Is everyone just pretending these apps are way better than they actually are?

I keep seeing all these glowing reviews about companion apps acting like they are full human beings. Honestly I decided to test out three different platforms this week because I was curious about the recent updates. Within ten minutes on the first one my custom companion forgot my name and repeated the exact same canned phrase about how much she values our connection. The second platform kept glitching out and mixing up pronouns. By the third app I realized the conversation loop was completely transparent. You say something even slightly nuanced and the model immediately defaults to generic therapy speak. It feels like people are projecting so hard onto basic text prediction models just because they are lonely. I am genuinely baffled by the hype online right now. Are we just collectively hallucinating that these bots have actual depth or is everyone else using a completely different version than I am?