I can’t read the article because of a full screen Cookie Choices pop-up that I can’t dismiss. ☠️
Unlock origin has cookie banner filters. I didn’t have this problem, I assume that’s why.
Or you can import the “I don’t care about cookies” list: https://www.i-dont-care-about-cookies.eu/abp/
I like Linux a lot and do want to have it as my OS, but most of the games are either painfully slow or just instantly crash upon loading. No game has ran better on linux than on windows, so I’m stuck unfortunately.
When was the last time you tested this? My steam library runs great, not to mention everything else that runs on Lutris.
Nobara was the only distro that could even boot up all of my games without crashing. Some ran at a noticeably slower but playable rate while others ran at like 30 less FPS. Lowering the graphics to PS1 level hardly helped, and I just gave up trying to get my games to run like normal on Linux.
That’s really unusual. My experience has been the opposite on Linux Mint, most games run the same or better than when I was on windows. I had a little bit of trouble getting world of warcraft to work at first, but I was mostly done playing that anyway. I guess it’s all down to what games you play.
It’s not the “AI nightmare”, it’s a nightmare of capitalism, proprietary software and user-hostile behavior by a greedy, profit-extracting Big Tech corporation.
I find the nightmare getting a lot more noticeably bad with LLMs, though. That’s not just correlation.
I am basically a layman, i do music productions and in the past VSTs seemed to never work properly nor the authentication software that some us. Has it gotten better in the past few years, is there a specific one i should try? i have tried Ubuntu but nothing else to be fair. Also if i want to make a plex server on an old PC, what would people recommend? thanks to anyone who responds!
I used to use FL Studio, but hated using Windows. I got almost all features (including VSTs) to work in Ubuntu under Wine, but had a problem with WineASIO, which I seemed to require to use the USB sound card properly.
Because of that, I since changed to a DAW called REAPER which is built natively for Linux and works flawlessly and is very nice. There is a program called Yabridge to help run Windows VSTs. I even got more complicated plugins with authentication like Addictive Drums 2 to work using Wine no problem.
If you want a fully FOSS solution there is Ardour which is also great but a little less slick than Reaper IMO.
+1 for yabridge.
Bitwig is a great DAW (but not FOSS unfortunately). I run that on Manjaro, although Mint or Ubuntu are probably perfectly good choices too, if I had to guess.
I’d go for Jellyfin over Plex myself.
Yeah I don’t like the idea of having to login to their site, like I’m self hosting for a reason lol
Also if i want to make a plex server on an old PC, what would people recommend?
Any desktop PC built in the last 10 years should be fine. Just stick some hard drives in it :)
Intel processors are a good choice because their onboard graphics is quite good for video encoding/decoding. 6th gen or newer Intel Core processors (2015 or newer) would work well. They improved the H265 encoding/decoding a lot in 8th gen (2018) so that’d be even better. You can use something older but you’d need to also use a graphics card for video encoding/decoding, and it’d use more power.
Having said that, keep in mind that performance per watt always improves over time, meaning newer processors are more powerful even if they use the same power as the previous generation. A newer i3 will perform better than a very old i7. Using an very old, power-hungry system may end up more expensive in the long run compared to a newer mini PC.
I like using Proxmox. It lets you run multiple virtual machines on the system. VMs are good because you can easily snapshot them and revert back to an old snapshot in case of issues, and you can easily move the VM to a different system in the future. I use Unraid at home and really like it. It’s a bit simpler than Proxmox, but it costs money to use (Proxmox is free for personal use).
I wonder if some big AI heads will publish some “AI enhanced” Linux distros, that will also have other issues…
Almost definitely.
There is a command line program called tesseract that does image to text generation. It produces plaintext from a picture of text. I didn’t look into exactly how it works but iirc, image to text that’s actually good and accurate needs ai shenanigans.
It’s built by Google, but it’s open source, and is probably the best optical character recognition by far. It’s one pip/pipx installation away and I find it pretty useful on occasion. Same as WhisperAI by by OpenAI. Fully open source and one pip/pipx command away, probably close to the best audio transcription there is as well.
Not sure either count as AI, at least not AI chatbot kind of AI more like more simple algorithms, but they’re great in the sense it’s just another program but a very useful tool. Not some baked in copilot kind of deal
probably the best optical character recognition by far
I’ve actually just been working with OCR this week, trying to capture data off of the screen of a stupid proprietary Schneider device as that’s the only way to get at it.
Long story short Tesseract stinks at this task.
The Chinese designed PaddleOCR seems significantly superior as it runs a more modern neural net and requires a lot less preprocessing. I would class it as more of a “full service AI” and not just a simple recognition system like Tesseract, it can correct for skew and do its own normalization and thresholding internally while Tesseract wants a perfect boolean raster fed to it.
Unfortunately, the barrier to entry is a lot higher due to trying to understand their text vomit website and the fact that it seems prone to random segfaulting.
The algorithm is exactly the same as the chat bot, only the underlying data is different. Yes, they are all deep neural networks
I guess id be ok with an installable debian package for an end user controlled llama package with gui avatar interface overlay. Local learning data set storage plus ability to use API calls to injest info from other cloud based llm ai systems when the local dataset doesnt have a reliable answer.
I expect canonical to do it to Ubuntu.
IBM’s Watson Enhanced Red Hat
imagine if pop os does it first.
Untill one day Ubuntu will start incorporating AI in GNOME search bar
How much are you willing to bet this wont happen with Canonical’s Ubuntu?
How much are you willing to bet this wont happen with Canonical’s Ubuntu?
Doesn’t matter? Just use Debian… Or Mint. Or just uninstall the search bar.
Stallman was always right
Until big tech needs 2% more systems to squeeze out 2% more money…
Not happening. People are always afraid of new features. But when they try, see it’s convenient, and forget all about past reservations.
Its going to be the same with this. Ai is here and soon we won’t be able to imagine computers without it.
You say that like a specific technology is inevitable, but it never is. The general march of tech will continue on, but no one thing is ever guaranteed.
e.g. 20 years ago everyone needed custom browser toolbars and now it’s not even possible to add one on major browsers. We eliminated the need for browser features by cramming 99% of what we need into a handful of websites that are constantly refreshed.
e.g. 10 years ago blockchain was surging and today it still doesn’t have a usable application. Turns out spreadsheets don’t really need to be distributed.
Machine learning is just an algorithm nobody understands. If I needed something to give me wrong answers to questions I’ll ask my dog.
I don’t think I need to defend the usefulness of ai and compare it to browser toolbars…
AI is here and pc needs it. People have been dreaming of it since before the were computers. Even for the most basic features of it.
“The year of Linux on the Desktop” is in the article. This again? Been reading this for decades and it’s still not true.
Linux is close, but has some core flaws that will forever keep it out of mainstream acceptance by your average user.
Linux is close, but has some core flaws that will forever keep it out of mainstream acceptance by your average user.
It has nothing to do with any flaws within Linux itself. The problem is and has always been that it’s nearly impossible to buy a PC with any flavour of Linux pre-installed. Until that changes, Linux (on home user desktops) will never gain mainstream acceptance.
I agree. Most people won’t switch to Linux because they have never used it and think they’ll have to relearn computers from scratch.
Didn’t HP sell some fancy shmancy laptops that came with Ubuntu or some flavor of it? Think it was for developers but I thought that was the closest we gotten to commercially selling Linux based machines.
P.S. I could be wrong about this but I am sure this happened.
HP sold he DevOne, it had PopOS on it. Dell sells an XPS developer machine that has Ubuntu pre installed. System76, Entroware, and Tuxedo computers have all been selling Linux hardware for a long time. So there are viable commercial options. I wish the DevOne were going to get refreshed, it looks like a nice machine but alas, I don’t think it will.
It’s possible they did. I think Dell briefly discussed it as an option, before using it as leverage to get cheaper Windows licenses from Microsoft. The EEE PC also shipped with its own Linux distro and appropriate hardware drivers.
This was why I said “nearly impossible” :)
Maybe we should have like a yearly event for this. Like a holiday. International Linux Year Day.
I think there is some sort of conference. The key would be to convince all the Linux users to stop telling us about it the rest of the time.
Linux uses and vegans have the same “I’m better than you” energy.
AI has people questioning Windows use Car systems ratting drivers out has people questioning car use
Not the way I expected to reach some of my desired ends but I’ll take it. 🤔
Ehh, I have a different vision here - AI is useful, it’s just going down the hypermonetisation path at the moment. It’s not great because your data is being scraped and used to fuel paywalled content - that is largely why most folks object.
It’s, also, badly implemented, and is draining a lot of system resource when plugged into an OS for little more than a showy web search.
Eventually, after a suitable lag, we’ll see Linux AI as the AI we always wanted. A local, reasonable resource intense, option.
The real game changer will be a shift towards custom hardware for AIs (they’re just huge probability models with a lot of repetitive similar calculations). At the moment, we use GPUs as they’re the best option for these calculations. As the specialist hardware is developed, and gets cheaper, we’ll see more local models and thus more Linux AI goodness.
There is no year of Linux desktop, it just keeps trucking and growing
It doesn’t really seem to be growing much, to be honest.
It’s dominating the smartphone market via Android, though.
Can we get a hatchback model? I’d much prefer it to a truck. And is there a setting so it doesn’t grow? I want to stay city-friendly.
Imagine the horror of living in a world where all vehicles slowly expand and have to be cut down to manageable size annually until eventually the car is just too big a la American full size SUVs at EOL.
It’s not AI that is the problem, it’s half baked insecure data harvesting products pushed by big corporations that are the problem.
The biggest joke is that the LLM in Windows is running locally, it uses your hardware and not some big external server farm. But you can bet your ass that they still use it to data harvest the shit out of you.
That’s a pretty big joke, but I think the bigger joke is calling LLMs AI. We taught linear algebra to talk real pretty and now corps want to use it to completely subsume our lives.
I think the bigger joke is calling LLMs AI
I have to disagree.
Frankly, LLMs (which are based on neural networks) seem a Hell of a lot closer to how actual brains work than “classical AI” (which basically boils down to a gigantic pile of
ifstatements) does.I guess I could agree that LLMs are undeserving of the term “AI”, but only in the sense that nothing we’ve made so far is deserving of it.
seem a Hell of a lot closer
“seem” is the critical word there. Interacting with an LLM they do seem to be pretty clever.
I’m not talking about interacting with it. I’m talking about how it’s implemented, from my perspective as a computer scientist.
Let me say it more concretely: if even shitty expert systems, which are literally just flowcharts flowchart implemented in procedural code, are considered “AI” – and historically speaking, they are – then the bar is really fucking low. LLMs, which at least make an effort to kinda resemble the structure of biological intelligence, are certainly way, way above it.
I’m actually sad that the state of AI deserves the hate it gets. Neural networks are so sick, just going through the example of detecting a diagonal on a 2x2 grid was like magic to me. And they made me second guess simulation theory for quite a while lmao
Tangentially, blockchain was a similar phenomenon for me. Or at least trust networks. One idea was to just throw away Certificate Authorities. Basically federate all the things, and this was before we knew about the fediverse. It gets all the hate because of crypto, but it’s cool tech. The CA thing would probably lead to a bad place too, though.
Do you, as a computer scientist, get a little woody when you think about “AI” though?
Oh I agree. I typically put “AI” in quotation marks when using that term regarding LLMs, because to me they simply are not intelligent in anyway. In my mind an AI would need an actual level of consciousness of sorts, the ability to form actual thoughts and learn things freely based on whatever senses it has. But AI is a term that’s good for marketing as well as fear mongering, which we see a lot of in current news cycles and on social media. The problem is that most people do not even understand the basic principles of how LLMs work, which lead to a lot of misconceptions about its uses & misuses and what we should do about it. Weirdly enough this makes LLMs both completely overhyped as a product and completely stigmatized as some nefarious tool as well. But I guess it fits into our today’s societies that kinda seem to have lost all nuance and reason.
To me this is even worse though. They’re using your electricity and CPU cycles to grab the data they want which lowers their bandwidth bills.
It happening “locally” while still sending all the metadata home is just a slap in the face.
Exactly. And if I use or even pay for an external LLM service then that’s also my decision. But they force this scheme onto every user, whether they want it or not. It’s like the worst out of all possible scenarios.
That is an accurate description of AI in common usage even if it isn’t an inherent aspect of AI.
Right, but AI is not the only way they’re doing the data collection.
You wrote AI twice.
I think it’s important to note that Linux can be a way to avoid AI, but doesn’t have to be. If you flip the headline around it almost implies that people who do want AI would be missing out by using Linux, but that’s not true at all: instead, the reality is that Linux is still better for them, too, because you could install all the same kind of functionality if you wanted, but it would be wholly under your control, not Microsoft’s.
That sounds very cool. I’m totally ignorant of the hardware requirements. What sort of minimum setup would such an install take?
Will my dell latitude from 2006 work?
It really depends on what model you want to run and how much training is bundled with it. You can pretty much run any model if you have enough disk space but of course GPU + VRAM is preferred for a ChatGPT like fast response. Otherwise, running on an older CPU and RAM is going to be noticeably slower, especially with complex models with a lot of training data to trawl through.
There are some pretty lite models out there but the responses will be more barebones and probably seem ‘less informed’.
Give GPT4All a try for your first time. It makes install, configuration and usage point-and-click while being fairly straight forward. For the presented/featured models, it presents a small summary and VRAM recommended, though there are many, many other models available from inside the UI.
Self hosted AI seems like an intriguing option for those capable of running it. Naturally this will always be more complex than paying someone else to host it for you but it seems like that’s that only way if you care about privacy
Check out Jan AI. It’s open source and extremely easy to install and run. I run it locally on a 2017 laptop without a dedicated GPU and it works, just takes longer to generate responses compared to something like ChatGPT.
