The training data thing is so frustrating.
Every service just changed their terms of service to allow it and nobody got in trouble.
Comment on Most Americans don’t trust AI — or the people in charge of it
hendrik@palaver.p3x.de 1 week agoAt the current state of things, AI just feels like being forced on people. There isn't much transparency and a lot happens without people's consent. Training data is taken without consent, and they display AI-written text, often riddled with msinformation to me withoit being upfront. I also stop reading most of the times, unless there is a comment section beneath for me to complain 😉
wise_pancake@lemmy.ca 1 week ago
Jimius@discuss.tchncs.de 1 week ago
My feeling is that it’s an AI bubble right now. The value seems apparent and money is being trucked in. But the uptake is lagging. Humans don’t need a piece of software that can write an essay for them. I want an AI that can find this obscure comic I read 10 years ago. That can order tickets for me. Find me the cheapest flights/connections to get from A to B. Summarize a text for me. My feeling is that it’s generative features are the least important.
It’s very telling that smart speakers are also in a very different place now. They were supposed to make shopping easier. That was how they were going to make money. But people just used them for music, asking for the weather and setting timers.
hendrik@palaver.p3x.de 1 week ago
Sure. I think you're right. I myself want an AI maid loading the dishwasher and doing the laundry and dusting the shelves. A robot vacuum is nice, but that's just a tiny amount of the tedious every-day chores. Plus an AI assistant on my computer, cleaning up the harddrive, sorting my gigabytes of photos...
And I don't think we're there yet. It's maybe the right amount of billions of dollars to pump into that hype if we anticipate all of this happening. But for a lame assistant that can answer questions and get the facts right 90% of the times, and whose attempts to 'improve' my emails are contraproductive lots of the times, isn't really that helpful to me.
And with that it's just an overinflated bubble that is based on expectations, not actual usefulness or yield of the current state of technology.
Jimius@discuss.tchncs.de 1 week ago
Other examples are drone deliveries. Was supposed to be the next big thing, but even more than 15 years later most companies are gone. And mainstream drone delivery is not a thing.
Or take AR/VR glasses. Supposed to revolutionize how we work. But in practice it’s mostly used to play games. First Google Glass and then the Apple Vision Pro gathered quite some attention but is already mostly forgotten. The VR space is still thriving, it’s just not the paradigm shifting technology the early investors wanted it to be. Facebook’s Metaverse cost 36 billion dollars and was a complete flop.