Viewing profile — hombre_fatal

Exactly. Though I was seeing 130 MB PSS when I did the same experiment last year. Not sure if it had something to do with my env (NixOS), but either way it was a turn off since I s…

That's better than the GTK4 hello world I tried out last year.

Long ago I added an instruction to my global AGENTS.md that every time the agent asks me something, make it multiple choice that ranks the top options, give a quick trade-off summa…

Agreed. The experiment is still in its infancy but the direction is great. For example, I want to be asked about general shapes/categories of commands as they first appear for a pr…

We have this idea of AI software development where a human watches over its shoulder and shouts out things like "use uuidv7 for the id since timing is important" or "use a sum type…

As opposed to the norm in computing where the average user is expected to just trust rando software, the AI auto-approver that classifies actions the agents wants to take is a huge…

Because we decide on the quality of community, discussion, and engagement that we want to have, not the government. HN has a lot of moderation just to keep this place a pleasant pl…

One of the weirdest parts of X is seeing Nikita go ham on some random account. Nikita shows up in the replies of a guy with 12k followers complaining that he doesn't get his small …

No, you unlock the ability. I think it's just account age + submitting ratings on existing community notes.

I think for those of us who got into software initially for the end result -- being the only way for the thing in your head to ever exist -- enjoying the engineering part was a nic…

We're pretty far past this if you're using anything close to the sota models. But you could be defensive with a security checklist in agents.md and have adversarial review, if you …

The cool thing is that you can use and adapt the tests of all major third party terminal implementations. Come up with your own invariants (brainstorm them with a smart LLM), and n…

Maintaining a fork seems trivial for an LLM. Replaying each upstream commit against your fork is exactly what even dumb models can do without issue. But at the same time LLMs also …

That’s becoming as illuminating as saying you produced that comment word by word.

I limited to 1-2 servings of coffee by switching to an aeropress and a hand cranked grinder. When I brew one big carafe, then I have this compulsion to drink it, like every new tas…

The examples kind torpedo the whole post. > The reviews are irrationally nitpicky, choosing one quibble and making the entire review about that. Somehow the apps deserve an overall…