dcre

19 ms·

451.
by dcre 1y ago
Both, but it’s mostly the models. The programs like Claude Code are actually simpler than the ones from before because of this.
452.
by dcre 1y ago
This is a great example of how no example provided is ever good enough. There’s always an argument that it doesn’t really count. Yet you just said the computer is doing what you did as a junior developer.
453.
by dcre 1y ago
The Eno quote is good, but the post adds nothing to it! If anything, I think the sections after the quote make the post worse. This shows poor judgment on the part of the author.
454.
by dcre 1y ago
"By matching the locations of Superfund sites and their pollutant accumulation periods with the birthplaces and birth years of these American-born CEOs, we identify 734 unique Superfund CEOs and 2,267 non-Superfund CEOs" Certainly
455.
by dcre 1y ago
That’s interesting, though it doesn’t undermine the point that it’s information-dense, especially if you are coming at it from outside the field.
456.
by dcre 1y ago
These interviews are very, very good. Hearing these people talk casually in their own idiom about their work is the most efficient way I know of to get a feel for how things work at the cutting edge inside the big labs. That is useful wheth
457.
by dcre 1y ago
Beautiful work. I tried other jj TUIs and found them clunky and counterintuitive. This one feels like they took the output of jj log and simply made it interactive. That said, I switched to jj after using GitUp https://gitup.co&#
458.
by dcre 1y ago
I bet it’s stronger than that, and they anchor a lot of the alignment training to the unique (ish) token of Claude.
459.
Highlights from the Claude 4 system prompt(simonwillison.net)
9 points by dcre 1y ago|1 comments
460.
by dcre 1y ago
Typo in headline (currently): Micro Execises
461.
by dcre 1y ago
This looks fine to use, but considering you can a achieve the same with a git repo and a 20 line bash script to make the symlinks[0], it seems to me precisely the kind of thing beginners should learn to do themselves rather than trust the m
462.
by dcre 1y ago
Not your fault. They are the worst.
463.
by dcre 1y ago
It is a thread. You may have only seen the first tweet because Twitter is a user-hostile trash fire. “The lesson: on a fundamental level, solutions to these games are low-dimensional. No matter how hard you hit them with from-scratch traini
464.
by dcre 1y ago
But many "agentic" tools are model-agnostic. The question is about what the tool itself is doing.
465.
by dcre 1y ago
You can use Aider with Gemini. All you need is an API key. https://aider.chat/docs/leaderboards/
466.
by dcre 1y ago
It's very surprising that it has taken this long to see a first-party CLI like this.
467.
by dcre 1y ago
I took a version of this away from Sandi Metz’s 99 Bottles of OOP. It’s not really my style overall, but the point about moving logic forks up the call stack was very well taken when I was working on a codebase where we had added a ton of f
468.
by dcre 1y ago
In this case we are not talking about beginner devs (to their credit, the React docs are pushing people toward frameworks now) — these are literally the developers of a framework, fucking it up in their own docs!
469.
by dcre 1y ago
If you render an <a> in React (or Angular, which I think they're using here), it's just an <a>! You have to do extra work to fuck it up!
470.
by dcre 1y ago
This is not really malicious compliance because it is not aimed at the boss who ordered the policy. It’s more like chaotic neutral compliance.
471.
by dcre 1y ago
As others have said, you are describing a totally imaginary world where money is the only cost. “Artificial” is doing all the work. But the very investments you’re describing are “artificial”, and more than that, they require constant spend
472.
by dcre 1y ago
I’m impressed. This is one of the strangest opinions I’ve ever seen. What is special about “at use”? Presumably because it lets you avoid the question of whether everyone should get a free car. Does a monthly car payment count as “at use”?
473.
by dcre 1y ago
I believe they use the LLMs to generate a set of things to search for and then run those through existing search engines, which are totally opaque and use whatever array of techniques SOTA search engines use. They are almost certainly not &
474.
by dcre 1y ago
As I say in my comment, they may even care about it but not have time to churn while best practices are figured out.
475.
by dcre 1y ago
It's not about it being locked to a commercial product — whatever they built would be provider-agnostic. My understanding is the decision is more about not wanting to build things into core that are evolving so quickly and not wanting
476.
by dcre 1y ago
Agreeing with the other replies: Helix. https://helix-editor.com/ Here's a nice recent post about it: https://felix-knorr.net/posts/2025-03-16-helix-review.html
477.
by dcre 1y ago
Seconded, everyone should be using CLIs. Simon: while that example is impressive, it is also complicated and hard to read in an HN comment.
478.
by dcre 1y ago
No need to write your own whole thing (though it is a good exercise) — the existing tools all support ways of customizing the prompting with preferences and conventions, whether globally or per-project.
479.
by dcre 1y ago
The Aider leaderboards are quite helpful, performance there seems to match people's subjective experience pretty well. https://aider.chat/docs/leaderboards/ Regarding choosing a tool, they're pretty ligh
480.
by dcre 1y ago
That's a good point, though it's worth thinking about what would count as a demonstration. Pointing to a PR that an LLM helped with is not enough because you don't know how long it would have taken without the use of LLMs. Th

More ›