Yes, it works out to a ton of power and money, but on the other hand, 2x the computation could be like a few percent better in results. so it’s often a thing of orders of magnitude, because that’s what is needed for a sufficiently noticeable difference in use.

basing things on theoretical tops is also not particularly equivalent to performance in actual use, it just gives a very general idea of a perfect workload.

At the datacenter scale Gaudi 3 was pretty good, at least when it came out.

Intel GPU support?

ZLUDA previously supported Intel GPUs, but not currently. It is possible to revive the Intel backend. The development team is focusing on high‑quality AMD GPU support and welcomes contributions.

Anyways, no actual AI company is going to buy $100M of AI cards just to run all of their software through an unfinished community made translation layer, no matter how good it becomes.

OneAPI is decent, but apparently usually fairly cumbersome to work with and people prefer to write software in cuda as it’s the industry standard (and the standard in academia)

Intel’s Gaudi 3 datacenter GPU from late 2024 advertises about 1800 tops in fp8, at 3.1 tops/w. Google’s mid 2025 TPU v7 advertises 4600 tops fp8, at 4.7 tops/w. Which is a difference, but not that dramatic of one. The reason it is so small is that GPUs are basically TPUs already; almost as much die space as is allocated to actual shader units is allocated to matrix accelerators. I have heard anecdotally.

It’s not even a pivot. They’ve been focusing on AI already. I’m sure they want it to seem like a pivot (and build up hype); the times before apparently just having the hardware and software wasn’t enough. nobody cared when the gaudi cards came out, nobody uses sycl or onednn, etc

Freecad is pretty good, but unfortunately there’s no foss cad software that’s better. If you don’t care about foss, I would recommend onshape if you’re fine with the “public by default” thing, else fusion360.

For art, blender is great. Plasticity seems neat too, it’s a more traditional software licensing model (pay per version I think, not cheap not insanely expensive)

It says in the article

15% of these subreddits contained content likely posted by bots or corporate trolls specifically designed to promote companies or organizations.

The title is just misinfo I guess

i do want to point out how hard it is to even find out about the views of these people, if you just look up the names of the projects and aren’t specifically looking for this information there’s no way you’ll find anything about it

even looking up the name of David Heinemeier Hansson, the more vocally bad of these, i had to go to the 5th link to find anything even vaguely mentioning his views

Samsung, Xiaomi and OnePlus have removed the option of bootloader unlocking on all of their devices.

Got me worried (bc i have a newish oneplus phone) but apparently OnePlus is only doing that in China for now. Still not a good sign for the future…

I’ve used some slightly weird hardware but haven’t experienced anything of what you described. Across the whole range from the lab server with 3 3090s and 500gb of ram to my $40 Chromebook I got on ebay

Not a petition exactly but I found https://fightchatcontrol.eu/

There might be others, I didn’t look that long and I’m not in the EU so idk

The linked petition says:

The European Parliament MPs must vote AGAINST Chat Control legislation during EPlenary vote in June-August 2021.

It’s not even up to date. And it only has 1.8k signatures.

Not going to surprise anyone but Windows Mixed Reality VR headsets aren’t great on Linux, at least with controllers

Although that is improving!

Arch also can absolutely be installed just as quickly as any other distro if you use the archinstall script. I used it recently to install KDE plasma onto a Chromebook from 2017 and everything worked exactly as expected, I haven’t had any issues with stability so far. Can absolutely be done in under half an hour. It ofc doesn’t come with the advantage of understanding exactly how your system is set up, like you would if you did it yourself.

The last time I did that (slightly different setup with xfce) though I broke it somehow and ended up with if freezing often when booting, although I’m still not sure if that was a hardware problem or not, but it doesn’t seem to be happening anymore. I also broke something with the audio jack somehow around then during an update, but chromebooks have weird audio drivers and you need to use this script maintained by (afaik) one person in their spare time. Anyways I would expect a framework laptop to handle it better as it’s newer and more common hardware.

This is slightly misleading. Even if you can’t achieve “agi” (a barely defined term anyways) it doesn’t mean AI is a dead end.

I actually had a problem where on Chrome, I would be signed out of my google account every time I restart my computer, while on Firefox, everything works normally. I use Firefox now lol.

Ram is slower than GPU VRAM, but that extreme slowdown is due to the bottleneck of the pcie bus that the data has to go through to get to the GPU.

Yeah, you also have to deal with the latency with the cloud, which is a big problem for a lot of possible applications

well, i think a lot of these cpus come with a dedicated npu, idk if it would be more efficient than the tensor cores on an nvidia gpu for example though

edit: whatever npu they put in does have the advantage of being able to access your full cpu ram though, so I could see it might be kinda useful for things other than custom zoom background effects

Yeah, I think it could be useful for CAD or 3D art (with proper software) but I can’t think of many other jobs where it would be all that helpful.