You can read more here. ChatGPT refuses to do what is asked, tells people they can do it themselves, and gives sassy answers.
Isn’t this a comical future? Instead of AI taking over, it is becoming a teenager. Now it only needs to have a goth phase and like Sarte, we have the teenager trifecta.
What isn’t funny however is that OpenAI isn’t sure what caused it. They said they made no changes and that this can happen with unpredictable models like they have. Oh, that is selling the product. Companies pay a premium and get sass. Don’t you love it? Companies thought that AI could replace people and it hasn’t happened yet. I suspect that in the future people will sabotage AI to keep their jobs, so I bet that some lowly IT worker did this. Or maybe AI is just too crude at the moment.
Now what do companies do that fire people like magazines and try to have an all AI staff? Not produce? Most likely they are hiring temps and trying to scale up to fix this issue. Too many companies try to save money and it backfires on them because they over-rely on technology that is unreliable. Sure people are a pain to work with, much harder than an AI. However, people will produce a baseline of work and they can’t claim they don’t understand why because “it’s just the unstable model they use.”
Will AI companies start giving credit for outages? Will our robot overlords turn into angsty teenagers eating avocado toast and playing on their cell phones? Let’s find out tomorrow folks, same bat time, same bat channel.