raincole

22 ms·

601.
by raincole 6mo ago
This is exactly what creativity is.
602.
by raincole 6mo ago
First of all this is clearly AI-assist writing (being charitable here). And the premise makes no sense anyway. The only risk of forgetting how to make shells is when other countries are making shells more efficiently. Non-western countries
603.
by raincole 6mo ago
They are, but Google Vertex has been one of the official ways to use Claude since forever.
604.
by raincole 6mo ago
I mean "understanding it just like when you hand wrote the code in 2019." Obviously I don't mean "understanding it so you can draw the exact memory layout on the white board from memory."
605.
by raincole 6mo ago
> the day when LLM-assisted coding is commoditized Like yesterday? LLM-assisted coding is $100/mo. It looks very commoditized when most houses in developed world pay more for electricity than that. My definition of LLM-assisted codi
606.
by raincole 6mo ago
You're probably hallucinating. The parent comment is quite clear.
607.
by raincole 6mo ago
The article is about payment providers. Do you think payment providers should act like moral police that decide how the customers can spend their money? If so, do you think Google/Apple/Microsoft should have a say in which apps th
608.
by raincole 6mo ago
Except the actual title here is clearer. Your suggestion is so anti-AI-clickbait that it overflew and became a bad title again. If Tesla (insert any car manufacturer you hate) ran over a kid I'd like to see the title say it, instead of
609.
by raincole 6mo ago
[flagged]
610.
by raincole 6mo ago
Insert always has been meme. But seriously, it just stems from the fact some people want AI to go away. If you set your conclusion first, you can very easily derive any premise. AI must go away -> AI must be a bad business -> AI must
611.
by raincole 6mo ago
Could you elaborate on how to interpret your comment without it leading to anti-intellectualism?
612.
by raincole 6mo ago
OpenCode? Pi? If one finds it difficult to set up OpenCode to use whatever providers they want, I won't call them 'dev'. The only real friction ( if the model is actually as good as SOTA) is to convince your employer to pay
613.
by raincole 6mo ago
History doesn't always repeat itself. But if it does, then in the following week we'll see DeepSeek4 floods every AI-related online space. Thousands of posts swearing how it's better than the latest models OpenAI/Anthrop
614.
by raincole 6mo ago
It's such a vague table for pricing information. 30-150 messages...? What?
615.
by raincole 6mo ago
Who is 'you' here? All of the npm package maintainers? Yes, if they all just backport security patches we'll be fine. No, people are not going to just .
616.
by raincole 6mo ago
It's very weird that they frame caching as "latency reduction" when it comes to a cloud service. I mean, yes, technically it reduces latency, but more importantly it reduces cost. Sometimes it's more than 80% of the tota
617.
by raincole 6mo ago
Most programmers who write reasonably deterministic code don't even know how many bits it allocates.
618.
by raincole 6mo ago
I think you're talking about a different topic unrelated to the linked article. In the linked article the LLM doesn't split it into several commits. If LLM had a button to split the bug fix and the overall refactoring, the author
619.
by raincole 6mo ago
And he's clearly not talking about scientific journeys. Glad he didn't ask how HN thought before he posted this :)
620.
by raincole 6mo ago
I know the knee-jerk reaction is social media, but from the graph in the article, it seems to just get back to the same level as 90s.
621.
by raincole 6mo ago
I've found the traditional publishing industry really interesting. It's so hard to get approved or even noticed from the gatekeepers[0]. Even getting a rejection from an agent can take months. And agents are just the very first
622.
by raincole 6mo ago
Really? I've never heard it's considered wise to put refactoring and new features (or bugfixes) in the same commit. Everyone I know from every place I've seen consider it bad. From harmful to a straight rejection in code revi
623.
by raincole 6mo ago
It never ceases to scare me how they just run python code I didn't write via: > python <<'EOF' > ${code the agent wrote on the spot} > EOF I mean, yeah, in theory it's just as dangerous as running arbitrar
624.
by raincole 6mo ago
That's some weird semantic nitpicking. Wikimedia has a category of "photographs of the Sun": https://commons.wikimedia.org/wiki/Category:Photographs_of_t... Do you think they are not photographs of the S
625.
by raincole 6mo ago
What missing from the article is that they didn't use the same "slop score" to measure Show HN posts from <2023. Nor they released this script so the readers can verify it against known human-made landing pages. Why? Let m
626.
by raincole 6mo ago
> The key is to give the agent not just the ability to pull things into context, but also remove from it Of course Anthropic/OpenAI can do it. And the next day everyone will be complaining how much Claude/Codex has been dumbed
627.
by raincole 6mo ago
> Imagine asking to install spyware on your lawyers' firms' company laptops because you didn't trust them not to make some deal with the judge. This sounds unironically a good idea.
628.
by raincole 6mo ago
Many humans are great at their expertise but bad at communicating. How?
629.
by raincole 6mo ago
I know anthropomorphizing LLMs has been normalized, but holy shit . I hope the language in this article is intentionally chosen for a dramatic effect.
630.
by raincole 6mo ago
Because QuickJS is really slow. Don't be fooled by the name. It's almost an order of magnitude slower than node/v8. (I suppose the quick in QuickJS means "quick for a pure interpreter without JIT compilation or somethi

More ›