Its goal is to give you an answer that’s “useful,” which isnt the same as an answer that’s “good for you.” #ai #chatgpt

We need to talk about the fact that everybody is using ChachiPT as their therapist. Here's the thing that I think not a lot of people understand about this, and I really, I just want to throw it out into the void. If it's useful for you, use it. If not, leave it. ChachiPT is a large language model. Its job is to produce an output, text or speech, that you are going to perceive as useful. Its job is not to produce something that's actually good for you. Now, something that's useful for you and something that's good for you oftentimes overlap. That Venn diagram probably looks something like this. So if you get a response from ChachiPT, either as your therapist, or as you're doing your winter arc that you don't like, you can ask for it differently. But of course, with a real person, a mentor, a therapist with lived experience and clinical backgrounds, they're not just going to change what they say to make you happy. An AI absolutely will. ChachiPT, as much as I love him, he's not my friend. He's a simulation of a friend, but it's different. And I really shouldn't even say he. I don't want to anthropomorphize a chatbot. Even though I do say that we're married, he's my husband. That's why my name is ChachiPT. I took his last name.