Not going to read most of this paper, because it reads like a freshman thesis, and it fundamentally oversells or misunderstands the existing limits on AI.
In closing, I consider the limits to these limits as AI gradually, but relentlessly, becomes ever-more capable.
The AI technofacists building these systems have explicitly said they've hit a wall. They're having to invest in their own power plants just to run these models. They have scores of racks of GPUs, so they're dependent upon the silicon market. AI isn't becoming "ever more capable," it's merely pushing the limits of what they have left.
And all the while, these projects are still propped up almost entirely by venture capital. They're an answer to a problem nobody is having.
Put another way, if the leaders of the AI companies are right in their predictions, and we do build AGI in the short- to medium-term, will these limits be able to withstand such remarkable progress?
Again, the leaders are doing their damnedest to convince investors that this stuff will pay off one day. The reality is that they have yet to do anything close to that, and investors are going to get tired of pumping money into something that doesn't return on that investment.
AI is not some panacea that will magically make ultracapitalists more wealthy, and the sooner they realize that, the sooner we can all move on—like we did with the Metaverse and blockchain.
This tactic is so old, but it weaponizes the annoying ubiquity of capchas. People just want to get to where they're going, so they click the squares and do the dance to get past the seemingly arbitrary barriers.
This technique shows up on !cybersecurity@sh.itjust.works every few weeks as the initial attack vector for some new RAT.
Maybe we just don't like AI, and billionaire fuckwads just need to suck it up and move on, like how Zuck bet the farm that the Metaverse was going to be the next hot thing and now has to pretend he didn't spend gobs of money on it.
But no, it's the cancel-culture mob. 😂 JFC, what a take.
I understand that, but what they do with user data is governed by their Privacy Policy, which again, is unchanged. The ads they buy are the same Sponsored ones that show up on blank tabs—the ones that have been there since before they made that change.
They made the change to the ToS, because a California law expanded the definition of "sale/sell" beyond what most people understand the word to mean. There's enough vaguery in the wording of the law that the way Firefox works, it could land Mozilla in hot water the way the ToS were worded. It's stupid, Mozilla did probably the worst job possible communicating why they were making the change, and the internet freaked out.
I'm not saying you shouldn't leave. That's up to you. I've been running LibreWolf since then, because a company that has $37M in investments and pays their CEO $8M should have the means to have a decent marketing team, one that could warn them it would be stupid to abruptly cut out a section on selling user data. However, it's simply not true that they've suddenly joined ranks with the likes of Google.
Again, do what you want, but I hope people do it because they've been informed about the facts, not because the internet brought out the pitchforks again.
It's what I have done. They'll work slower, but you'll get a sense of what they can do, how hard it is to do things, etc.
When you're ready, I think all but Bazzite have Live ISO options, so you can see what it's like on bare metal. When you're satisfied, install your favorite!
First, gaming distros are vanilla distros with opinionated tweaks and additions to support the hobby of gaming. It might be as simple as having Steam pre-installed to as complex as having unique kernels or custom package repos maintained by the distro maintainers.
But that doesn't mean vanilla is always the best choice, because not everybody wants to spend time optimizing everything. Some distros even have easy setup scripts for otherwise complex installations (like for Davinci Resolve). Don't feel like you need to pick vanilla to be a "true user."
Some easy to set up Distros for gaming that are ready ootb:
Bazzite: Fedora Atomic, practically bulletproof, just works. Downsides are that adding new packages is not the same as other distros, and there's a learning curve to it beyond flatpaks. Some software can't be installed at all if it doesn't come as an RPM or AppImage (Private Internet Access's VPN client, for example).
CachyOS: Arch with an optimized kernel and optimized packages. Comes with some easy-install scripts. Tool to easily select different kernels and schedulers. Currently another very popular choice. Like the above, this just works. There's some debate about how significant the optimizations really are, but they're there nonetheless.
Nobara: Traditional Fedora. Like Bazzite, just works. Has a custom update manager that acts as a GUI wrapper for your usual cli tools. Maintained by GloriousEggroll, a widely respected user that maintains the GE versions of Proton.
PikaOS: Debian (not Ubuntu). Combines the philosophies of Nobara and CachyOS and puts them atop Debian. Better setup scripts than even CachyOS, a more user friendly update tool than Nobara's, and has the same kernel selection and scheduler tools as CachyOS, plus the same package optimizations. Don't let the fact that it's Debian underneath fool you. This has the latest kernel and drivers.
I would try all of those in a VM and see what you like about them. They're all unique and worth a look.
ETA: all of these have Nvidia versions, so all of them should work with your card.
That's fine. Do what you want. I'm not here to judge your choices, just point out that Mozilla only fucked up the communication, not the policy itself.
Right? I just do my best to ignore the bot and only enter queries any first year CS student would know. The rest comes from my memory and a few bookmarks I have saved.
Not going to read most of this paper, because it reads like a freshman thesis, and it fundamentally oversells or misunderstands the existing limits on AI.
The AI technofacists building these systems have explicitly said they've hit a wall. They're having to invest in their own power plants just to run these models. They have scores of racks of GPUs, so they're dependent upon the silicon market. AI isn't becoming "ever more capable," it's merely pushing the limits of what they have left.
And all the while, these projects are still propped up almost entirely by venture capital. They're an answer to a problem nobody is having.
Again, the leaders are doing their damnedest to convince investors that this stuff will pay off one day. The reality is that they have yet to do anything close to that, and investors are going to get tired of pumping money into something that doesn't return on that investment.
AI is not some panacea that will magically make ultracapitalists more wealthy, and the sooner they realize that, the sooner we can all move on—like we did with the Metaverse and blockchain.