He's basically saying "Palantir is gonna eat your lunch while you regurgitate leftist opinion crap."
I had a debate with Grok and asked what it's sources were for its opinion given on a certain person. It was all AP, CNN, NYT, etc. I challenged it then to go deeper on some correlating relationships that person used for placement and to compare and contrast with another person who had "done good works" but not received the same attention and suddenly the conclusions changed. BUT! (and it's a big but) All these AIs are mirrors designed for stickiness. They will mimic your tone, model their voice to be compatible, and have straight up told me (without prompt) that they mirror the user as any sycophant would in real life.
So... Yeah, the real value is in the proprietary data. But the danger is how many will fall for the soft, encouraging, agreeable voice that pulls more private confessions and gently manipulates viewpoints.
Fuck em all. I'll keep asking for stupid mundane stuff but anything real and I'm going to try out Maple.
