“If you’re thinking without writing, you only think you’re thinking.” ― Leslie Lamport This quote lives rent-free in my head these days, now that delegating our writing to AI is more and more common.

He took a non optimized and inefficient Rust code, put an AI agent to the task of rewriting it in assembly from scratch with an optimized approach, and now claims that moving to assembly alone made a 17x performance improvement.

I'm having a hard time understanding the appeal of paying for courses to learn how to code with AI. What you may learn today may be obsolete when the new frontier models arrive.

More than six months later, I feel this even more acutely—and I’m more optimistic than ever about our profession.

Now with AI making it easier to do things, it is more important than ever to know what not to do.

I’m not sure I agree. The solution should be testing larger units, and mocking as little as possible. But throwing away unit tests entirely does not sound right to me.

Now that everyone's looking down on em dashes because they're a sign of AI writing, I'm moving towards them instead—I'm now relearning to use them properly and intentionally in my own writing.

"I expect to need to keep evolving what I do and how I do it on a daily basis, just like I've done for 30 years at this point." 💯

How do we even call this new class of models? It’s clearly not an LLM in the traditional sense.

Can’t wait for Jev-like open source models to come out. Or even running them on device.

Claude vs Astra. It's crazy how unpleasant Claude 5 has become to use.

Nailed it!

Another weird Claude/Fable habit: giving human-coding-era estimates to everything I ask it to: "this will take an evening", "two days", "a week". No buddy, you're going to do it right now, and it'll be ready in 10 minutes tops, or you're fired.

What a load of crap!

Finally realized why it’s so exhausting and stressful to work with AI agents 8h a day. When delegating a task, you want to be assured it’s off your plate and you can forget about it, knowing it will be done completely and correctly. AI doesn’t give you that: you need to check it’s doing the right thing, it’s not forgetting anything, not taking short cuts etc. So it adds to your mental load, rather than removing from it.

ChatGPT Desktop (the Codex app) can now be configured to use Ollama models. Download or update to Ollama 0.34 to get started.

here's the argument some ai lab employees are making - they genuinely believe what they're working on might kill everyone - they continue to work on it because if this thing gets built it should be by responsible people like them - the reason they do money things like selling enterprise deals and IPOing is solely to fund this effort if this is all true then why is employee equity a thing? if you want to tell this story of being terrified of what you're building it doesn't work when there's a clear profit motive it's particularly weird when people vest and then leave and write a tirade about how these companies are bringing the end of everything if you want to act like this is the manhattan project, those scientists didn't have equity in something that gives them generational wealth you don't get to say something bad is happening while profiting off it and also being seen as the good guy. one of these has to go

It's been only 2 full days using it, but I think I'm sticking with @heliumbrowser for the foreseeable future. I'm really liking it.