t took it. Moved on.

Anyway, I'm not too worried about the DoorDash robot, but it won't be completely smooth sailing, but I think they have to get there eventually. Got to do it.

OpenAI has released a list of work tasks it says ChatGPT can already do in replacement of human activity. Now apparently OpenAI is sort of trying to walk this fine line between scaring you that AI will take all your jobs and then reassuring you that it will allow you to keep your job but do it better. So I think this is along the mode of telling people that some of their tasks might go away, but you might still need the person to oversee the task. So they're not going so far as to say here's a whole bunch of things that these people will lose their jobs. It might be just their jobs will be different.

They got 44 occupations and a whole bunch of tasks. And I think the permanent situation will be you need a human to closely watch almost everything the AI does and you can't get rid of the human because the LLM AIs, the ones we use, are prone to hallucination and you really can't have anything in the workplace that you're responsible for that could be a hallucination. You can't do that even once. I mean, that would be terrible for your career.

So, I'm going to say that what AI would like you to believe is that getting really really close to being completely able to do a human's job means that if you just wait a little bit longer, it's going to close the gap and then for sure it'll be able to do the human's job without intervention. I'm going to say no. I'm going to say that the closest you can get is somewhere around where we are, which is you can't trust it. I think "can't trust it" is its cap with current technology. And I'm no