This is absurd:

it is important to note that Valve’s promotion of games that glorify violence and guns helps fuel the dangerous epidemic of gun violence, particularly among young gamers who can become numbed to grave violence before their brains are fully developed.

Quoting this tired trope from the early 2000 [External Link]'s is bizarre and ill considered.

Sure, go after loot boxes, they're rotten to the core, mostly. Although as someone has already pointed out, it would be nice to see some consistency - also get rid of Magic the Gathering packs, Pokemon packs, Pannini packs, or anything else that's marketed to kids that is a lottery.

View this comment - View article - View full comments

Quoting: ItsRainingSomewhere

There's fixes for the Discover software centre with Flatpaks

I've come across random weird issues with this on my Fedora KDE boxes. Sometimes the Flathub source is listed under Settings, sometimes its not, sometimes its not listed and packages still update normally. It's hit or miss. Noticed it first on 6.5, but seems to be better/normal/fixed now on 6.6.

What distro are you on? The issue I had with any distro that wasn't Arch-based was that Plasma was typically at least one major version out of date! Even Siduction, which is based on Debian Sid, tended to be months behind the release schedule.

View this comment - View article - View full comments

Honestly looks like about 75% of contributions mention Linux support. Maybe more - incredible support from the penguins. On the back of that response I hope they devote some time to the Linux version instead of just shoving out untested builds.

View this comment - View article - View full comments

Quoting: mr-victory

Quoting: scaineBut... again, this isn't what I'm talking about when I disparage genAI. From the article:

I tried to find a short source as I recalled that something akin to GenAI was used in this specific discovery, instead of checking the content of the source. My bad.

PS: If I use at least 2 quotes or links, do I end up in the mod queue?

It's just any external link, or embedded picture, I believe. Sadly, quoting a post with a link in it will also trigger the queue.

View this comment - View article - View full comments

Quoting: mr-victory

Quoting: ScottCarammellstill waiting for that legitimate use case the technology is sure to get. I mean this is supposed to be the next motor vehicle or whatever, right? there's gotta be something huge right around the corner. not like this was all a giant, unbelievable waste of time, money, life and resources for absolutely nothing of value. that's impossible. the people on wall street said otherwise. they're supposed to be smart, right?

I'll drop another one: generative ai found an algorithm to multiply matrices faster

https://spectrum.ieee.org/matrix-multiplication-deepmind [External Link]

But... again, this isn't what I'm talking about when I disparage genAI. From the article:

The scientists developed an AI system dubbed AlphaTensor based on AlphaZero, which they earlier developed to master chess, Go, and other games

They didn't use ChatGPT, or Claude, Gemini or Perplexity here. The kind of genAI that makes people's blood boil is very different from the useful, targeted ML instances that have existed for decades before all the genAI hype kicked off around 2022.

The actual technology of LLMs might have merit. But the over-hyping and over-investments in genAI... the lack of governance and ethical standards... the illegality of the training data... the inefficiency in the middle of a climate crisis... does not have merit.

And you know what? Finding one of two loosely related wins for LLMs won't balance the scales here anyway. Too much damage has already been done.

View this comment - View article - View full comments

Well okay, I guess, but that feels a bit pedantic? Did you read how they used genAI here? They designed their own ML model using medical data. Technically, okay, sure, that's genAI, fine. But it's hardly the kind of genAI that everyone is talking about and hating on. It's not plagiarising books, pirating material, talking kids into suicide or whatever.

It's probably also not quite the planet burning water hog either, because the data sets are so targeted.

Maybe you're right, and I'm the one being pedantic here, but I honestly don't feel like we should conflate these two things and thereby justify the car crash of genAI over-investment by pointing at fringe, targeted medical uses.

View this comment - View article - View full comments

Quoting: Eike

Quoting: ScottCarammellstill waiting for that legitimate use case the technology is sure to get. I mean this is supposed to be the next motor vehicle or whatever, right? there's gotta be something huge right around the corner. not like this was all a giant, unbelievable waste of time, money, life and resources for absolutely nothing of value. that's impossible. the people on wall street said otherwise. they're supposed to be smart, right?

https://asm.org/articles/2025/august/ai-next-frontier-antibiotic-discovery [External Link]

Oh man, please don't confuse AI, which is vital to this kind of research you've linked, to genAI, which what we're generally complaining about here in articles like this one (even though it's often just termed as AI).

That conflation of AI and genAI is often how the techbros will try to justify the ridiculous investment in this tech, even though medical research AI won't likely see a single penny from those investments, which are all focused on genAI pish.

View this comment - View article - View full comments

Quoting: vic-bayIf you see ai slop long enough, you will learn to identify it immediately. Yes, it is still a time waster, but it is nowhere as bad as it looks at the first look. Creating slop commits and pull requests takes way longer than reading it for half minute and smashing ban button.

That might be true, but Godot currently has nearly 5000 PRs awaiting review, and 50K closed PRs overall. I think that when you're dealing with those kind of numbers, it still adds up to a near-insurmountable problem.

View this comment - View article - View full comments

using AI to fight AI which they said "seems horribly ironic" but they might have to eventually

A couple of years ago, I was on a cyber security panel which asked "what is the best use for AI, in your business". We were a panel of investment managers, and the top answer was "writing out responses to client due diligence questionnaires". The second-to-top answer was "reading responses from our own vendor due diligence questionnaires".

You couldn't make it up.

But no, fighting AI with more AI is a poor choice that simply feeds the beast (in this case, the beast being AI itself). A ban on AI is ideal... except it will become increasingly difficult to know when AI has been involved.

I wonder if it's possible that new PR requests (that is, PRs from new contributors) go into a queue, and a voting system is introduced among Github users. This wouldn't be the Godot team themselves, just interested parties. As more people vote on the "good" PRs, or at least the desirable PRs, they rise to the top, and only then get reviewed by the Godot team. Once they've contributed (well) once, they skip the queue for future PRs.

It's far from simple, but there's a crowd of interested parties out there, and it would be a shame not to give that crowd some agency on the prioritisation of new contributions.

View this comment - View article - View full comments

Quoting: Leopard

Quoting: hardpenguinThis game has kinda fallen into obscurity since they moved to Epic 🤷

Lol, not correct at all.

Correct for many. Correct for me. I don't have an Epic account and never, ever will.

View this comment - View article - View full comments