Windows 11 will soon be able to describe images on your screen using AI — and it'll all be done locally
Windows 11 will soon be able to describe images on your screen using AI — and it'll all be done locally
Just had a chat with a gamer friend who was ditching Windows for good and switching to https://bazzite.gg/. I'll probably do so as well. Apparently I've been needlessly scared of Linux and I had missed that SteamDeck encourages devs to make games work for Linux.
If you aren't playing a game that has kernel mode anti cheat, like League of Legends, then your gaming experience will probably be the same if not better.
Bazzite is great.
Linux has come SO far. I first tried it about a decade ago, wrecked my install, and came crawling back to Windows. A few months back, I installed Ubuntu (tease all you want, peanut gallery, it's better than Windows) on my laptop and it's so much better. It just works great, I have my system customized to how I want it to work, and the poor CPU fan is no longer blasting even when the system is idle. My battery life also gained almost an hour.
I've been using Ubuntu since 7.04, it's all Linux underneath just with a slightly different look
An image is worth a thousand words. How is reading a text describing what is on the screen going to be better than just looking at the screen yourself, something you'll need to do to read the description anyway? Aside from accessibility for the blind, the practicality such a technology is questionable.
The motivation behind this is obviously to facilitate the collection and reporting user profiling data. Accessibility for the blind is only a side effect. Tech companies have been doing it with automated audio transcriptions for years already, now they're after what you look at on your screen.
Blind people exist
Read the whole post. I acknowledged them already and am expressing my doubts over the true motivations that drive Microsoft to force a tech like this upon all their users and express my concerns over the real use they will make of this technology.
Don't you try to change the meaning of my post just so you can have a cause to white knight over. This isn't Reddit.
Sometimes things are unlabeled, and this is for that
this is 100% right, you don’t need an AI to describe something you’re already looking at. This is an absurd feature (again aside from the accessibility portion but that’s not what this is).
Why would I, or anyone want this?
The blind or vision impaired would appreciate this. Past that, hard pass.
I use a screen reader due to being blind, and this would be incredibly helpful. Mind you, I sure as hell wouldn't use Windows in order to get it. But if I had the option to do this on Linux, I would totally go for it.
My wife, when she's not wearing her prosthetic corneal lenses.
Never been happier that all my computers run Windows 10 or Linux. Windows 11 is dead to me, and if anything happens to accidentally get it installed somehow, it's going to be replaced with Linux going forward.
Go and fuck yourselves.
A small conversation with Copilot makes my laptop choke on RAM, and that stuff is processed on cloud. No way I will allow Microsoft to run an AI locally.
I never cease to be impressed by the continual enshittification of Microsoft products, and the bar was pretty low to begin with.
Why.
Do I look as if I need that?
Multimodal models has a lot of potentials in terms of accessibility.
But fuck Microsoft. you're not fooling anyone.