Whenever AI is mentioned lots of people in the Linux space immediately react negatively. Creators like TheLinuxExperiment on YouTube always feel the need to add a disclaimer that “some people think AI is problematic” or something along those lines if an AI topic is discussed. I get that AI has many problems but at the same time the potential it has is immense, especially as an assistant on personal computers (just look at what “Apple Intelligence” seems to be capable of.) Gnome and other desktops need to start working on integrating FOSS AI models so that we don’t become obsolete. Using an AI-less desktop may be akin to hand copying books after the printing press revolution. If you think of specific problems it is better to point them out and try think of solutions, not reject the technology as a whole.
TLDR: A lot of ludite sentiments around AI in Linux community.
I have yet to see any benefit to AI beyond the current browser UIs. The MS Paint image generation feature is neat for creating some quick clipart, if you mind the plagiarism, I guess.
Windows Recall shouldn’t be too hard to copy (it’s just OCR + CLIP on pediodic screenshots, after all) for those who want that sort of thing. Perhaps excluding private browser windows will be more of a challenge, especially on Wayland, but if the feature is built as an extension/plugin to the DE then I don’t think that’ll be impossible either.
Currently, the power and hardware requirements are too high for me to run anything useful locally, though. Even low-res image generation takes half a minute on my gaming GPU while burning a steady 180W of power.
The kind of text reformatting Apple has shown (selecting text and allowing a quick “make this paragraph more professional” in the context menu) takes forever on my hardware. Granted, it’s a few years old, but at 3 tokens per second I’m not exactly ready to install an AI addon yet.
I look forward to the Qualcomm and Apple advancements on this area, though. If the AI hype doesn’t die down, we may just see affordable and usable local AI in end user devices in a couple of years, and that’s pretty neat.
Hell, we may even see useful AI accelerator cards like that Coral thing or whatever it’s called, but with a usable amount of RAM. An upgradeable, replaceable AI accelerator could do a lot if AI stuff is going to be a hit in the future.
Like always, I expect Linux to be ahead of the curve when it comes to the technical ability (after all, Stable Diffusion ran on Linux long before Microsoft added it to Paint) but actually user friendly implementations will lag behind several years. Especially with the current direction of AI, basically advanced plagiarism and academic dishonesty machines, I don’t expect the free software community to embrace LLMs and other generative AI any time soon.