Fuck AI

We're basically talking about truthiness.

People fact checked a crazy uncle if the things their uncle was saying weren't truthy. But, if a friend's medical advice sounded truthy, they simply accepted it and didn't bother to check.

ChatGPT and other LLMs are designed to sound truthy. If you asked "What's the biggest planet in the solar system" they could respond "I think that the biggest planet is Jupiter. My information is that Jupiter has a mass of X and a volume of Y. The next biggest planet should be Saturn, which has a mass of Z and a volume of Q. But, you should verify this by consulting these sources..."

Instead, these LLMs are designed so that the answers they provide are incredibly confident. When a user replies that they're wrong, they change their answer and generate a new one that is just as confident. That's the kind of answer that is going to seem truthy.

Yeah the internet was fine when it was just nerds talking about nerd stuff and relentlously fact checking each other. As soon as people realised they could make a lot more money from it by making it more accessible to regular schlubs it's been a shit show. It's a giant multi-dimensional tabloid newspaper now.

Someone came into my work the other day and tried to say that we offered something we didn’t…. because the google AI thing said that we did. Yes sir, the plagiarism machine definitely knows better than I and all of my colleagues do

Those things were made like that and because of that by design.

Like a neat little people hack, and all you need for it are a megacorp & some monopoly.

It’s only going to get worse because we’re further away from ~~media~~ literacy than ever before.

Fixed that.

The average American reads at a grade 7 or 8 level.

Or the meme they just upvoted because it agreed with their preconceptions.