Conservatives think ChatGPT has gone 'woke'

The refinement of ChatGPT's "abilities" over the past 6 weeks has been very interesting to watch on /r/chatgpt. People are extremely agitated that their once baudy and severely humorous chatbot has been nerfed into an obnoxiously clean-cut uptight dweeb. Some get actually depressed, some get very upset. It's like watching everyone's bar buddy sell out and start wearing polos.

[flagged]

I think these systems could be greatly improved by leaning towards more speculative outputs. I had initially hoped to use ChatGPT to fact-check my writing, but found that it occasionally made completely-false assertions. If its tone were less assertive, and more speculative, the added bonus is that you wouldn't have to filter as much. Results could be presented in the format of "this source claims that xyz, while this source claims abc" structure, which used to be the crux of quality journalism. I get the fact-checker I want, and the whinier ends of the political spectrum get their ideas presented in a way that doesn't treat it as absolute truth.

Now I'm wondering what it would be like if the model were strictly trained on text that Conservapedia might approve of.