Latest 30 Comments

Quoting: PlayingOnLinuxphoneThere is no real exponential trend. I did the math some hours ago and there was a linear trend before Steam Deck release, than an accelerated trend between Steam Deck Release and Win10 EOL and a 5.5 times increase after Win10 EOL. Steam Deck OLED did not change the linear trend. November and December 2025 had exactly the same 5.5 times increase as the whole 2026 until April. January and February were Chinese New Year, March somehow bad data, so these should be ignored.

If it would be a real exponential curve, March data would be real and April even above that. So it would be better to use 3 linear lines, clamped between those relevant dates.

PS: If it is linear we will see on Windows 10 extended support end of life. It should be around 6%. It should be far above 6% if it is exponential.

Here's a visual aid.

As you say, it's all pretty flat before the Deck, then we get added a pretty linear Deck signal, and then we get a pretty linear desktop signal as the Deck flattens out.

I've found the 5-month moving average to be quite good at separating out the signal from the noise.


There's no real justification for splatting the data onto an exponential curve, although we get someone demanding that it's done pretty much every month so I'm not surprised that Liam's given it to them to make them stop. One could splat it into a logarithmic, sinusoidal or tangent curve just as easily.

View this comment - View article - View full comments - Report this comment

Quoting: PlayingOnLinuxphone

Quoting: mattaraxiaIt's really wild. 10% by ~2030 seems just about inevitable now.

With the current speed, 9.5% is reached by end of 2027 and considering Steam hardware it should even be above 10% next year (if nothing slows down).

I'm as excited as every other Linux enthusiast but that's very unlikely.
We're probably not going to see such big jumps frequently.

View this comment - View article - View full comments - Report this comment

Quoting: mattaraxiaIt's really wild. 10% by ~2030 seems just about inevitable now.

With the current speed, 9.5% is reached by end of 2027 and considering Steam hardware it should even be above 10% next year (if nothing slows down).

View this comment - View article - View full comments - Report this comment

Quoting: KimyrielleIt only takes "AI" somewhere in the first line to make these people go reach for the pitchforks. Objectivity and reason has left this discussion a while ago.

...which is a good thing because we've learned very quickly that giving an inch has them take several hundred miles. There's nothing wrong with rejecting satan technology in all of its forms and not just the most cataclysmicly evil version of it, in fact I'd argue that's exactly what you SHOULD do.

View this comment - View article - View full comments - Report this comment

There is no real exponential trend. I did the math some hours ago and there was a linear trend before Steam Deck release, than an accelerated trend between Steam Deck Release and Win10 EOL and a 5.5 times increase after Win10 EOL. Steam Deck OLED did not change the linear trend. November and December 2025 had exactly the same 5.5 times increase as the whole 2026 until April. January and February were Chinese New Year, March somehow bad data, so these should be ignored.

If it would be a real exponential curve, March data would be real and April even above that. So it would be better to use 3 linear lines, clamped between those relevant dates.

PS: If it is linear we will see on Windows 10 extended support end of life. It should be around 6%. It should be far above 6% if it is exponential.

View this comment - View article - View full comments - Report this comment

It's really wild. 10% by ~2030 seems just about inevitable now.

Even as someone who's used Linux for literally decades now, I'd have never thought this would be such a thing.

View this comment - View article - View full comments - Report this comment

Hopefully the dip brought the numbers closer to reality.
That's still a 1.5% increase since the Windows 10 EoL. Roughly 2% over the last year. Despite the Steam Deck being out of stock for many months now. The future of Linux gaming is looking bright, folks.

View this comment - View article - View full comments - Report this comment

0 64 bit? 😆

View this comment - View article - View full comments - Report this comment

​This wine version in my case install mono 11.1



​😀

View this comment - View article - View full comments - Report this comment

FNA still delivering good stuff, thanks flibitijibibo et al.!

View this comment - View article - View full comments - Report this comment

Quoting: GerarderloperBet Sony spending a fair bit of money trying to patch all these holes out.
Bit of a scenario of them shooting themselves in the foot because having a way to install Linux on these consoles will make them MUCH more popular and people will naturally buy the games for them anyway.

BUT they won't see it that way; gotta waste money and ruin everyone's fun, the corporate motto!

I've read they are now putting more DRM into consoles; wonder how long until Age Verification comes and you are required to have Face and ID scans to even turn the things on.... what a world!

I do wish Steam Machines were going to be affordable alternatives, but it seems that won't be happening for quite some time! Onward and Downward into the fiery pits we go!!!! :)

we used to have a lot of companies making their own proprietary operating system for phones with their own proprietary app stores, but they failed and now all companies use android, except for apple with iOS.

we used to have a lot of proprietary formats for pcs, nowadays we have ibm open pc and apple , basically.

I think the same will happen for consoles, a lot of companies making windows or linux devices, and sony will be forced to chose between the 2 or be kicked out of the market.
I think Nintendo might survive as "the apple of gaming consoles", their exclusive brands are too strong to simply disappear, the same cant be said about sony games, at least in terms of sales they arent such a big deal as fanboys think.

View this comment - View article - View full comments - Report this comment

It even has the iconic Tattooine track from SW Episode I - Racer!

View this comment - View article - View full comments - Report this comment

Glad to see you enjoyed it!! I helped with this release

View this comment - View article - View full comments - Report this comment

Quoting: Craggles086So not all the Studios that got bought up by Epic had half their workers sacked.

Nice to know.

Sorry, still feeling a bit bitter about some of the Epic purchases.

Yeah it sucks what they did to Fall Guys

View this comment - View article - View full comments - Report this comment

Quoting: NezchanAs much as the original Subnautica was one of my favourite games of all time, I'm gonna sit this one out and see what the reviews say.

I've heard from people that claim to have had access to it (yes, have a bag of salt handy) that the core game is very good, but that a lot of content still has to be added. Which sounds very positive, but is still a reason to not be in a hurry.

Quoting: Nezchan... Krafton's declaration they're now an "AI First" company a few months ago does not fill me with confidence.

The Krafton CEO couldn't even resist using ChatGPT to hatch his plan to not pay out the bonus to the Unknown Worlds founders (and luckily hide it badly). Personally I'm more worried about the chaos than the AI use. On the other hand, Krafton has been behaving terribly for a while and I'd have a good chuckle if they'd be forced to have to pay out an extra 250M. That alone might be worth 30 dollars. 😄

View this comment - View article - View full comments - Report this comment

Damn I didn't think this was going to be quite this big of a deal. I was happy to just say 'Look it's funding, we can take money from anyone, doesn't mean the company will have any influence', just like we do every time a dodgy company gives money to Blender Foundation.

I would have thought if the Blender community was happy to accept money from Adobe that they'd be happy to accept money from anyone.

Don't get me wrong, I hate OpenAI and Anthropic with the fiery passion of a thousand stars burning brightly, but sometimes it's hard for me to tell what outrage is just me being me and what outrage is actually universal. Seems this was a universal outrage. It really says something out of all the companies Blender Foundation has taken money from, THIS is the one which people couldn't stomach. Jeez, just shows how absolutely poisonous 'GenAI' is right now in public opinion.

View this comment - View article - View full comments - Report this comment

Quoting: Renzatic Gear

Quoting: EWGIf a higher price tags equates to higher quality builds and wages for those making it, then I'd be happy to eat a few more peanut butter & jelly sandwiches and wait another pay cheque if necessary to buy something.

You make it sound like having more PB&J sandwiches is a bad thing.

I’ll tell you right now that it is not.

Fair. I do often either way. It's one of those wonderful cravings and allows for a little creativity. Triple decker with cinnamon sprinkled on the bread? Crushed/sliced/chopped nuts? Alternative to peanut butter like almond butter? Frrsh fruit? Dried fruit? What bread is available today?
🤤

However, it's also really nice to have a fat veggie heavy sandwich or cheezy panini. Something hot, grilled, with more vegetables and some protein.

View this comment - View article - View full comments - Report this comment

I'm whiling away the hours on my Deck with Pictopix and Nordic Ashes. They've both served me well in scratching my Picross and Vampire Survivors itch.

View this comment - View article - View full comments - Report this comment

So wait, I don't get it. Are they just updating the ISO to ship with the latest kernel that's regularly available to update to? So, basically, during install it'll have broader and better support for hardware depending on the kernel? Is that it?

View this comment - View article - View full comments - Report this comment

As I understand it, some games (such as Arc Raiders?) will group players with similar skill levels, which seems like a good way to limit frustration. Rather than banning cheaters, or someone who *looks* like a cheater, just make them play with other cheaters. If it's a false positive, it may be annoying, but way better than being banned?

Edit: I guess this is a bit off-topic, since the article is about DRM and online checks. But the conversation has drifted a bit into anti-cheat...

View this comment - View article - View full comments - Report this comment

Quoting: LinuxwarperEventually even Microsoft will support Linux

Soon after the Steam Deck came out, Microsoft was actively promoting their published games that worked on it. I think that was when they were under scrutiny for wanting to buy Activision, so that may have been part of the motivation...

View this comment - View article - View full comments - Report this comment

That behavior agrees with what I wrote below the previous article: Anthropic just wants to keep Blender healthy. They have an interest for plugins, not to destroy Blender while Blender requires money to make Blender better than the proprietary concurrences. I am one of the first who criticize things related to so called AI-companies, but this action was totally fine.

However, it is good that they address the complains in a good way. People who (still) complain do it for an uninformed opinion in this specific case. I am Blender user myself and don't want to get unethical neuronal network tools implemented by default. I am also fine with keeping the door open for implementing such tools in future, if it is done in a fully ethical and transparent free software way.

Hint: ethical and transparent includes:
- All trainings data is public available and published under free licenses as GPL or CC.
- All licenses of trainings data are respected (as naming sources/creators).
- Tools running locally, offline and do not collect any data.
- Tools are not implemented if there is no real benefit (like replacing totally fine tools) and also not to produce art-slop (can be done with plugins anyway).

View this comment - View article - View full comments - Report this comment

Business isnt going too great for Epic while Linux platform is growing. It was a matter of time. Ask yourself, if Epic's tactics worked, exclusivity and free games, and Linux didnt reach 4℅ market share, would they have supported EAC for Deck/SteamOS/Linux? I highly doubt it. Eventually even Microsoft will support Linux. When that time comes please dont act shocked, just look at the number "oh its market share of 10%".

Profits, or loss thereof, will often take precedent over lock-in measures. If you can no longer make DirectX work for you, stifling Linux ports AND compatibility, you take next best thing; be in control of software stack. If Microsoft doesnt provide DirectX on Linux, when market share is 10%+, it will push devs more towards adoption of competing software. If you're more familiar with Vulkan, and it works for your projects, why bother with Direct3D on Windows as well when you can reuse Vulkan code.

Fortnite and Epic's games will be negatively impacted if a scenario of 10% market share of Linux gamers happens and they dont support the platform. Game is on decline and with alot gamers moved over to Linux, a unsupported platform, the pool of players they can attract will be reduced and for a "dying" game thats bad. Moreover if the players who like Fortnite, but cant play it because of support, find alternative, a similar game that supports Linux, then the growth the Linux players give to said new game will boost its presence among gamers (social media). Social media will bleed to Windows gamers who play Fortnite and they may stop playing Fortnite and try the new hot thing. Then Epic/Fortnite player base will bleed (more).

View this comment - View article - View full comments - Report this comment

@LoudTechie: I don't discuss such meta nonsense with you. I don't even understand why you are arguing here something that is clearly out of context. So I did not make a mistake, you are just looking for a failure that does not exist in the underlying context. All neuronal networks are just bad in this job, no matter if LLM in specific or not. And "AI" that is no neuronal network (even if it is a smart algorithm) is no AI or not relevant to prevent cheating, because the term is meant to be connected as "artificial animal/human being" like NPCs (behavior simulation) or neuronal networks (algorithmic brain simulation).

Quoting: sarmadyou need math + patterns.

Patterns are nothing else than pure math. For simple things neuronal networks can be better indeed. But with growing complexity the failure rate is also growing significant. And games are just too complex to work well (you don't want false positives). So it is better to stay at simple rules and some games proofed that these are good enough for a cheating free experience. Your example is something I am pretty sure I could realize in my own game with classic algorithms. Cheaters that do such obvious things are even more easily to filter using kill-death-rate with a value of 10+:1 (because they are stupid enough not to stay in the range of best legit players).

Moreover classic data analysis can also handle complex data structures (and especially patterns) easily. That's the way search algorithms, social media profiles and a lot of other stuff was done before neuronal networks became an useful thing.

But let's say there is a way using neuronal networks in a good way to prevent or ban cheaters while the failure rate is 0.0%. Then it becomes accepted and lazy companies build in these tools in a lazy way and so the failure rate is not 0.0% any longer. So even in the theoretical best case scenario it is better to not open the door. I don't want to have the fear that AI can always take away my progress by auto-banning when I was playing totally fair (which I always do, because cheating takes away the fun of challenges).

And again, why you would want to implement such tool, if it brings no benefit over classic algorithms, but potential risks? The only point is "costs less money".

View this comment - View article - View full comments - Report this comment

KKnD was my favourite RTS back then. Followed by Dark Reign. I loved the KKnD 2 demo as well, but it crashed a lot on my computer so I didn't play it to death like the first one.

Anyway, this should be awesome. I've never had these working properly on proton. Always a tiny window that's about 1/24 size of my screen and I can't seem to change it.

View this comment - View article - View full comments - Report this comment

Stardew. I'm one of those numbers.

Also Arkham City. My daughter just started watching the Bruce Timm animated series and I thought she'd like watching me play it. She said no, she wants to try playing games herself, oh well she's learning and having fun.

View this comment - View article - View full comments - Report this comment

Quoting: PlayingOnLinuxphone

Quoting: sarmadFor example, it should be able to detect that you are aiming precisely at a target behind a wall, or other similar things that are impossible without cheating.

That is a good example why classic algorithms are better than LLMs which are basically algorithmic lossy archive files. It is something like MP3 for all kind of data where the timestamp is the input (prompt, randomization number and so on). The important part is "lossy", because it loses information on the training process. This causes fail predictions as not being able to do correct maths. I just asked CGPT on duck.ai "What is the math result of 13/73²?" and the result was "13 / 5329 ≈ 0.002438 (rounded to 6 significant figures).", while the real numbers are 0.00243948... this was not even a rounding failure.

If we take this approach to your gaming situation, there are precise position parameters and time frames. The server can calculate precise if such an action could be possible - easy triangulating math. Why would you want to predict if it was possible or not if you just can calculate it? And AI probably does not even go the math route, because another one is "easier" (which does not mean more correct).

AI in games is good for things like animation prediction, can save resources while being much more natural than classic systems. But there is no benefit in using it for anti-cheat (except it may costs less at development, but it can damage the companies reputation in return).

Yes, you should not depend on AI for doing math at all, but math alone cannot easily determine cheating; you need math + patterns. In other words, you need math to convert the data into a form that can then be consumed by AI. For example, assume you are aiming at someone and moving sideways at the same time, then you go behind an obstacle and you continue aiming at the same spot which remains precise then you continue moving until you are away from the obstacle and you shoot immediately and get a perfect shot. This is not cheating because you had an initial line of site that you used to aim, but doing math alone will determine that you are cheating because you are aiming behind a wall. On the other hand, if the subject moves while you are behind the wall and you successfully follow its movement while still being behind the wall then that's likely cheating. I say likely because this can very well be just a coincidence or it could be experience like knowing the direction of the subject and predicting it's movement. But if this happens all the time with high precision then the person is more likely cheating. So, the proper solution is to use math to convert the data into high level data, then feed that to AI for a smart cheat detection.

View this comment - View article - View full comments - Report this comment

I wonder how much of that is backlash and how much is being conscious of an issue. A few years back I would not have had those doubts.

At this point, "money is money", and Blender could certainly use the boost in that regard. But it's very easy to slip towards justifications like "see? We might be destroying everything, but the money is used in a good cause!". Time will tell.

View this comment - View article - View full comments - Report this comment

Quoting: Ehvis

Quoting: hardpenguinWhat kind of silly moron cheats in Rocket League of all games???

Look at the screenshots in the article. That right one is highly suspicious! 😄

Honestly, for a studio now owned by Epic, it still has a pretty decent implementation. Still able to turn it off and do your own thing outside of the official servers is something that most devs would never do. Especially if they rely on skins for monetization.

So not all the Studios that got bought up by Epic had half their workers sacked.

Nice to know.

Sorry, still feeling a bit bitter about some of the Epic purchases.

View this comment - View article - View full comments - Report this comment

Just got back in to Last Epoch with Season 4.
Bought Company of Heroes 3 in the Steam Christmas sale, so that one have been fun. Had over 2000 hours in CoH2, I guess time will tell if the CoH3 will keep me going as long.
And of course World of Warships.

View this comment - View article - View full comments - Report this comment