News for hackers. Stuff that matters.

Disposal8433

Which ones and what are those exclusive features that GitLab doesn't have?
Re:Using Claude Code SDK to reduce E2E test time on Saturday September 06, 2025 @03:02PM (#45151975)
Flagged because:

> what if we could run only the relevant E2E tests

The real title should be "Using Claude Code to Reduce E2E Tests by 84%."

Re:Ask HN: Who is hiring? (September 2025) on Saturday September 06, 2025 @11:44AM (#45150279)
Killing robots made by slaves. Where do I sign up?
> If outages [...] stop whole companies in their tracks

They should fucking learn how to code because no one in their right mind would depend on such an external service that can be easily replaced by cloning repos locally or using proxies like Artifactory. Even worse when you know that Microsoft is behind it.

Yes, most companies don't have good practices and suck at maintaining a basic infrastructure, but it doesn't mean GitHub is the center of the internet. It's only a stupid git server with PRs.

Re:The repercussions of a typo in C++ & Rust on Saturday September 06, 2025 @04:38AM (#45147637)
What do you suggest? Some kind of std::const_reference<Type>? Clang-tidy is enough in addition to the reviews.
It's perverse and disgusting. Also what about lazy students, neurodivergent, or poor people that are most likely not having a good outcome despite learning? In the end even average students would be ignored by the teachers.

What about everyone having the same education? What about not putting capitalism in everything?

Re:Type Checking Is a Symptom, Not a Solution on Friday September 05, 2025 @02:58PM (#45142289)
He's even more wrong when the Python typing system allows us to catch bugs in small scripts.
Like obfuscating text on Twitter, GitHub gists, or using an unreliable E2E encrypted service?

> without creating accounts or relying on any platform to keep it safe

You're still restricted to places where your JS decryption functions can be used. A novel idea would be to put both text and decryption inside a simple URL like a bookmarklet.

> The encryption method could be anything

Insecure then.

Re:simonw has vibe-coded 124 useful tools on Friday September 05, 2025 @01:01AM (#45135128)
I had the same reaction. It's a big stretch to call a "tool" 500 lines of JS that could be replaced with a few Bash calls, a regexp, or a small self-contained Python+UV script. Even worse when it's limited in functionality.

We failed at creating automation tools for the users (even if Apple tried that with AppleScript or Automator), but vibe coding throwaway scripts is not the answer.

It is prohibiting in a few ways: relying on a company in a foreign country, and no alternative if your account gets randomly banned. Also what would happen to developers in embargoed countries?

Those are unacceptable to developers outside of the USA.

Or yours...
They remove your privacy to make you think that your not worthy and that there is this pyramid. It worked!

I don't use Google's products either. As for Microsoft and other big companies, they deserve the same hate.

Re:First attempt will be 95% garbage: 6 weeks with Claude Code on Wednesday September 03, 2025 @12:17AM (#45112214)
Will the LLM take responsibility for the bugs and bad code introduced by the review? If it does and I'm free, then go for it.
Re:Jujutsu for Everyone on Sunday August 31, 2025 @02:56PM (#45085945)
> deep understanding

It took me less than one day to switch from git.

Re:Jujutsu for Everyone on Sunday August 31, 2025 @02:53PM (#45085897)
> Since jj doesn't have names branches

False. You need to call `jj branch set -r@ XYX` manually which can be a PITA but you only need to do that once you push. Or there is `jj git push --named XYZ=@` which moves the branch.

Re:Jujutsu for Everyone on Sunday August 31, 2025 @02:17PM (#45085517)
IIRC it's push force with lease, ie non destructive push force. No one will be bothered or notice what you did.

And if you have conflicts, it's really easy to rebase and fix any issue.

Re: on Sunday August 31, 2025 @11:28AM (#45083935)
It's called "reviewing code." I'm not taking any kind of responsibility for code that I haven't written myself.
> AI-enabled <something>

1. Scrum-enabled 2. Blockchain enabled 3. Web3 enabled 4. AI-enabled

Re:Are You Tired of AI? on Sunday August 31, 2025 @04:51AM (#45081591)
You created an AI-generated coloring book. Is it a way to farm karma?
Re:Show HN: OAuth for AI Agents on Saturday August 30, 2025 @02:41PM (#45076988)
You didn't answer OP's questions. It's not authentication at all.
Re:Ask HN: Do you have CPU or network spikes when opening HN? on Saturday August 30, 2025 @02:22PM (#45076883)
What's your astrological sign?
Re:Cognitive Load is what matters on Saturday August 30, 2025 @11:52AM (#45075660)
> three personas for software engineers

The kind of psycho-bullshit that we should stay away from, and wouldn't happen if we respected each other. Coming from Microsoft is not surprising though.

AI + subscription + tool that can be replaced with a simple notebook: spam.

Also, what happened to https://contactconverter.app/? Were you bored already or wanted to scam more people with an AI application?

Re:Agent Client Protocol on Saturday August 30, 2025 @11:39AM (#45075560)
> the obvious problems can mostly be solved by having the agent try to compile the program

The famous "It compiles on my machine." Is that where engineering is going? Spending $billions to get the same result as the laziest developer ever?

Re:Agent Client Protocol on Saturday August 30, 2025 @11:31AM (#45075488)
You can't disagree with facts. Every time I try to give a chance to all those LLMs, they always use old APIs, APIs that don't exist, or mix things up. I'll still try that once a month to see how it evolves, but I have never been amazed by the capabilities of those things.

> with popular languages

Don't know, don't care. I write C++ code and that's all I need. JS and React can die a painful death for all I care as they have injected the worst practices across all the CS field. As for Python, I don't need help with that thanks to uv, but that's another story.

Re:Agent Client Protocol on Saturday August 30, 2025 @10:16AM (#45074878)
> Nowadays, it is better to write prompts

Very big doubt. AI can help for a few very specific tasks, but the hallucinations still happen, and making things up (especially APIs) is unacceptable.

What's wrong with LibreOffice?
Re:Show HN: Generate human-sounding tweets from current news on Saturday August 30, 2025 @09:08AM (#45074319)
> I’d really appreciate feedback

It's a spam generator. I'll stay polite and stop there.

Re:Are You Tired of AI? on Saturday August 30, 2025 @09:06AM (#45074299)
I don't care about LLMs being integrated everywhere. It's a stupid idea but we can't prevent it. As for the AI-bros full of FOMO, they are wrong as usual. IF it stays as it is, anyone will learn to use those tools in a few hours (like Google in the past).

What pisses me off is the mountain of code that I will have to debug in the future because a bored lazy developer decided to take shortcuts. And I'm also annoyed that 50% of the posts on HN are selling subscriptions to dumb wrappers around the LLM du jour. "Subscribe to my birthday card generator web site $5 a month it's great" Fuck off.

Re:Is it possible to allow sideloading *and* keep users safe? on Saturday August 30, 2025 @09:01AM (#45074267)
Sandboxing should prevent most of those issues. We can't control the users giving permissions to everything, but with more control on those permissions, or disabled by default, a phone should stay pretty safe, or am I missing something?

Older »