> I've long maintained that the threat from AI to workers isn't that AI can do your job – it's that an AI salesman can convince your boss to fire you and replace you with an AI that can't do your job:
Yes, entropy always wins. But at the very local scale, and in the immediate term, we fight the chaos. It means things like trying to not burn the economy to the ground while things sort themselves out.
So obviously a technology to get a specific person to buy a specific thing at a specific time isn't going to happen. But the last ten years has taught me two things:
Many social media users are gullible enough to be convinced to act irrationally.
A well-trained AI on great data (which Facebook has) is pretty good at sorting signal from noise, i.e. advertisements that would or would not appeal to a particular user.
Sometimes I watch people 'use' Facebook. It's just endless scrolling...pause to watch a 10 second video...scroll...pause to click thumbs up...scroll...ad infinitum. Modern Facebook is hell.
Facebook is awful to use. I organise a group so I have an account, and there's things like the local food truck that only has a presence on fb so that's where you need to go to get the menu and phone number. Also my kids are in a bunch of activity groups like karate. So I use it once or twice a week to find some info or post some info.
The feed is pointless. It would be useful to see things from the groups I'm in. That's not what it shows.
If you're looking at a post and reload the page who knows where you'll end up. Why would you reload? Because it doesn't update the comments. Similarly the back and forward are broken. So I'll be looking at one thing, click to look at the next, click back... be lost, have to go back to the homepage, find the group again, find the post.
It's not that complex a site to be that complex or broken. I'm sure it's broken on purpose (ad serving, engagement etc), but because my usage of it is so narrow I don't see what they're trying to do.
I wish there was a good alternative. There are, of course, alternatives for all the things I do there, but they're not popular enough that I can direct people there.
What i would like to know is my digital footprint in the form of cookies that google and meta have access to. That digital image is a form of mental projection, and I presume meta's AI can target ads to that image.
Recommendations in the form of ads targeted to humans is a perfect usecase for LLMs. There is no right answer, and the interpreter is a human. Hallucinations dont matter, and if the targeting is a tad bit better, that justifies the investment.
I think the key feature of AI driven exploitation of cognitive flaws for the purpose of influencing people is the ability to remove any trace of hesitation or empathy from the process. Ai can adhere to the letter of ethical behaviour but remain blind to the the human consequences of manipulating a society into believing untruths and basing their resource choices on these lies about worth, beauty, desire, safety, or prosperity. This enables the zuckstrocity to turn up the cognitive dissonance up to eleven without creating an internal army of dissension and whistleblowers.
What are you saying is that AI can now be using to take "manipulation" of humans to the next level. and I agree. Thats how advertising works - ie to influence people and their purchasing behavior.
Oh irony. The article talks about exaggerated attention to success of individuals, like Zuck, being an issue in Silicon Valley. And yet the article itself talks about this AI Ad direction as something invented by evil Zuck.
This isn't really about Zuck. Ultra-convincing ads will happen soon regardless of what Zuck does.
Posting photos of people on a website so that random strangers can assign numbers to their appearance is not the same as discussing who you find attractive in private.
I'm confused how you would even come to compare these two scenarios?
This post is on to something.