I've busted AI models saying some outright false things. Even for simple things like the actors in a movie. It is a real danger if one doesn't verify because I've never had one admit to being wrong until I challenged it. I've had it happen while checking my math homework too.

There will probably be a day when verification is a lot more difficult, especially as more people become dependent and stop duplicating information based upon their own work. I see a future where we have centralized caches of information that AI agents access to serve us. The problem will be who controls those databases.

I saw a YT video today where a guy was explaining how he would use AI agents to summarize sources for his articles. Cool, but my agent could just replace him. At some point everybody will replace everybody, leaving fewer sources of data over time. That's a potential outcome I'm seeing anyway. We could probably mitigate the risks and have abundance instead. But will we? Doubt.

End AI rant.