- 0 Posts
- 35 Comments
Concept artist Sam Santala pointed out that you can’t raise a support request to discuss the terms without first agreeing to them. You can’t even uninstall the apps!
Well, at least they made this as easy as possible to challenge in court.
But given the way AI bros behave I suspect this is a smash and grab. The penalties don’t matter if they get to keep the goods in the form of generative AI training.
The biggest problem with AI alt text is that it lacks the ability to determine and add in context, which is particularly important in social media image descriptions. But people adding useless alt text isn’t exactly a new thing either. If people treat this as a starting place for adding an alt text description and not a “click it and I don’t have to think about it” solution I’m massively in support of it.
This is from the account that spread the image originally: https://x.com/ai_for_success/status/1793987884032385097
Alternate Bluesky link with screencaps (must be logged in): https://bsky.app/profile/joshuajfriedman.com/post/3ktarh3vgde2b
I’m not using “AI” in web searches no matter how much any VC bro’s golden parachute depends on it, sorry. Refusing to partake or even using tools to filter out LLM trash are perfectly fine ways to adapt to search engines leaning on AI hype to try to convince you that their inability to combat SEO spam is good, actually.
“How would these students know how to barricade a door?” Daughtry asked on Newsmax, as he charged that protesters wouldn’t have been capable of measures like locking doors with chains, blocking them with vending machines, or disabling security cameras.
Come the fuck on. I’m too old to have been through school shooter drills, but I’m not too stupid to remember that for the last 20 years this is what kids have been taught to do to deal with psychopaths with guns on campus. They start teaching it in kindergarten! At this point a kid not knowing how to do this stuff is like a kid that doesn’t know to stop, drop, and roll when on fire.
Apologies for being so sketchy on the details but I really can’t remember too many of the specifics. I’m fairly certain it wasn’t that his family name came first, because that’s fairly straightforward. I think the author might have been from an east or southeast Asian culture? I think that part of the essay might have been about how addressing him as Mr. Firstname is actually more formal than Mr. Lastname, even though Firstname is not his family name. I don’t want to keep guessing on more details about how the naming conventions were different because I’m probably going to get it wrong, I have fairly low confidence in what I remember from it.
Because I have been completely unable to find it again and this seems like a relevant place to ask: does anyone have a link to an article similar to this, that I believe might have been titled ‘My First Name is My Last Name’? This is made extra hard to look up because I’ve forgotten the specific culture and details it’s talking about, but it’s about the same basic issue with cultural conventions on names.
Not only is T-Mobile using these models when customers call in, but we’ve heard via multiple tips that T-Mobile is planning to use AI learning to predict why a customer might be visiting a store, too. There has allegedly been discussion of using the T-Mobile app to detect a customer’s account when they enter a store, and have AI predict why they might be visiting.
God help me if I have to worry about some issue with my phone service being even more annoying to get fixed because the AI spits out I’m really there for an unrelated reason and just need to be cajoled into admitting it.
The way some people in this thread are responding to the issue of elder poverty is so nasty. “Well they should have just bought a house when they were younger and since they’re old it’s probably their own fault anyway” is some shit I’d expect to see out of conservatives. There are SO many reasons an elderly person may be living in poverty, and none of them mean they don’t deserve secure housing and food.
TikTok is still home to ridiculous (and even deadly) misinformation, especially vaccine related misinformation, and it’s still a corporation that makes billions of dollars per year. They also have a history of a litany of issues with LGBTQ users. It’s kind of ridiculous to compare scientific study of it as a social platform to invading the home of a private citizen.
DeWine is a fascist shitbag, there’s zero chance this starts and ends with “protecting kids.”
What’s most concerning is how this “restoration” process can be done multiple times if the victim never becomes aware that they’ve been compromised. Even worse is how even after a Google Account password reset, this exploit can be used one more time by the bad actor to get access to your account.
Does this mean if you change your password twice you’re fine? Or is the second sentence just worded awkwardly?
ChatGPT is OK at summarizing popular, low specificity topics that tons of people have already written a ton about, but it’s terrible at anything else. When I tested its knowledge about the process of a niche interest of mine (fabric dyeing) it skipped completely over certain important pieces of information, and when I prompted it to include them it basically just mirrored my prompt back at me.
Which has pretty much summed up my ChatGPT experience: it just regurgitates stuff I can find myself, but removes the ability to determine if the source is reliable. And if it’s something I’m already having trouble finding detailed information about it usually doesn’t help.
Have you heard of Hardspace: Shipbreaker? If anything it’s the opposite of a crafting-building game, it’s a disassembly game! It’s set in space and has a nice plot about corporate greed. And you have to manage your oxygen in space, so arguably it’s like diving, haha.
Another game I liked that’s made by a small studio is Carrion. You play as a The Thing-esque monster that has to kill, grow, and solve different puzzles to escape the lab it was created in. Very gory but fun.
Acknowledging this is a dragnet, a practice generally considered unconstitutional since the 1950s, actually illustrates pretty well why people are upset about it. Even if it would result in more easy prosecutions for cops, it doesn’t change that it’s mass surveillance and an unconstitutional practice.
For my money I would bet the issue stems from abandoning Google server hosting, either from arrogance or being unable to afford it.
Even if you perfectly solve crashes (doubtful, definitely not something that exists right now), the problems of microplastic pollution from tires and emissions from electricity generation are non-trivial.