this post was submitted on 28 Apr 2026
126 points (97.0% liked)
Technology
84222 readers
5618 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
If they get it right (opt in by default, respects privacy, appropriately sandboxed for security, clearly defined use case, etc.) then I can see how this could be useful.
But it's a big if.
I still don't understand how AI would be useful even if I owned the whole stack from the power station on up
People really have no idea what AI is. And in many cases those "AI features" are just machine learning. I'm currently running Ubuntu and I have several AI features running. One of them then is Handy.comptuer and it's a really nice push to talk voice recognition software. I'm dictating this post with it and it's awesome.
Another feature is image classifications integrated in Immich. It's nice to have your personal photo library and all the pictures that you have ever taken classified so that you can search for "dog" or "hamster" or "John". There are great text-to-speech programs out there. Everything is running locally and stuff like that is really useful to have in your OS. And it won't boil the planet, the energy requirements are quite low.
Text to speech already existed and doesn't need AI. Photos admittedly if I could have a sovereign version of Google photos that incorporates some of those features that was FOSS and offline, then ok maybe that's something useful "AI" can do.
I'm reasonably tech literate though and the idea of even exploring if such a thing is possible seems like more pain than it's even worth
It can barf out large outputs fast. As long as you review it and clean it up, it can get you 90% of the way there on a lot of things where you know what the final output should be basically. Need a quick bit of code for a common process? An email requesting a meeting? A picture of a dog burying a bone? A list of which people might be the suspect in the video? Then AI is great!
Need math? Accurate texts or image matching? Use a different tool.
I can envision use cases like wanting to be able to search through all your documents or photos when you can't clearly remember what you wrote (assuming it's properly implemented and well-secured, e.g. path-constrained time-limited read-only access), or local chatbot for brainstorming, or if you want to play around with agentic "AI" (automating small things on the local system that aren't easily scripted) then having one that is private and properly sandboxed would be helpful.
I can see that being useful but to me that's just like a sovereign/offline large language model basically.
We can call it AI but ultimately that's just a piece of software I could in theory buy once and forget about
yup. at the moment if you want a truly private, properly useful local AI agent you basically have to set one up correctly, manually, yourself. Having one come as part of the OS already largely set up for the user would be a massive W.
but yeah, If.
I use incus as a sandbox for claude. Previously I used podman, but podman in podman is a pain, so I gave it a full vm 😄