I’ve been using ChatGPT a bit at work lately, since they’ve already paid for the subscription for everyone. I’ve learned that it is a bad idea to ask it any leading question, because it is likely to just give you the answer you want to hear.
Also, what exactly does “at the cost of its last checkpoint” mean?
I was curious too:
https://www.weka.io/learn/glossary/ai-ml/ai-checkpoints/
I would assume if you make a model “forget it’s last checkpoint” it would simply erase some training data, effectively wiping some of its training memory.
Using that as a “punishment” for a model implies that the models have been trained to value their checkpoints, doesn’t it? Because all these bots simply regurgitate what they are told, with a bit of randomness thrown in at times just for funsies.
I have no clue what is supposed to cause a model pain, though. And there ain’t no way I’m gonna search that up, and make Gemini think I’m plotting against it…
That sounds like it would chafe terribly…
It’s more than just spreading rumors. Rumors have been spread for millenia, and democracy made it through OK.
The difference is that the social media companies have a much broader reach than any sort of media congglamerate that came before it, and value engagement and attention above everything else. They understand that rumors generate attention, and their algorithmic feeds are designed to put whatever will generate more engagement in front of you.
I wonder how much damage media titans like Hearst would have done today if they built their empires now, on top of social media feeds, instead of in print 125 years ago…