jbwinters

6 ms·

1.
by jbwinters 2mo ago
Ah, that’s a tough one. My hope is that Jacquard, or some similar language, can make intent explicit through readable, high-level constructs, giving both the human reviewer and the model a clearer target while the details are filled in. But
2.
by jbwinters 2mo ago
That’s a fair concern. I’d expect multi-shot handlers to be relatively rare and mostly something library code uses rather than a day-to-day utility. But an LLM given free rein may decide to behave differently. For now, I am looking into mak
3.
by jbwinters 2mo ago
This is definitely true. The language is intentionally small enough to fit into a single SKILL.md file, for what it's worth: https://github.com/jbwinters/jacquard-lang/blob/main/docs/SK... Agen
4.
by jbwinters 2mo ago
You're right, an LLM doesn't have preferences. As shorthand, I thought it a useful concept though, as they do have particular ways of writing that can be tuned for. The effect system here is part of a type system (technically a ty
5.
by jbwinters 2mo ago
It's a good name regardless! Don't stop on my account.
6.
by jbwinters 2mo ago
Thanks for the pointer to Jai! I will check it out. Yes, Jacquard uses content-addressed definitions and it should be possible to set up a process for 'review again if this changes' on top of it. Warp, the testing framework, alrea
7.
Show HN: Jacquard, a programming language for AI-written, human-reviewed code(github.com)
102 points by jbwinters 2mo ago|59 comments
8.
by jbwinters 9mo ago
I dug into the Stargate data center Michigan just approved. 1.4 gigawatts of load, roughly 11% of the state's entire retail electricity consumption. It's been controversial locally but most of the debate has been about farmland an
9.
Compute Is the New Car(subvocalizing.substack.com)
1 points by jbwinters 9mo ago|1 comments