Hacker Newsnew | past | comments | ask | show | jobs | submit | more nullbyte808's commentslogin

I wonder if the vale of Pixel phones that run GrapheneOS will rise dramatically?


Graphene is not a contender here, it's too niche. LineageOS is, otoh, going to become more popular.


Malware could still use DNS records for storage and access to bootstrapped payloads correct?


Yes, but it's not a problem, any more than downloading any arbitrary text is. You'd still have to have something execute the binary.


If anything, this would be more of a way to act as a command and control server


Does this work?


I almost switched back to Fedora Bazzite to get a working gamescope, but realized I can get HDR in sway and its actually more stable than Valve's mess of gamescope. Even though I have to use "--unsupported-gpu" flag, my Nvidia card works wonders in Sway, where as gamescope gives me a blinking cursor and segfaults.


Some things are not that simple and nix options come in handy automating other packages and services needed.


also the AI hallucinating nix options. I have to constantly check https://search.nixos.org/packages?channel=unstable


oh yeah AI realy does not seem to actually know which packages exist. I once asked AI to create a devenv for some Julia development and it pulled some packages out of its ass that just plain did not exist.


I'm overwhelmingly surprised about Claude's ability to know the package.

But the cut-off point in model / harness quality before it hallucinates everything but the general Nix syntax is staggeringly low.



I've been trying to get nixd LSP to work with Claude Code but I got stuck as they gatekeep it behind their "plugin" system and you can't just configure it in settings.json to point to a nix store path like mcps :(


My usual solution is to just clone whatever I need. e.g. in this case just clone nixpkgs and put in your instructions that it should do a git pull to make sure it's up to date and then refer to that whenever doing anything with nix. Agents are really good at using grep to explore repos even for something completely internal. Then you don't need any config or special tools. e.g. for work I just have a directory with like 30 repos cloned and my base AGENTS file I use says to refer to either them or live system state for ground truth. I basically never encounter hallucinations.

Same goes for the harness itself. Want to know how Codex works or whether it can be configured in some way? Clone it and ask it to look at itself.


There’s a NixOS MCP, it’s pretty good


I recently fixed the pipewire audio stutters by just giving gemini my flake and asking how to fix it. It suggested a few fixes and low and behold they were gone! Here's my flake with impermanence + yubikey login: https://github.com/leonewt0n/nixos


How do you know its AI? Who cares, sounds fine too me.


"Privacy isn't a feature — it's the foundation."

"No feed, no doomscrolling — just intention."

"Not your whole address book — just the ones you'd hate to lose touch with"

"You care deeply—you're just terrible at follow-through."

"You care deeply—your ADHD brain just doesn't..."


"You care deeply—your ADHD brain just doesn't..."

Well shit, that solves everything! What a revelation -_-


Even the word "intentional" alone is an AI tell.


"Quietly" is another watermark word I've noticed lately (though not in TFA to be clear).


The whole website was prompted. You can tell by the overload of emoji's on the page and every section having cards with hover effects. It's classic LLM design.


Funnily enough, while I definitely prompted but finding other website designs I liked and color schemes. I specifically wanted the hover effects because I love quirky animations. On the garden in the app try holding on a flower/ seed or click on a butterfly and enjoy the Easter egg ;)


Well, if someone can't bother to write something, why should anyone bother to read it?


that was one of the big reasons I dropped FB and Insta.

Low effort interactions just aren't worth it. IF it's not worth writing at least 3 sentences chances are I don't care - and won't ever care.

Like yeah it's cool to see what some guy I knew in high school is up to, but ultimately that connection ain't making my life better.

HN is an exception in the signal to noise ratio is far higher than a lot of other social media - though the bots are hitting it hard these days.


That's awfully able-ist isn't it?

How do you know that the author is capable of communicating fluently in english?

What if it were the case the the author was so excited about sharing a project but didn't know how to properly explained it and so took the extra effort to learn how to get a piece of software to explain it for them?

Would that then satisfy your requirement that the human behind the project has done enough work to earn your interest?


But AI provides the illusion of communication. Since the AI has no direct access to the user's brain, and has to go off the words they provide, if we're assuming that the person isn't capable generating words that accurately communicate their thoughts, the AI is getting all its information from the same flawed words we'd have access to if they didn't use AI, but destroying any signal encoded in the specific mistakes or choices they've made in its process of shaping their thoughts into something more polished.

AIs don't violate entropy, and can't create information from nothing. They can interpret, and expand, and maybe, just maybe, tease out meaning that a human would have missed. But the more sensitively they're tuned to pick up on small nuances, they more likely they're going to interpret a pattern that isn't there, and the more they're tuned to avoid over-interpretation, the more likely they are to miss something that is there, the same as how a human can aim to interpret something with high or low context.

The difference is, by filtering it through an AI, you're taking that capability out of other people's hands, you're (often intentionally) flattening and damaging signals people usually use to choose how to distribute their attention (often with the cry of "But it's not fair that people want to spend their attention on things that I'm not good at, I have to use AI to convince people to look at my work that they would prefer not to!!"), and when you do that without acknowledging the use of AI, it feels a lot like you don't care about any negative effects your actions have on the existing ecosystems of human creativity and communication, and you're going to get an appropriately hostile response.


> How do you know that the author is capable of communicating fluently in english?

Irrelevant; they can do the best they can and I'll do the best I can.

If the best they can do is have it ghost-written, then the best that I will do is not read it.

> What if it were the case the the author was so excited about sharing a project but didn't know how to properly explained it and so took the extra effort to learn how to get a piece of software to explain it for them?

That's not extra effort, that's less effort.

> Would that then satisfy your requirement that the human behind the project has done enough work to earn your interest?

Look, if someone isn't going to bother to write something, why would others bother to read it?


since when the world only has english speaking people?


Probably because its not War and Peace. It's the hook for some app you're probably going top spend all of 2 minutes on deciding whether it may or may not be functionally useful, and that equation is going to be largely solved independent of the quality of that marketing copy.


The avatars in the "testimonials" are all fakes by pravatar.cc, so take that into consideration. The site is bull shit.


I revisited the site and noticed that the testimonials section is now deleted. What a punk.

Please do not engage with this fraudulent clown-show of an app.


Deleted? Well, the Wayback Machine comes to rescue.

https://web.archive.org/web/20260305052854/https://poppy-con...


Yep, that was actually a placeholder (eventually with real people, and now soon with 500 downloads!) I didn’t comment out properly hence what someone else pointed out with the html commenting. Mistakes happen and I agree I should do a better job to proof read :)


I truly don't think it's possible for new software to have testimonials... especially good ones.


AI tends to be a buzz kill on products because it sends the signal "i can't be bothered to craft this deliberately."

So why then should we bother to interact with the product deliberately.

Around here most know how hard and time consuming it is to ship a production grade experience. AI helps a ton. it's not "wrong" per say, but it undeniably leaves an odor.


Btw, this project seems to be a solution looking for a problem.


We do mate, we do.


1. Makes no sense. 2. not true. 3. You can turn features you don't like off, like the AI 4. False. The bootloader is not locked. Linux does work, but it would be nice if they activly worked on kernel modules for their hardware. 5. Macbooks have the longest shelf life of any consumer PC. Period.


> Linux does work, but it would be nice if they activly worked on kernel modules for their hardware

Asahi Linux works on M1 & M2 macs, M3 and later are still being worked on. It should eventually get there, but we'll have to wait and see.


Like saying my PC boots up 2x faster so it must be 2x more powerful. lol


It was also one of the areas it was weakest in though, so this brings it way more in line with usable GPU territory.


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

Search: