Hacker Newsnew | past | comments | ask | show | jobs | submit | nathansherburn's commentslogin

I learnt a new word today. Thanks!


'Thence' is another nice word


I've actually been following this project for a long time and it's none of the above. They're simply testing what a set of frontier models can do when given a goal and left to their own devices.

I agree this outcome is very painful to see and I really feel for Rob. It's clear people (myself included) are completely at breaking point with AI slop.

In this specific case though it's worth spending 30sec to read the website of AI model village to understand the experiment before claiming this was sent by Anthropic or assigning malicious intent.


Thanks for this context.

Here is one specific link to the project by Adam Binksmith from April 2025.

https://theaidigest.org/village/blog/introducing-the-agent-v...

Would have been a safer experiment in a sandbox full of volunteers participant. This got messy and causes confusion.


This is the equivalent of releasing a poorly tested and validated self driving vehicle into general traffic. Of course nobody would ever do such a thing...


Wouldn't frequent reauth be beneficial for stolen sessions?

E.g. If you set your session timeouts to a ~1 day then by the time your session cookies are up for sale on the dark web, they will be expired.

The article doesn't mention this and it's the main reason I advocate for auth sessions that are as short as practical.


If your session cookies were stolen, they can be stolen again and again too? Timeouts of 1 day assumes the cookies can only be stolen once.


With who in control? A Russia, China and US alliance? This seems very unlikely to me. All signs indicate we're moving to a multipolar world as far as I can tell.

China owns an increasing majority of global GDP and Trump seems to be either taking a path of peace, populism and or stupidity by pulling America out of its global alliances.


That’s one possibility; a different one is that Trump considers China as the real contender for the US in the 21st century, and wants to pivot the US attention there (and out of Europe.)


How does divorcing Europe make the US stronger against China?


I'm not a military strategist but if one believes that the real power competition in this century is going to be played between the US and China, it would make sense to concentrate on that direction. And if so, Europe becomes a distraction.


What if this drives the eu into the hands of China? How does that help the US versus china?


I don't think this is anywhere remotely likely.

The US performs a two-fold security function in Europe: first, it protects Europe from extra-European threats. Second, it protects Europe from intra-European threats. If however the US exits Europe, both kinds of threats become a reality overnight. When you think of what Europe is going to do if faced with any of these situations, it becomes very clear that China has absolutely nothing to offer on either front.


The extra European threat is not protected by the us anymore. And what even is an intra European threat? US is gonna intervene in a German - French war?


Ideally there would be a 20 year plan to drawdown in Europe and a replacement security architecture. I don't think there's currently any strategic vision, not that Biden had one either.


Agreed. Biden however saw Europe as allies; Trump sees them as "free-riders".


The idea that Trump’s brain is capable of producing original, substantive thoughts resembling long-term geopolitical strategies based in reality is adorable.


It's not necessarily his idea.


That's not a good-faith argument, sorry. Please be intellectually honest by attacking the strongest version of Trump you can imagine, not the weakest, brain-dead, reality-detached version of him that's easier for you to attack.


A great way to get around this is with an edge function from deno deploy.


I think this is spot on. The confusion for me comes from the fact that, as far as I can tell, I've never met a prig in real life. And yet they seem to be the biggest political issue of our time. Is it because I live in Australia and it's more of a US thing? Or is it because I'm not online as much maybe? I find it really confusing.


What Graham means by "prig" is, say, an HR person who informs you that you need to use your coworker's preferred pronouns.


If you're calling a coworker something that they are uncomfortable with enough to get HR involved, HR may be the prigs but you're being an asshole.


Yes on a pragmatic basis, if the coworker is male and gets upset at being referred to by "he", but it goes against your own personal beliefs to refer to him as "she", it's best just to refer to him by name and practise wording your sentences to be pronounless. And, where possible and not disadvantageous, to avoid situations where he's involved in your own work.

With this approach, he's less likely to make a complaint to HR about you (though he might notice the careful lack of "she", but that's much more difficult to make a substantiated complaint about), and you still get to stick to your own beliefs.

It's still somewhat vexing to have to do this, but at least it prevents you from getting in the crosshairs of HR.


If it goes against your personal beliefs to call someone what they prefer to be called at no expense to yourself, then you need some new personal beliefs.


What about neopronouns? Can someone just make up a new set of words like zi / zim / zis and expect you to remember them whenever talking to or about that person?


You already remember what standard pronouns to use to refer to each person in your life. They're words you learned right from the start of learning the language. The same isn't true of "neopronouns".


You seem to be in agreement with my sentiment.

There is also additional work in remembering which pronoun to use with which NB person. With most people, you just automatically say the right thing.


What have you done so far in cases where a coworker has asked you to refer to them only using a neopronoun?


As others have suggested I just avoid pronouns. This seems to be a common approach.


Seems like a practical solution. So what’s the problem?


Probably that gender ideology played a sizable role in getting Trump re-elected. It was the topic of his most popular ad, and IIRC the plurality of swing state voters said it was the most important issue for them.

So getting people to use preferred pronouns was a bit of a Pyrrhic victory in my book.


>IIRC the plurality of swing state voters said [gender ideology] was the most important issue for them.

I don't think that you RC. Citation very much needed here.

The issue of neopronouns is largely theoretical, since almost all users of neopronouns also accept the gender neutral 'they'. The Trump ads weren't about neopronouns, and I doubt that most Trump voters (or indeed most Democratic voters) could tell you what a neopronoun is.

Using people's preferred pronouns out of he/she/they is just common courtesy. It's essentially what everyone, social conservatives included, already does whenever they take someone's word regarding their gender rather than looking down their pants before talking about them in the third person.


Eh, ok. [1] I guess there must be another reason Trump's campaign spent hundreds of millions of dollars broadcasting that ad. It was way more than they spent on any other ad.

You're moving the goalpost here and pretending that neopronouns is not that big an issue. But it's obviously just part of the gender ideology issue, which was clearly part of the reason Trump won.

I can see that there are still people out there with their heads in the sand. I wonder who you'll help elect next time around?

1: https://www.megynkelly.com/2024/11/11/survey-finds-trans-iss...


You said ‘swing state voters’. The survey you indirectly link to is talking about ‘swing voters’:

>Our definition of swing voters includes those who are undecided in the presidential race, have changed their voting preference since 2020 (voting Democrat in one election and Republican in the other), or are independents who either indicate they split their votes between Democrats and Republicans, or who hold either favorable or unfavorable views of both Trump and Harris.

The sample of voters is weighted towards swing states, but judging by the numbers for 'All voters', gender wasn't the predominant issue for swing state voters in general.

I don't think that neopronouns are a big issue outside of hypothetical arguments on the internet. You can certainly link neopronouns to a broader issue that people care about. But this thread was just about neopronouns (starting from your question "What about neopronouns?") before you brought Trump into it!


Ok, well you're doing a great job of splitting hairs but have not brought any evidence. Good luck in 4 years, to us all!


I already have to remember people’s names, what’s the difference?


Or you could just call people what they want to be called when it does not inconvenience you in the slightest.


Personally, agree with calling people what they want to be called. That said, here's a thought experiment: What if someone is inconvenienced? What if someone feels uncomfortable using pronouns that don't match the sex of the person? What about uncommon "neopronouns" like "zhe", "xe", or "fae"?

Whose comfort gets priority in this situation?


We expect people to say things that make them uncomfortable all the time. I don't feel comfortable telling my boss that I'm the one who wrote the buggy code that caused the incident, but I have a responsibility to do it regardless. I might be expected to thank everyone involved in a project, even if I don't feel personally grateful to them. And so on.

Obviously there's no easy way to reason these cases from first principles. As it is, I'm aware that being affirmed in their gender identity is recognized as therapeutically important for trans people. On the flip side, I'm not aware of any condition that causes people to suffer significant distress due to using a particular pronoun. So in this case, I feel like it's a pretty easy decision.

EDIT: The "neopronoun" question was added after I replied, or I missed it. I have never met a person who expected me to use them, nor have I ever encountered a workplace environment where policies required their use, so I haven't formed an opinion.


I've never met anybody who used neopronouns either, I've only heard about it online.

I wonder if there are any long term effects of forcing someone to say something that they consider to be untrue? Taken to its most hyperbolic extreme, it could be used as a form of psychological torture, like something out of 1984, where Winston is tortured for not accepting that four fingers being held up is five, or "Four Lights" from Star Trek.

To get one to renounce what they know to be true and accept whatever you say without question is probably the ultimate form of control and subjugation.

For emphasis: "taken to its most hyperbolic extreme".

edit: more realistically, you could say that transgender or gay people might feel like they are compelled to lie about who they are in order to fit in, or in certain circumstances. Surely, if we recognize this as psychologically damaging, then we should recognize all other types of forced lying to be similarly damaging.


I don't think so. If someone shows you their baby you say "how adorable, how beautiful" no matter how ugly the baby is. If you haven't learned to accept that by the time you're an adult you're going to live a miserable life.


[flagged]


That analogy doesn't work. A closer analogy would be a vegan refusing to talk to a coworker who eats meat. That would be unacceptable to me, too.


The analogy in my comment was about adherence to philosophical beliefs, in response to what appeared to be a suggestion by you that such beliefs should be ignored if someone else finds them to be an inconvenience.

Could you explain why you think your analogy works, please?


A vegan doesn't eat tofu at you.

Your right to untrammeled adherence to your philosophical beliefs ends the moment that those beliefs result in conduct affecting other people. After that point, some form of balancing occurs.


[flagged]


Noticeably treating a coworker differently because of your beliefs about their gender identity is not a balanced and pragmatic solution.


[flagged]


Do you ask everyone you work with for a birth certificate so you can ensure you're referring to them using the terms you consider appropriate?


Yes absolutely, and I take a blood sample from each of them for karyotype testing while doing a thorough inspection of their genitals.

Any more daft questions?


So is it a problem to use the "wrong" pronouns for someone or not? You take people at their word on their gender every day. Why go out of your way to fret about what you call people only when they tell you they're trans?


You know we're people right?

Sorry to interrupt your slapfight about me, but you do know transgender people are... people... right?

Like we're not just abstract things for you to argue about on the internet. I have blood. And feelings. And goals. And dreams.

It's so bizarre to see people talk about people like me like this, like you've never actually interacted with one of us.

Like we're some kind of philosophical concept or something.

It's weird.

I don't expect compassion or self-reflection out of a green name on the orange site, but I felt the need to say this for some reason.

Have a nice rest of your argument.


That is just basic human decency, not being prig.


No, to the whole sibling thread. He’s talking about the “pledge of allegiance” required to get hired in a university or like company, circa 2020. Also posts that imply you’re a monster if you don’t conform.

Sticklers for rules are the traditional definition. I think most of us have met a tyrant before, ruler of a very small kingdom. Often in a government position.


Not sure if relevant but I thought this looked very cool.

https://www.squiggle-language.com/docs


I don't think the purpose of the checkbox was to collect useful data but an any case storing it as a date would probably solve the issue.


I'm still using the built in MacOS terminal. What am I missing? I mostly do webdev - maybe terminal features are more important for other types of programming?


Apple Terminal has a lot of problems. As others have mentioned, it lacks support for 24 bit color, enforces minimum contrast ratios without any ability to disable them (meaning you cannot set arbitrary color themes), is hopelessly bad at Unicode rendering (particularly with multi-codepoint graphemes, see [1]), and in general misbehaves in other myriad ways [2][3][4].

With both Ghostty and iTerm2 now freely available, there's really no reason to use Terminal.app.

[1]: https://mitchellh.com/writing/grapheme-clusters-in-terminals

[2]: https://github.com/neovim/neovim/issues/26093

[3]: https://github.com/neovim/neovim/issues/28776

[4]: https://github.com/neovim/neovim/pull/28453


Personally, my main reason to prefer Kitty over Terminal.app are (I believe Ghostty support all of them as well, so looking forward to try it out):

- “Kitty Keyboard Protocol” means that a TUI app can detect all keybindings. For example, if you install kkp.el in Emacs, then Emacs running in a terminal will pick up ctrl+shift keys, super keys, etc. on par with a GUI app. I believe NeoVim supports this out of the box now as well, so if you ever felt like binding Cmd+S to :w<cr> you now can. - “Kitty Graphics Protocol” means that I can let e.g. Matplotlib show images inside a terminal, even over SSH connections. If you’re annoyed at pop-up GUI windows, or struggling with viewing remote images often, this is a nice workaround. There are even attempts at making terminal PDF viewers (like termpdf.py); IMO that’s a game changer, even though the app itself is still in a “proof of concept” stage IMO. - Terminal splits. If you work a lot in a terminal, it’s nice to be able to full-screen a terminal and view many different shells or processes side-by-side. Last I checked, Terminal.app just doesn’t have this feature. Sure, you could use multiple windows or a multiplexer like tmux, but that comes with different trade-offs; for example, a native terminal offers smooth scrolling with a trackpad whereas tmux doesn’t. Personally, I use tmux remotely, but stopped using it locally. - I see a lot of people mention 24-bit colors as a main reason to not use Terminal.app. For me, I’m actually pretending that I have a 16-color terminal, because I’m tired of having to theme every command-line utility individually, I’d rather they all just respect my 16 chosen colors instead. The only exception is my editor, just because there are unfortunately few good 16-color themes these days, so I instead change my terminal program to be consistent with my editor theme and then let every other TUI utility believe the terminal only supports 16 colors to match.


Built-in terminal doesn’t support truecolor. Many CLI programs rely on that to look good.


What CLI programs? I have yet to run into any issues after using the built in terminal for its entire lifespan.


As an example, here is the Helix editor (neovim looks just as bad) in Terminal.app vs. ghostty (would look just as good in any other modern terminal).

https://imgur.com/a/terminal-app-left-vs-truecolor-tddRL0C


I appreciate you making an effort to visually display things, but FWIW I don't think that (or sibling) really gets to the root issue, which is less of capability then practical convenience and compatibility. Mac native Terminal perfectly well supports 256 colors (or arbitrary 16 colors with themes), which I've used, so you could absolutely make both sides look identical there. Unless someone is regularly viewing photos inside their terminal (which some do support and can actually sometimes be handy!) then it might not be immediately clear what 24-bit would bring to the table over 8-bit, regardless of editor or shell themes.

But in practice 24-bit was an easy lift for terminals under active development ages ago, and in turn made it trivial to have everyone across any platform specify exact colors more easily without any end user customization or arguments about "not quite what I wanted" in an 8-bit palette or whatever. Thus a lot of the ecosystem now makes use of it. Being able to replicate anything yourself, or get close enough, in a smaller colorspace is still extra grunt work for no particularly valuable reason, and could actually add up to be fairly significant work if one has a lot of more complex code coloring themes and such.


My example is precisely about practical convenience. If I try to do something very normal in Terminal.app, namely use a text editor with a theme, it will look like garbage.


Off topic but what Helix theme are you using here?


Some small tweaks on the built-in theme ayu_evolve.


You wouldn’t realize because 24-bit colors are silently dropped. You don’t get the full experience but you also won’t “run into any issues”, unless you start using something that relies on 24-bit color coding for core functionality, like my own log viewer.


How do you think reasoning happens in our brains? I wonder if it's more like an LLM than we realise?


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: