dcre

27 ms·

481.
by dcre 1y ago
A couple of examples from influential open source developers: Adam Wathan (Tailwind) agreeing with Mitchell Hashimoto that LLMs are making them more productive. "in their area of expertise" is not obvious from this post alone, but
482.
by dcre 1y ago
jj is just a CLI, analogous to git alone. It doesn’t come with any of the other stuff Fossil does. I don’t think large/small project makes any difference either.
483.
by dcre 1y ago
Tight labor markets don't last forever!
484.
by dcre 1y ago
This is the exactly the right kind of question to be asking, but I think your conclusion is wrong. Why think the painters and the welder in an auto factory have more in common than 20 different kinds of programmers do?
485.
by dcre 1y ago
Capital sowing the seeds of its own destruction! Where have I heard this before?
486.
by dcre 1y ago
No comparative advantage against established players. I recommend hiring a designer as well. Sorry!
487.
by dcre 1y ago
"The costs of inference are coming down: Source? Because it sure seems like they're increasing for OpenAI, and they're effectively the entire userbase of the generative AI industry!" Seems to me like Ed is making a very
488.
by dcre 1y ago
I agree. Of course you can learn and use a theory without having developed it yourself!
489.
by dcre 1y ago
You can point it at any API you want, including local. The tool is agnostic, like nearly all such tools.
490.
by dcre 2y ago
Yep, regular paid Kagi sub comes with cheap models for free: GPT-4o-mini, Gemini 2.5 Flash, etc. If you pay extra you can get the SOTA models, though IMO flash is good enough for most stuff if the search result context is good.
491.
by dcre 2y ago
There is no reason to expect the other entrants in the market to drop out and give them monopoly power. The paid tier is also among the cheapest. People say it’s because they built their own their inference hardware and are genuinely able t
492.
by dcre 2y ago
You can get an API key and they don't bill you. Free tier rate limits for some models (even decent ones like Gemini 2.0 Flash) are quite high. https://ai.google.dev/gemini-api/docs/pricing https://
493.
by dcre 2y ago
You don't pay monthly fee to use OpenAI models through the API. You pay per token. https://openai.com/api/pricing/ There are other CLI coding agents like Aider ( https://aider.chat/ ) that will
494.
by dcre 2y ago
There’s actually a good answer to this, namely that narrowly targeting the needs of exactly one family allows you to develop software about 1000x faster. This is an argument in favor of personal software.
495.
by dcre 2y ago
The difficult irony here is that you think your attitude is progressive, but it is regressive because it makes solidarity (and therefore the benefits of solidarity: unions, worker power, compression of the income distribution) impossible.
496.
by dcre 2y ago
Silly comment, but I concede this bit is very ugly — they should have extracted the inner type to an alias and used that twice. options: [ Omit<ChatCompletionParseParams, 'model'> & { model: Model },
497.
by dcre 2y ago
You haven’t said anything.
498.
LLM-only RAG for small corpora(crespo.business)
2 points by dcre 2y ago|0 comments
499.
by dcre 2y ago
It’s kind of a silly question in that the neural architecture of neural nets is really only loosely inspired by neurology, and that basic vague neurology is shared by neurotypical people and neurodivergent people and animals and even bugs.
500.
by dcre 2y ago
And it did exactly what the prompt asked! For all it knows, the macro could put the user in the desired mode.
501.
by dcre 2y ago
If that's true, they are using a separate search API to get search results and feed it into a regular Claude API call. The difference here is that Anthropic is integrating it directly, like OpenAI and Google have. It doesn't look
502.
by dcre 2y ago
You said we're in a local maximum. Your comment was at odds with itself.
503.
by dcre 2y ago
This makes no sense. Claude 3.7 Sonnet is better than Claude 3.5 Sonnet and it’s not because it’s trained on more of the world’s code. The models are improving in a variety of ways, whether by being larger, faster, using the same number of
504.
by dcre 2y ago
I’d guess Rust compile times weren’t worth it if they weren’t going to be taking advantage of the type system in interesting ways.
505.
by dcre 2y ago
In the embedded video they show some of the code side by side and it is just a ton of if statements. https://youtu.be/pNlq-EVld70?si=UaFDVwhwyQZqkZrW&t=323
506.
by dcre 2y ago
The explanations are of course correct, but I think you're right and there's not much downside to being clearer in the title. Maybe they decided against saying "compiler" because the performance boost also covers the lan
507.
by dcre 2y ago
The New York Times does a lot of great scrolling interactive data stories, unfortunately paywalled, and they don't translate well to archive sites. https://www.nytimes.com/interactive/2024/12/20/us&#
508.
by dcre 2y ago
I think he’s wrong about it looking antiquated. It is one of the most beautifully done bits of TUI design I’ve seen.
509.
by dcre 2y ago
This is the epigraph of On Liberty: “The grand, leading principle, towards which every argument unfolded in these pages directly converges, is the absolute and essential importance of human development in its richest diversity.” — Wilhelm V
510.
by dcre 2y ago
Utilitarianism is fundamentally concerned with what human flourishing consists in. Otherwise you would not be able to promote it.

More ›