yjftsjthsd-h

That seems as likely as anything, but that just moves the problem to counting, doesn't it? The model gets the duration, but I was under the impression LLMs struggled with the arithmetic that would make that useful.

Wait, how did it `sleep 3`? I thought LLMs didn't have a concept of time

Oh, it's going to be one of those weeks.

I'm pretty sure it is? It's not validation, but what are tests if not automated verification of certain things?

It's true that a computer can do everything, as long as you carve out exceptions for anything the computer can't do, but it seems like a rather pointless position. I feel quite comfortable saying that "most actual workloads people have" includes the functions once embodied by radio, TV, record/tape/CD players, and more.

I favor simpler machines, but that seems a bit of a stretch. Most people would like to communicate over the internet without every intermediate hop seeing their traffic (so, their emails, their chats, their credit card number, ...), therefore they want TLS which I don't think is practical on a 6502? To say nothing of watching videos and listening to music.

It's traditional:) The original machines with simpler hardware largely converged to BASIC, so it's a nice benchmark for "is this a Real Computer™"

> talking about "LLMs" is kinda a yellow flag that you might be a bit behind the times, here: in ~April of this year[1], "agentic engineering" became truly tractable for serious work

It's still LLMs though? This feels like saying we've moved on from using compilers now that we have IDEs; that's just a wrapper and we're still using the first thing.

I guess my question is whether it increases the difficulty more for LLMs or humans.

A camera and some rudimentary software should be a pretty good fix for color blindness. A camera and modern AI is a decent mitigation to partial or perhaps even complete blindness (prior art: Be My Eyes). A microphone, STT, and display would help with deafness.

> Instead of targeting extremely ancient hardware that almost no-one uses

The author was targeting hardware that they owned. One of the big themes of FOSS is scratching one's own itch.

Is Linux riddled with remotely exploitable vulnerabilities? A router shouldn't be running a lot of untrusted code.

Which is sad; a Pixel phone using libreboot + EDK2 to offer a "normal" boot path to Linux would be sweet.

So what OS are you building on and running on the machine? https://codeberg.org/ethical-foss/open-slopware#kernels doesn't seem to list anything I would consider a good alternative to Linux, although illumos or 9front becoming the preferred build environment for libreboot would make my inner nerd happy.

Are you publishing anywhere that I can follow? In particular, it would be really helpful if you could publish the prompts and workflow, so that the rest of us can more quickly use it to replicate your work on other machines:)

(Signed, the owner of several Chromebooks in need of better firmware)

AIUI: A password can be entered into a phishing site, a passkey can't.

That seems needlessly uncharitable. Isn't the more obvious explanation that they tried using it in a program, hit a failure, cached that result and never retried?

> From the very beginning, software engineering has been about automating other people’s work.

Excuse you; since the beginning of my career I've been automating my work, thank you much.

Sadly I am not aware of any guides. It's an unfortunate gap.

Perhaps it's a matter of who hits the pain point? Ugly specs make annoyed devs, licensing problems make distributors unhappy, and ecosystem issues leave users unimpressed. (With some spillover between categories, of course.)

Is there a real distinction? It can be a great format, but if actually using it sucks then in practice it's a pain to use/encounter. (I don't actually have this problem personally - all my apps seem to handle it fine - but the argument seems reasonable.)

It's really not; you've argued extensively with me that the moment a user can run an app with root the whole system is insecure. LOS sadly doesn't ship anything but `adb root` by default (although... they do that, so yes they do ship "user accessible root"), but they're still less hostile about it.

Anyways, since you're here perhaps you can answer my question from the other subthread: If I flash GOS and then flash Magisk on it, how hard is it to stay unbricked? Is it as easy as declining to relock the bootloader once, or is the system going to actively fight me on every boot?

I'm mostly talking about their stance on root. Certainly I agree GOS is vastly better than stock.

Not all short options are portable, but all POSIX options are short.

I suppose it depends how hard it tries to relock the bootloader; if I can tell it once to not do that then perhaps it's fine, but I don't want to risk a misclick soft bricking the device.

Although as an extension of that - I'm hesitant to use software written by people with such a philosophical difference. It might work today, but I wouldn't trust it to work tomorrow.

It's not quite that simple. I don't use GOS because GOS and I have mutually incompatible views of user control. I prefer that I control my phone, they say I can't be trusted with that.

I dunno, from watching the news I trust my setup more than theirs

Oh, you mean bash specifically. Then yeah, that's reasonable enough. Although I agree with the other commenter that sh/bash have the advantage of ubiquity.

No, powershell runs on Linux, it just sucks because it wants objects and everything on *nix speaks streams of text.

It's an open source community project; the devices they support are what the maintainers own.