from Hacker News

Microsoft: 'ever present' AI assistants are coming

by mindracer on 10/3/24, 6:10 AM with 70 comments

• by peutetre on 10/3/24, 7:28 AM

  > But AI supporters argue that in order to be truly useful, these tools have to be deeply embedded into our lives: that they can only be really helpful if they know the history and context behind what they are being tasked to do.

  The problem is these AI assistants won't work for me, they'll work for Microsoft. They won't help me as much as they will point me in the direction that is the most profitable for Microsoft.

  When the agenda is not mine then these things are of no use to me.

• by myprotegeai on 10/3/24, 7:29 AM

  I can't wait until little Johnny's AI assistant (that's been with him since birth) can communicate to Big Tech's Global HR Overlord AI and they can decide together that a good job is not a privilege that a wrongthinker like Johnny should enjoy.

• by deergomoo on 10/3/24, 7:15 AM

  Given this is Microsoft, I’ll assume the “whether you want them or not” is implied.

• by fldskfjdslkfj on 10/3/24, 7:32 AM

  They misspelled "ever present user data collection"

• by noirscape on 10/3/24, 9:17 AM

  I think what Suleyman is missing here is that most people aren't fine with their "TVs, laptops, in-car cameras and earbuds" recording everything continuously.

  The most popular modification made to laptops is a small round sticker to seal off the webcam. The enshittification of the car market is something that only car manufacturers seem to like. The most common question with a smart TV is "how do I get that thing to just accept an HDMI input and not get in the way". Making a slight jump from car cams - things like doorbell cameras are facing increased scrutiny because of their inadvertent recording of the nearby street, which might violate privacy laws.

  Adding audio and the few seconds before and after to a photo isn't the same thing as a continuous recording; the choice to make the photo is both still with the user and what's happening doesn't feel like a black box. AI assistants are the black box.

  It was cute when Siri could tell you the nearest sushi bar or can check the schedule from your calendar (although I'll note that every use for Siri tends to be a party trick and half the fun people get from Siri and other assistants is when they don't work as expected; I don't know a single person who uses Siri to for example, read the news or give them a briefing on what's important for the day a-la Star Trek or other utopian scifi), it's creepy when Siri starts to suggest that you might want to visit a sushi bar because it's tracked your food habits for the past week and thinks you could enjoy some more fish food instead. AI assistants overwhelmingly are heading to the latter and now that the party trick element is cooling off, people get more cynical towards how they work (and how they don't work as advertised.)

• by fakedang on 10/3/24, 7:43 AM

  Microsoft can't even be arsed to implement proper search in Windows 10/11, and they claim this.

• by rsynnott on 10/3/24, 7:46 AM

  Oh, right, Microsoft says it? It’ll definitely happen, then. Their pronouncements about the future of computing are famously accurate. That’s why we’ve all been using Windows for Pen Computing since the noughties.

• by dspillett on 10/3/24, 10:42 AM

  > But AI supporters argue that in order to be truly useful, these tools have to be deeply embedded into our lives

  A tool controlled by corporates like Microsoft that deeply embedded into my life? If that is what it takes to be truly useful then I don't want it to be useful and will be opting out in every way possible. With force if needed…

• by croes on 10/3/24, 7:50 AM

  It will be interesting to see when the first conviction based on this feature will take place.

• by karel-3d on 10/3/24, 7:16 AM

  That's just Clippy

• by phkahler on 10/3/24, 8:41 AM

  >> For example, an AI diary manager can only organise your diary if it can access that diary, edit it, and retain information about your activities.

  Edit a diary? Retain the information? A diary IS a log of information as understood at the time. It's also not the clouds business, nor to be edited.

  My gosh these people sound stupid.

• by jasonvorhe on 10/3/24, 7:54 AM

  Never forget that most of Western governments pretty much run on Windows and Microsoft services.

• by drooopy on 10/3/24, 9:36 AM

  How about you go back to making an OS that takes less than 10 GB of disk space, doesn't spy on me and is not filled with useless ads everywhere?

• by jerpint on 10/3/24, 8:23 AM

  If these models could run offline, 100% locally, with expected privacy, Im up for it, i want to use a product, not be training data

• by NoGravitas on 10/3/24, 1:21 PM

  The headline feels like a threat.

• by GiorgioG on 10/3/24, 7:32 AM

  I expect we’ll look back on this in 5 years and laugh our asses off. AI offers glimpses of brilliance and then sprinkles a bunch of turds around it. 60% of the time it works every time.

• by Mountain_Skies on 10/3/24, 9:06 AM

  The only thing really keeping me on Windows now is Visual Studio. I know there are alternatives on Linux but for my particular work, there's no substitute for the actual full pro version of Visual Studio. But this non-stop parade of enshitification of Windows is really making me think about going back to Linux for my daily driver outside of work tasks. I don't want to have to keep up with what privacy invasion Microsoft has cooked up this week and keep track of the current way to opt-out, if that's even possible. Since first installing Slackware in the mid 90s, I've done several multi-year stints with Linux as my desktop OS, so it wouldn't be a completely alien change but in general I've found Windows to be a bit easier day-to-day and gets in the way less. All this AI and snapshotting garbage, on top of the existing telemetry, is adding up to be too much.

• by pixelpoet on 10/3/24, 8:30 AM

  I wish they would give a shit that nobody wants this and actively hates it. Please fuck off with your AI trash, big tech :(

• by agos on 10/3/24, 9:05 AM

  I wonder if they realized how ominous it sounds

• by kaimac on 10/3/24, 8:47 AM

  could be great news for linux adoption

• by JohnFen on 10/3/24, 5:20 PM

  No no no. There is literally nothing about this that I find useful or acceptable, and a whole lot that I find dangerous and frightening. It's handing far too much power over to companies that I fundamentally don't trust.

  I'm not getting on this train.

  > AI supporters argue that in order to be truly useful, these tools have to be deeply embedded into our lives: that they can only be really helpful if they know the history and context behind what they are being tasked to do.

  That cost is far too high.

• by xtiansimon on 10/3/24, 12:23 PM

  First it was every online support had the idiot computer voice, Uh. Sorry. Please tell me again in a few words…

  Now I’ve noticed all of these support services, once you get past the idiot computer, there is now an idiot human who rephrases your question or concerns to the point of stupidity, You’re calling because you’re having problems using the website…

  Meanwhile, I’m waiting for sentiment analysis and an idiot computer telling me I sound frustrated, How are things at home? Are you having a bad day?

• by xarope on 10/3/24, 7:32 AM

  MS Head of AI says there'll be more AI. News at 11.

  In other news, MS Head of AI says Clippy is making a comeback...

• by guidedlight on 10/3/24, 8:19 AM

  I don’t know why everyone is so concerned. Microsoft has been sending telemetry data and your photos to Microsoft for years.