Hacker News Reader

nvme0n1p1on Who should be held accountable when an AI Agent (accidentally) acts maliciously?

Okay fine, you caught me! There is, in fact, a giant global conspiracy to downvote your posts, qarl. It isn't just the fact that randomly-arriving independently-occurring events often arrive unevenly distributed. https://en.wikipedia.org/wiki/Poisson_distribution

This conspiracy goes all the way to the top. The illuminati assigned me personally to comment on your post. I've already said too much. If you never see me again, tell my wife I love her.

nvme0n1p1on Who should be held accountable when an AI Agent (accidentally) acts maliciously?

> how do you suggest I handle the following situation

The question is vague and doesn't seem related to the discussion. What do you mean "handle"? If you're having trouble handling it psychologically, see a therapist. If you're having trouble getting the output you want, look up guides on prompting. If you're anthropomorphizing the chatbot to the point you're worried about offending it... just don't worry? It's a computer program, don't overthink it, don't anthropomorphize it, just give it the input bytes you need to get the output bytes you want.

> I'm sure it's just a coincidence.

There isn't some grand conspiracy here. You just overindexed on the word "anthropomorphic" and didn't really understand what the discussion is about.

nvme0n1p1on US sanctions force The Netherlands off Microsoft and toward alternative NixOS

Yes, that was a pretty big dealbreaker... in 2011

nvme0n1p1on Who should be held accountable when an AI Agent (accidentally) acts maliciously?

You're discussing how to speak _to_ a model, but everyone else here is discussing how to speak _about_ a model.

I didn't downvote, but wow you're being aggressive, you have a lot to learn if you read the comments here with an open mind.

nvme0n1p1on GrapheneOS – When an app is slow

If you're expecting some leaked internal document that says "priorities: 1. benchmarks, 2. security", no, that probably doesn't exist. But their priorities are revealed by their actions.

GrapheneOS typically ships security updates within a day. But most of Samsung's phones get quarterly security updates[1]. There will never be a primary source with a Samsung insider saying "we're fine with our customers being vulnerable to widely exploited security bugs for the next 89 days, in order to save our massive corporation a few thousand dollars a month." But their priorities are revealed by their actions.

If the statement "OEMs don't care about security" is trivially proven false, as you said, then you should be able to prove it trivially by naming even just a single OEM whose update cadence matches GrapheneOS's.

[1]: https://security.samsungmobile.com/workScope.smsb

nvme0n1p1on So long Google, and thanks for all the nudes

And for the testers requirement? Should every hobbyist developer be required to buy a dozen phones and set up one of those click farms that fraudsters use? Or are we fine with ending the era of hobbyist development?

https://cheq.ai/wp-content/uploads/2021/12/clickfarm6-1920w....

nvme0n1p1on GrapheneOS – When an app is slow

If so, I'd love to know which OEM you're thinking of who ships an Android distro more secure than GrapheneOS.

nvme0n1p1on "As a Language Model": Chat Template Switches LLM Self-Referential Voice

THERE IS AS YET INSUFFICIENT DATA FOR A MEANINGFUL ANSWER.

nvme0n1p1on "As a Language Model": Chat Template Switches LLM Self-Referential Voice

The difference is

1. there weren't numerous real-life killings where the murderer credits GTA for coaching them through the crime

2. Rockstar Games didn't publicly say "sorry about the murders, but don't worry, we'll add more safeguards to GTA 6 to prevent even more people dying. GTA 6 will be the most aligned GTA ever!"

When a company admits to having blood on its hands, is maybe the point where things stop being "absurd" and start becoming real. But hey, what do I know. Maybe if HN existed in 2005 you'd have people commenting "who needs real people when we have computers, would it be so bad if lonely people have GTA as their only friend?" And I'd be the crazy one for engaging with them.

nvme0n1p1on "As a Language Model": Chat Template Switches LLM Self-Referential Voice

We should treat AI that recommends suicide to teenagers the same as we would treat a therapist recommending suicide to teenagers. I don't care if either/both/neither have souls.

nvme0n1p1on "As a Language Model": Chat Template Switches LLM Self-Referential Voice

Ah yes, my favorite LLM fallacy. "You used the wrong model! The newest fanciest model is perfect and makes no mistakes, have you tried it yet?" Let's force all of society's most vulnerable people to pay extra $$$ to Sam Altman, then surely all our problems would be solved.

It's just a convenient way to ignore the years of evidence of the harms. Any bad news can be swept under the rug, labeled outdated as quickly as it happens. Well here's one that just happened, maybe this kid should have used a fancier model too? Should OpenAI pay him a large sum of money for the good he's done? https://www.cnn.com/2026/08/15/us/arjun-aravind-massachusett...

nvme0n1p1on "As a Language Model": Chat Template Switches LLM Self-Referential Voice

It sounds like OpenAI forgot to include "make no mistakes" in the system prompt. Rookie mistake.

nvme0n1p1on "As a Language Model": Chat Template Switches LLM Self-Referential Voice

> We don't want destructive friends encouraging us to do bad things, but a real 'buddy', somebody who always has our best well-being in its interests?

GPUs are not people, and generated tokens can't have interest in a person's well-being. If you try to pretend otherwise, the results are not great. https://www.cnn.com/2025/11/06/us/openai-chatgpt-suicide-law...

nvme0n1p1on We Should Be Able to Change Our Languages

And even if LLMs do our thinking for us, it's still not a reason to fork the programming language you're using, unless you expect me to believe an LLM can understand `foo |> bar` but it can't understand `bar(foo(x))`. If you're not looking at the code, it doesn't matter how pretty the code is anymore.

nvme0n1p1on I'm tired of being on the network

Right you are, my mistake.

nvme0n1p1on I'm Tired of Being on the Network

It certainly does seem like he has a very specific axe to grind

nvme0n1p1on I'm Tired of Being on the Network

You're really moving the goalposts here. You started with "anyone who receives their packages is extremely privileged" and ended with "theft is something that retailers must be aware of in the course of their business". Obviously I agree with the second statement, but not the first.

Going back to your original post, now that we have some numbers to work with, do you really believe that 70.3% of Americans are "extremely privileged" people living in gated communities?

nvme0n1p1on Goodbye Google

You still haven't named an LLM. You know that Go is a board game, right? We're talking about LLMs, not games. Try to stay focused.

Name an LLM from 2016.

nvme0n1p1on Goodbye Google

> huge swathes of the world ride free

You really can't think of any way the rest of the world is paying for their dependence on US tech, besides monetarily?

nvme0n1p1on I'm tired of being on the network

https://www.oversight.gov/sites/default/files/documents/repo...

> Overall, total estimated annual losses range from a low of $5.2 billion in 2023 to $16 billion in 2024

> In 2024, the U.S. ecommerce market reached an estimated $1.22 trillion in revenue

So... between 0.4% and 1.3% of packages by value are affected, concentrated in specific geographic areas. A narrow segment of society highly affected, complaining often and making the issue seem wider than it is. The rest of us living our lives as normal.

nvme0n1p1on I'm tired of being on the network

It's not a problem for 99% of the country. No "crackhead" is going door to door between cornfields with a mile between houses, scoping out everyone's porches daily. I'm sorry you're having those problems, but the world is a big place, don't assume everyone lives the same life as you.

Even for the narrow segment of society in that situation, you're not required to give your money to Apple and then stay home 24/7 to immediately scoop up your packages. You could get a lockbox too, and guess what, they're cheaper than phones.

nvme0n1p1on Goodbye Google

Those don't mention any models.

You said, and I quote, "Each and every model"

So I'll ask once again. What models?

nvme0n1p1on Goodbye Google

Ten years? Can you refresh my memory, what were the top LLMs in 2016?

nvme0n1p1on Goodbye Google

There's no free lunch. You're paying one way or another.

nvme0n1p1on Meta takes down a critical video about meta AI Glasses after filming at Meta

https://www.engadget.com/2227710/a-dollar2-sticker-let-me-by...

nvme0n1p1on Swap, ZRAM, Zswap and Hibernate on NixOS

But why should it swap things out even when I'm not crunching data and there's no memory pressure? I'd rather everything stay in memory and stay fast, since I have enough RAM.

nvme0n1p1on Swap, ZRAM, Zswap and Hibernate on NixOS

Agree. I tried using a swapfile last year-ish, as the internet experts tell you to. I have 64GB of RAM. If I fresh boot and start a few apps I'll be using maybe 16GB RAM + 0GB swap. Then a few hours later I see I'm using 8GB of RAM + 8GB of swap... why? Why did Linux waste IO moving data to disk when there's no memory pressure?

That's why I don't use swap.

nvme0n1p1on Why Omarchy Exists

Having fun alone in your homelab is one thing. But encouraging others to use the thing you made, despite the systematic lack of attention to security, is quite another. I don't want YouTube video titles to have arbitrary code execution on my machine and I'm not sure why anyone would.

Though if it really is an OS dedicated to AI, as the article says, maybe the vibe coded security holes are what give it a reason to exist at all.

nvme0n1p1on Tokens too cheap to meter

Grep (or ripgrep at least) is i/o bottlenecked at this point. It's impossible to process data at faster than i/o speeds, since you have to get the data to the processor somehow. That doesn't change whether that processing is grep on a CPU, or LLM on an ASIC.

nvme0n1p1on An update on how we confirm your age group on Discord

https://man.openbsd.org/passwd.5

  The gecos field normally contains comma separated subfields as follows:

    name    User's full name.
    office  User's office location.
    wphone  User's work phone number.
    hphone  User's home phone number.

I see you support identity verification after all. They're collecting all this personal information! What's next, a felon flag?

Maybe one day you'll switch to a freedom respecting OS.