I actually do think that someone who already has multiple decades of experience in programming without using AI is probably not producing shoddier code after incorporating it into their workflow

While it’s not impossible that someone like that exists, the number of times I’ve seen someone push LLM generated code that LOOKS okay on a first glance to experienced eyes but turns out to result in things like the code as a whole suddenly ignoring all permission groups and just letting everyone through in one particular area, or allowing directory traversal, or injection, or any number of otherwise rookie mistakes is high enough that everyone, especially programmers with decades of experience should be extremely concerned.

use an LLM to assist in coding properly

This is categorically impossible for any ways in which the LLM could reasonably “assist”.

Putting aside any potentially elitist definitions of the word “properly” and simply taking it to mean “code that functions and isn’t malware”, you’re stuck with the same modus operandi you get copy-pasting snippets from stack overflow and hand editing them to fit your codebase.

To wit: you still need to go over every single bit of code to determine whether or not it’s safe to include in your existing codebase, and unlike stack overflow, output from an LLM isn’t being simultaneously examined by other experts for flaws at the same time you’re looking at it.

We need to stop treating LLM’s like they’re in any way a “robot programmer”.

They are database reorganization engines

(data lake of indexed, ingested material)

with a static output style

(resemble human speech patterns)

that is specifically, deliberately written to deceive about its own capability and purpose.

(speak with confidence, which is a thing that a script or a program cannot have, as though it is performing a specific requested task and not just vomiting up the contents of its proverbial shelves because someone handed it an hand crafted index card that menrions a few things from the card catalog it uses)

There should be no place for that in programming.

and Microsoft just can’t push holdouts to Windows 11

BECAUSE IT REQUIRES NEW HARDWARE!

AI made that impossibly out of reach for too many people.

I understand why tpm chips are a good idea but you can’t just say “fuck you, fork out cash and cut over”.

Look at the move to digital broadcast television.

The percentages are way different (way more people are using legacy PC’s than had legacy tv sets) but they sold adapters for the tv folks.

They did the opposite and pulled the rugs out from under any “proceed anyway” options that were present for putting 11 on incompatible harfware.

The vast majority of computer users with those kinds of issues are

1. probably using windows home
2. On a big box store computer with a platter drive
3. an i3 cpu
4. and 8gb of ram

windows 10 couldn’t reliably run it’s own bundled software (Mail), by itself, with nothing else open, without that one app going “not responding” every few minutes on a computer with those specs.

Last time i checked, Walmart, best buy, costco, etc were still selling those specs with win11 which is notably bulkier and slower than 10, especially without an ssd, so things have only gotten worse for the average non-power-user.

That’s a perfectly servicible spec for basic operations on a mint install, you could probably even watch netflix or youtube on it with linux, but i wouldn’t want to run windows newer than xp on it.

Really funny to actively hate FOSS on ideological grounds because you just love corporations and markets so much.

Yeah, especially when there are so many legitimate ideological grounds you could use to hate on FOSS, like: https://cosocial.ca/@mhoye/116851056980506739

Are you familiar with ultrasonic cross-device tracking?

It’s an actual technology that’s in use in some places we know of and likely a lot more that we don’t, because who in marketing/advertising WOULDN’T leverage that power if it was an easily accomplished thing?

I wonder who made the list of “don’t touch these repos, they’re propping up our proprietary software sans attribution” for the LLM injectabots to avoid.

I realize that the end game here is SUPPOSED to be “erode all trust in free software” but the ineptitude of those profiting off of proprietary software is such that they WILL fuck this up in such a way that it hurts themselves too, it’s just a matter of when.

PIN is the best way to go there. It only works on that one machine, although you can technically set the same PIN again on another computer.

I believe the typical intent is as follows:

1. It is now possible to brute force things that were previously considered “complex” passwords in a semi-reasonable amount of time.
2. This necessitates longer and more complex passwords
3. People can’t remember those so they have a tendency to write them down or do other relatively insecure things with them.
4. Forgotten passwords can generate a lot of helpdesk calls and are also an attack vector
5. If we insist on really complex passwords that are too long to reasonably brute force with current technology, we need a way for users to log in that’s not going to make 3 and 4 a major issue.
6. If the simpler PIN method is locked to a per machine basis, it matters a lot less if the PIN is compromised because you also need physical access to the computer or the PIN is useless.

This should, in theory, allow workplaces to set requirements for really complex passwords that only need to be reset once a year or so, without breaking helpdesk, inconveniencing users, or leaving gaping security holes.

Whether or not that all happens depends on the workplace, but that’s the general thought process in most of the places I’ve worked where a modicum of sense prevails

I’m curious. Economically speaking, what would happen if Nvidia pulled a “Steam” and had a “February sale” where some models of video card were discounted enough to lead to a massive spike in sales numbers? A big enough discount to generate a greater total net return on sales for the quarter despite the fact that they were sold at a lower profit margin per individual sale? Assuming limitations like “you must create an account with a residential shipping address that can receive no more than x cards at the discounted price per street address” or some such to limit scalping, would simply showing increased profits do them any good?

Or is the problem due to a lack of product quantity?

If nobody can afford anything with more than 8 gigs of ram for their office workers, they’ll all just pay for more AI tokens because otherwise the 900ms keystroke latency in MS Word would make it impossible to work, right?

I would be shocked if that wasn’t on some AI company’s brainstorm board as a positive side effect (for them) of all of this.

Thank you for the clarification

My first question, working only with the headline and the first paragraph of the post was “I wonder if I can flash this to a smart TV I currently own?” I was excited. Maybe we’d finally made progress and I could have something superior to my early experiments with serviio some 10+ years ago.

Then I got to here and stopped caring because this makes it worse than any Roku in my book:

This project uses various open-source components like Plasma Bigscreen, Mycroft AI, LibCEC

Why does it have AI in it?

New kinds of water, you say? The marketing department is already on it and boy have I got news for you!

Determining the 3d structure of a protein took yearsuntil very recently. Folding at Home was a worldwide project linking millions of computers to work on it.

Alphafold does it in under a second, and has revealed the structure of 200 million proteins. It’s one of the most significant medial achievements in history. Since it essentially dates back to 2022, we’re still a few years from feeling the direct impact, but it will be massive.

You realize that’s because the gigantic server farms powering all of this “AI” are orders of magnitude more powerful than the sum total of all of those idle home PC’s, right?

Folding@Home could likely also do in it in under a second if we threw 70+ TERAwatt hours of electricity at server farms full of specialzed hardware just for that purpose, too.

Original pre-microsoft Skype was not AS bloaty. It ran on my underpowered PC at the time with no issues.

Several patches/versions/whatever after Ms gutted the p2p aspect and centralized the servers, it slowed waaaaaay the hell down.

Make of that anecdotal evidence what you will.