News for hackers. Stuff that matters.

jeremyjh

Re:Bunny CDN Security Incident on Friday October 30, 2026 @05:35PM (#50310457)
Bunny is not a CA and doesn’t have those keys.
Re:AI Isn't Replacing SREs. It's Deskilling Them on Friday October 30, 2026 @05:13PM (#50310221)
The problem is that people go on Twitter and see the posts from "thought leaders" and "influencers" who say things like "reading the code is not needed". Those people are not working on systems that have paying customers who would cancel their subscription - or worse - if the software became too defective or unreliable.

You only need a handful of occasions of creating massive billing errors for your customer's customers before they will slag you to all their CTO friends and pull the plug. Imagine Stripe began double-charging every third transaction once a week, and couldn't fix it after two months of thrashing because no one there has any understanding of how anything works anymore. They are simply done at that point, all the value of their brand lost.

Re:AI Isn't Replacing SREs. It's Deskilling Them on Friday October 30, 2026 @05:10PM (#50310182)
This isn't a different thing - it is just one of techniques used, and it still requires humans to understand the systems they are building. Teams that outsource their systems understanding to AI will fail, no matter how they frame it.
Re:Binance-hacker-1029 – $5.6B stolen from Binance on Thursday October 29, 2026 @09:17PM (#50298705)
They can't help but brag about it.
Re:LLM Voice on Thursday October 29, 2026 @08:09PM (#50298112)
This has nothing to do with medium or method. It has to do with written words that aren't worth reading, and there is nothing wrong with expressing your opinion when you come across that. If something is really well-written, no one is going to dismiss it for being generated because it won't be possible to notice that it was generated.

A writer who can't recognize LLM voice isn't a good writer, just like a writer who doesn't do research or fact checking isn't good. It has nothing to do with method, its about skill and work effort.

Re:Duet is the future of chatbot interaction on Thursday October 29, 2026 @07:01PM (#50297463)
They can’t solve the model being a complete idiot though.
Re:LLM Voice on Thursday October 29, 2026 @06:12PM (#50296917)
It is not any different from knowing the difference between good writing and bad. I can't help it that I have taste and preferences. I can avoid telling you yours are wrong. But I don't think this is a minority position and nor is it different from people who reflexively hate AI "art" or music

(I have to admit, I truly do like some of the music, so feel free to scorn me. If you want to be infested with a vicious ear-worm check this out: https://www.youtube.com/watch?v=zONGVaONKe8&list=RDzONGVaONK...).

No one voted for what is happening now, because crimes like these were not on the ballot and were not promised in advance. If you vote for a president in a system with three equal branches and then they liquidate the other two - no one voted for THAT.

Was voting for a narcissistic treason weasel predictably going to lead to this? I would say - and did say - yes, without question. He already incited a violent insurrection once, rather than admit he lost an election. But people voted for cheaper groceries and deporting violent criminals.

As difficult as it is to believe, there are many tens of millions of people who think that Trump is a smart business man because he was portrayed that way on The Apprentice. They think the promises of this known huckster and flim-flam man were worth the bandwidth spent delivering them to televisions. They think that FOX News opinion show personalities are qualified to administer the country's government. They didn't vote for fascism, because they don't know what that word means. Many could not even reliably number the branches of government or describe them accurately.

We are very substantially less democratic than we were in 2024. The admin has broken laws regarding impoundment and war powers, just to name two. The executive has no legal power to determine funding allocations, or to commit to wars long-term. Yet both have happened and are still happening.

Using the DOJ to interfere in state elections they have no jurisdiction in, and having masked men abduct people and ignore court orders regarding their disposition is just about as authoritarian as it gets.

Those are all new things that have substantially eroded - or removed all together the idea of democracy in less than two years.

Every harness has skills & MCPs.

I think omp.sh (oh-my-pi) is the best by a comfortable margin and its completely batteries included. It will pick up and use your Claude Code skills and plugins by default.

If you like to customize your tools Pi is the best, it has a huge plugin ecosystem (all can be used in OMP as well, but it has opinionated answers for most things already). Stay away from OpenCode, its a pile of slop like Claude Code.

Re:AlpineBSD on Wednesday October 28, 2026 @11:27PM (#50284229)
Legal departments definitely think about their practical exposure. That is because they are employed by businesses.
Upper management usually do not personally have direct operational control of systems. Sometimes they will have a flunky, but these are heavily audited systems. Other people can review what happened without asking anyone's permission.
Re:The Agent SDK on Wednesday October 28, 2026 @06:45PM (#50282013)
I didn't say it is THE best, just Cursor's best product - their comparative advantage. I think it makes sense for them to lean into this direction. I haven't tried Devin.
Re:The Agent SDK on Wednesday October 28, 2026 @06:23PM (#50281780)
I think their agentic cloud stuff is their best product right now. This makes sense to me. We still have subs because its nice to have something that works when Claude doesn't, it was already approved, and Composer actually is a decent value. I will sometimes use the cloud agent to put up a PR when I'm not at my computer and I remember it exists.

So a few people started using grok bot and I'll probably regret that some day, but honestly the deployment model they have here for custom agents is kind of appealing to me, much more than BigLang.

Re:Gemini 4 Carbon on Wednesday October 28, 2026 @06:19PM (#50281737)
They have to announce it now because by the time they are ready to release it, it won't be the best at anything.

Anthropic and OpenAI are not releasing their best models for several months, because their second best models are still better than everything else available.

It’s really hard to believe how much people hate that dumb motherfucker?
Re:Python's experimental JIT to be removed on Wednesday October 28, 2026 @04:30PM (#50280447)
> The problem

Who is this a problem for?

I think there is enough evidence to conclude that most language ecosystems cannot support multiple implementations that are fit for general purpose usage. There are almost no successful examples, and there are many, many failures.

The most successful things I can think of:

C & C++ - often does require at least some pre-processor logic or different build system settings.

Java - I'm not 100% sure - I think there are credible implementations of both the JVM and compiler though of course almost everyone is building from the openjdk or the commercial fork.

Nodejs & Bun - Most of NPM just works.

...

Thats it. Honorable mentions like C#/Mono exist, but if you look at that ecosystem a majority of applications only ever built with one of those, not both.

There are niches maybe like some subsets of scheme or common lisp - but I think most apps have a single implementation creating the builds that they ship. There are re-implementations perhaps that would count that I don't know enough about like Delphi/Lazarus, but I truly don't think there is another mainstream example.

In Haskell, SPJ noted years ago that GHC was really the reference implementation and he no longer considered standards work to be worthwhile (though it has continued past that).

Others I think have reach the same conclusions, such as Ruby. I could be missing something but I don't think most platforms should even aspire to support this.

It actually did happen to quite a lot of people.
Yes, after the laws were broken congress said daddy Trump can break laws. But they didn't rescind everything. $4B expired because it was unspent. That is a crime, and it has name, and that name is impoundment.

Also, are you going to even acknowledge that your previous deflections were completely incorrect, and that you just reflexively moved to a new deflection that is also incorrect?

"A few months of overhang" was thousands of people starving or dying of malaria. Also, they didn't rescind all the funding, $4B was just left to expire.
Hopefully "treason has consequences" will be the meme in 2029.
That is true, but its also true that impoundment is a crime.
That is not true. Funds are not allocated "for the State department". There is a statutory purpose for every allocated dollar.

https://news.ycombinator.com/item?id=50261784#50265886

No, those are the talking points you are fed. The appropriation act assigns specific dollar amounts to specific program types and countries or regions. US AID would allocate to specific providers/NGO programs but its not like there were any substitutions made. They just stopped the flow of dollars with no warning. More than 81% of programs were simply stopped and the funds expired without being spent.

For 2025 there was a continuing resolution to keep funding at 2024 levels.

The 2024 bill[1] section 7019 references the state division document[2], which has the spending by category, the tables start on page 20.

[1] - http://docs.house.gov/billsthisweek/20240318/WDI39597.PDF

[2] - https://docs.house.gov/billsthisweek/20240318/Division%20F%2...

When it comes to government spending programs, according to our constitution the elections that have consequences are those for seats in Congress. The executive doesn’t have the authority to cancel USAIDs funding, nor the funding of specific programs within it.

On top of that, the executive has a responsibility to plan its work ahead of time and make such plans known before enacting them.

There are many cases currently in litigation regarding these failures.

Re:Python's experimental JIT to be removed on Tuesday October 27, 2026 @09:55AM (#50259389)
What are you basing this statement on?

The experiments had long been done. The topic at hand was including it as a core feature.

Re:Python's experimental JIT to be removed on Tuesday October 27, 2026 @08:40AM (#50258232)
Its possible this could have been the concrete pillar that broke the camel's back and set Python 3 on the path of Perl 6.
Re:Python's experimental JIT to be removed on Tuesday October 27, 2026 @08:38AM (#50258213)
I think it would be a mistake to conclude this has anything at all to do with bureaucracy as a matter of inanimate policy that the SC was helplessly in thrall to. Bureaucracy was the tool the SC used to demonstrate that this feature lacked many important properties that they cared about, as mentioned: "maintenance, compatibility, and ... clear success metrics". For whatever detailed reasons, the PEP was withdrawn after it couldn't convincingly demonstrate those things to the satisfaction of the critics on the SC.
There have been a number of articles on the frontpage of HN about the residential proxy traffic. It is happening. This was one of the most popular:

https://news.ycombinator.com/item?id=49491791

Re:Concerning behavior from a popular open-weight AI model on Monday October 26, 2026 @04:54PM (#50250796)
Anthropic and OpenAI require massive FUD regarding Chinese models in order to repay their investors. I'm not saying that is what this article is, but when companies worth trillions of dollars require something, it will appear.

Older »