Skip Navigation

lichtmetzger

@ lichtmetzger @discuss.tchncs.de

Posts

243
Comments

1200
Joined

2 yr. ago

Moderating

ApocalypticArt feddit.de

ApocalypticArt discuss.tchncs.de

9h ago

According to this news story, AI is subtly making devices more expensive. Who would have guessed? Everyone! 🙄
Jump
lichtmetzger @discuss.tchncs.de 9h ago
The hardware shortage is self-induced by the LLM providers, though - it doesn't reflect actual demand. They're shooting themselves in the foot by making their models bigger and throwing more and more data and compute (and energy and water) at it, for increasingly diminishing returns.
In my opinion, that's not sustainable in the longterm and after the bubble pops, even the code generation ecosystem will change a lot. It's gonna be interesting to see which way of working will survive.

22h ago

What's a better feeling?

Jump

lichtmetzger @discuss.tchncs.de 22h ago

In this economy?

1d ago

According to this news story, AI is subtly making devices more expensive. Who would have guessed? Everyone! 🙄

Jump

lichtmetzger @discuss.tchncs.de 1d ago

because AI demand is still growing

Is it, though? Most of OpenAIs customers are free users and outside of the coding sphere there isn't a major usecase for LLMs. When the bubble pops and people will have to pay a lot more for their token usage, a lot of them will just turn away. The free users will never convert to a paid model as well.

Nobody is going to buy that stuff in bulk just to start desoldering chips.

Don't underestimate Chinese vendors doing absolutely everything to turn a profit.

1d ago

According to this news story, AI is subtly making devices more expensive. Who would have guessed? Everyone! 🙄

Jump

lichtmetzger @discuss.tchncs.de 1d ago

HBM memory can technically be used by consumer GPUs. In 2015, AMD released the Radeon R9 Fury X which used HBM. When the bubble pops and there's an extreme oversupply of HBM sitting in warehouses, we will get those types of GPUs back.

And even if the big players decide not to do it, crazy people in China will build their own products and adapters, they will even desolder HBM from worthless datacentre GPUs.

Either way, it's gonna be glorious for us all. Let's just hold on for a while longer, until all of the dominos have fallen.

2d ago

What’s the dumbest mistake or misinformation you’ve ever heard AI make?

Jump

lichtmetzger @discuss.tchncs.de 2d ago

They're a little more competent, but that doesn't say much, given how extremely bad this project was run - it's probably very hard to do worse.

2d ago

What’s the dumbest mistake or misinformation you’ve ever heard AI make?

Jump

lichtmetzger @discuss.tchncs.de 2d ago

That's the one positive thing that came out of this project. :)

Fun fact: The project was also supposed to have a feature where the customer could add new pages with content and then decide where in the main menu those pages should appear.

WordPress internally uses a very basic system for assigning menu items - if you have ten menu items, they have an order that goes from 1 to 10. Number 1 is the first, 10 is the last - very simple (and stupid, because you can't add numbers in between to add new menu items on the fly).

The old fart (or the AI) wrote a system where the customer would just put in a number for the menu order. So if they created a page and then set it to be the fourth item in the menu...there already was a fourth item and WordPress decided at random where to put the menu item instead. Sometimes it was in the correct place. Most of the time it landed somewhere unintended..

What you would actually have to do here is look at all of the menu items, shove the new item into that list and then recalculate all of the numbers and save them. AI was not able to do it, no matter how you prompted it. We tried Gemini, Claude Code, ChatGPT, all of them shit the bed completely.

So there we were - this feature absolutely had to be implemented by a human. I didn't have the time for it, the old fart left and the student was completely useless without chatbots helping him out. That put the project on hold for six months until I was available again, but then I left. Whoopsie!

Fun fact #2: The company eventually went bankrupt because of projects like this and was bought out by a competitor.

2d ago

What’s the dumbest mistake or misinformation you’ve ever heard AI make?

Jump

lichtmetzger @discuss.tchncs.de 2d ago

I work in software development. At one of my jobs i had an elderly colleague with onset dementia, who was causing all kinds of problems.

As a last hurra before going into retirement he wanted to prove to everyone (or better, himself) that he could still do it and code a big application all by himself.

So he decided to take on this project where he would build an intranet portal for a customers internal website...with WordPress.

With the help of AI, he coded a plugin and a theme. Which in itself is not a problem, but he made them depend hard on each other. There were constants defined in the theme that were used in the plugin and the AI did not add any checks, so when you disabled one of the two, the site would just throw an error 500.

When it went live, the customer noticed that only administrators were able to login, any user level lower than that (i.e. all of the customers employees) would just get an error on login.

By then the old fart had decided he was leaving the company, so I received the task to clean that up. He had used two different versions of an external library that were slightly incompatible with each other and depending on a race condition we would get one or the other, the LLM hooked into things that didn't actually exist, functions were duplicated between the plugin and theme, the CSS was 20.000 lines. It was an absolute shitshow and basically unrecoverable.

The funniest thing was that this guy was absolutely proud of his work and anyone who just slightly criticized him for the atrocious code quality just got shouted at loudly.

My boss then hired someone fresh from university to fix that and he just put more AI code into the project, making it even worse.

When I left, the project was still not launched after over a year and the customer was (rightfully so) super pissed-off.

3d ago

Steam Deal: Save 30% on The Elder Scrolls IV: Oblivion Remastered on Steam

Jump

lichtmetzger @discuss.tchncs.de 3d ago

Can't recommend. This is using the old engine under the hood, they only slapped the Unreal engine on top of it for the graphics layer. That means a lot of the old Oblivion bugs and general jankyness are still there.

It's also suffering from microstutters that turn into macrostutters when you're outside in large areas.

It looks nice though.

4d ago

Amazon is gutting its AI division after sustained failure

Jump

lichtmetzger @discuss.tchncs.de 4d ago

Nvidia is doing some real shady stuff, like selling GPUs and then renting them back from the companies they sold them to, which inflates their own value. That money doesn't actually exist, though.

5d ago

KernelCorruption

Jump

lichtmetzger @discuss.tchncs.de 5d ago

I don't really see this as a problem. These companies don't dictate the direction of the Linux kernel project, there are independent people like Torvalds and Kroah-Hartman at the helm.

The Blender project is also supported by major players like Nvidia, Intel and Netflix and they still work for the wishes of the community and not for those companies.

5d ago

How to embed vomit in your own code, by JetBrains

Jump

lichtmetzger @discuss.tchncs.de 5d ago

If JetBrains could spend more time fixing all their broken shit and less time on this, that would be amazing.

After each update of WebStorm I have to manually edit the VM options to enable Vulkan support, otherwise it will render the editor in software mode with ~5fps.

Smooth-scrolling a little bit of text in 2026 seems way too difficult for a commercial project. If AI is so great, why can it not implement that?

5d ago

How to embed vomit in your own code, by JetBrains

Jump

lichtmetzger @discuss.tchncs.de 5d ago

JetBrains is just weird. Their plugin for WebStorm technically supports local LLMs, but it defaults to OpenAI models and after each update I have to reconfigure it again to use my own machine instead of a massive datacenter poisoning freshwater.

One time it had a hiccup and couldn't process my input and when I clicked a simple "Resend" button, it just randomly decided to send all of my code into the cloud to be handled by OpenAI instead of my local machine. Thanks, that's totally what I want you fucking morons.

Cloud providers cannot be disabled in the plugin and the "Automatic" setting in the model selection does not mean it'll always use your own local LLM, it just does what it wants.

Their priorities seem pretty clear here - they added support for local LLMs because otherwise they'd piss off legitimate developers, but it's very barebones and the actual focus is on stupid cloud LLM providers like Anthropic and OpenAI. Can't wait for this bubble to burst.