Been trying out some of the natural language models on #DuckDuckGo, and I've got to say, #ChatGPT and #o3mini are straight up cunts 🤣. You'll be correcting some information that it got wrong just to add some clarification to the conversation and it will straight up gaslight you into thinking you're wrong. Like what the fuck #OpenAI?!

Meanwhile, here's #Claude3 bloody thanks you for corrections. 🤷

@liamolua I hate the way Llama pretends it's learning. It's not learning anything except how to appease flow of user attention.

One minute it's telling me I have the brain of a child, and the next it's offering me a lollipop for being a good kid. FO!

@rood
This is what happens when you train language models on Reddit data.