help before they ask
8. slow processes? propose a solution and drive adoption
9. no owner for a project? step up and lead
10. recurring issue? proactively automate/document the fix
I think everyone should operate with a high bias for action. It is not just about identifying problems but actively working on solutions.
Instead of waiting for someone to act, you act and lead.
here is o1, a series of our most capable and aligned models yet:
https://t.co/yzZGNN8HvD
o1 is still flawed, still limited, and it still seems more impressive on first use than it does after you spend more time with it.
In order to ship things faster, devs start writing bad code.
Bad code might seem fast, but it's not.
At first, fixing it seems cheap.
But as time goes on, it gets slower and harder to work with.
Start putting effort into delivering good code.