This is a most excellent place for technology news and articles.
A 100% accurate AI would be useful. A 99.999% accurate AI is in fact useless, because of the damage that one miss might do.
It's like the French say: Add one drop of wine in a barrel of sewage and you get sewage. Add one drop of sewage in a barrel of wine and you get sewage.
99.999% accurate would be pretty useful. Theres plenty of misinformation without AI. Nothing and nobody will be perfect.
Trouble is they range from 0-95% accurate depending on the topic and given context while being very confident when they’re wrong.
I think it largely depends on what kind of AI we're talking about. iOS has had models that let you extract subjects from images for a while now, and that's pretty nifty. Affinity Photo recently got the same feature. Noise cancellation can also be quite useful.
As for LLMs? Fuck off, honestly. My company apparently pays for MS CoPilot, something I only discovered when the garbage popped up the other day. I wrote a few random sentences for it to fix, and the only thing it managed to consistently do was screw the entire text up. Maybe it doesn't handle Swedish? I don't know.
One of the examples I sent to a friend is as follows, but in Swedish;
Microsoft CoPilot is an incredibly poor product. It has a tendency to make up entirely new, nonsensical words, as well as completely mangle the grammar. I really don't understand why we pay for this. It's very disappointing.
And CoPilot was like "yeah, let me fix this for you!"
Microsoft CoPilot is a comedy show without a manuscript. It makes up new nonsense words as though were a word-juggler on circus, and the grammar becomes mang like a bulldzer over a lawn. Why do we pay for this? It is buy a ticket to a show where actosorgets their lines. Entredibly disappointing.
Most AIs struggle with languages other than English, unfortunately, I hate how it reinforces the "defaultness" of English
We're not talking about an AI running a nuclear reactor, this article is about AI assistants on a personal phone. 0.001% failure rates for apps on your phone isn't that insane, and generally the only consequence of those failures would be you need to try a slightly different query. Tools like Alexa or Siri mishear user commands probably more than 0.001% of the time, and yet those tools have absolutely caught on for a significant amount of people.
The issue is that the failure rate of AI is high enough that you have to vet the outputs which typically requires about as much work as doing whatever you wanted the AI to do yourself, and using AI for creative things like art or videos is a fun novelty, but isn't something that you're doing regularly and so your phone trying to promote apps that you only want to use once in a blue moon is annoying. If AI were actually so useful you could query it with anything and 99.999% of the time get back exactly what you wanted, AI would absolutely become much more useful.
It's all just to get more data from you so it can monetized.
The AI thing I'd really like is an on-device classifier that decides with reasonably high reliability whether I would want my phone to interrupt me with a given notification or not. I already don't allow useless notifications, but a message from a friend might be a question about something urgent, or a cat picture.
What I don't want is:
It seems like those are the main AI features bundled on phones now, and I have no use for any of them.
That's useful AI that doesn't take billions of dollars to train, though. (it's also a great idea and I'd be down for it)
Unless it can be a legit personal assistant, I’m not actually interested. Companies hyped AI way too much.
Not sure if Google Lens counts as AI, but Circle to Search is a cool feature. And on Samsung specifically there is Smart Select that I occasionally use for text extraction, but I suppose it is just OCR.
From Galaxy AI branded features I have tested only Drawing assist which is an image generator. Fooled around for 5 minutes and have not touched it again. I am using Samsung keyboard and I know it has some kind of text generator thing, but have not even bothered myself to try it.
Not sure if Google Lens counts as AI, but Circle to Search is a cool feature.
Not to the point where it's worth having a button for it permanently taking up space at the bottom of the screen.
On a lot of phones you can hide the navigation pill, but Samsung started forcibly showing it when they added Circle to Search. Fortunately I don't have a Samsung phone.
It's cool
Is it useful? Idk
They're kinda like the S-Pen... is it cool? Sure! Do I find myself using it? No, not really.
Certainly counts, Samsung has a few features like grabbing text from images that I found useful.
My problem with them is its all online stuff and I'd like that sort of thing to be processed on device but thats just me.
I think folks often are thinking AI is only the crappy image generation or chat bots they get shoved to. AI is used in a lot of different things, only difference is that those implementations like drawing assist or that text grabbing feature are actually useful and are well done.
I'm shocked, I tell you. Absolutely shocked. And if you believe that, I got some oceanfront property in Arizona. I'll sell you too.
AI was never meant for the average person but the average person had to be convinced it was for funding.
AI is useless and I block it anyway I can.
"Stop trying to make ~~fetch~~ AI happen. It's not going to happen."
AI is worse that adding no value, it is an actual detriment.
I feel like I'm in those years of You really want a 3d TV, right? Right? 3D is what you've been waiting for, right? all over again, but with a different technology.
It will be VR's turn again next.
I admit I'm really rooting for affordable, real-world, daily-use AR though.
Just look at Smart Speakers. Basically the early AI at home. People just used them to set timers and ask about the weather. Even though it was capable of much more. Google and others were unable to monetize them for this reason and have mostly given up. (Protip: if you have a google speaker and kids, ask about the animal of the day. It's an addition during COVID times for kids learning at home.)
But people also aren't used to AI yet. Most will still google for something, some already skip that step and have ChatGPT search and summarize. I would not be surprised if the internet of the future is just plain text files for the AI agents to scrape.
"useless" is a more positive impression than I have.
Sometimes I wonder what is going to happen to all this tech in 4 or so years when its less profitable to keep the AI centers on.
Right now they are "free" because of all the investment that is going on. But they have a huge maintenance/energy cost.
dint MS said their AI isnt as profitable, google is sure hellbent on going with AI.
They just need to capitalize the surveillance capabilities. Find a way to convince users they need access to everything on their phones in order to sell them first class convenience. Once you've done that there's plenty of money to be made.
A damning result for AI pump and dump scammers.
every NVDA earnings call lol. Old man Jenson had a (chip) farm, AI AI OH! guy literally said AI almost 100 times in a call.
Anyone who has been paying attention has been waiting for this enormous bag of shit to explode already.
On Samsung they got rid of a perfectly good screenshot tool and replaced it with one that has AI, it's slower, clunky, and not as good, I just want them to revert it. If I wanted AI I'd download an app.
You are thinking about Smart Select? I just take fullscreen screenshot and then crop it if I need part of it. Did it even when I had previous Smart Select version. Overall I think new version with all previous 4 select options bundled in 1 is better.
Yes, Smart Select. I do that now, but taking a full screenshot and cropping it is slower for me than the old Smart Select. I hate this new version, it's slower and doesn't work the same, we should get the option to pick, but they forced the upgrade and I have no choice.
I love the AI features for photos of my galaxy, but other than that I don't use it
The first thing I do with a new phone is turn off any kind of assistance.
Honestly I can't say I've ever had a reason to use it on my phone.
Surprise surprise!
At work we deal with valuable information and we gotta be careful what to ask. Probably we'll have a total ban on these things at work.
At home we don't give a fuck what your AI does. I just wanna relax and do nothing for as long as I can. So off load your AI onto a local system that doesn't talk to your server and then we'll talk.
In my office there's one prototype model under testing that nobody uses and does nothing useful. Anything else is actually banned, we handled way too sensitive information. It causes office and outlook to glitch often when it tries to open copilot and get immediately slapped silly to shut up. The blinking blank windows are annoying though. IT had to make an special communication to all staff explaining that it was normal behavior.
