I'm sorry, you make a lot of points correct. It's true that AI will have impact. Translations, chess, medical analysis. Yet, for many many tasks, it is plainly

I'm sorry, you make a lot of points correct. It's true that AI will have impact. Translations, chess, medical analysis. Yet, for many many tasks, it is plainly unreliable. We always think of machine being autonomous 24/7 and perfect in their job. This is a new kind of machine. It may be good at translation, but I wouldn't send a translated message without checking it first. This is what hinders its supposed capabilities. To software developers, this is immensely evident. I can chat with it for hours, no way it's going to help me with my depression more than 30 minutes with my therapist. It can analyze MRI scans, but it doesn't know me or have an holistic approach to health. It may translate and hallucinate. It often throws nonsense in the slop. I would not read an AI book that is not redacted because the average of its training materials... is bad. And it's not a technology's fault, it's just a limitation. It is not fixable: it's the way the drill drills. This makes it fundamentally good for very limited and controlled aspects in our job. It cannot code, it cannot review, but yeah, it can probably help you learn a programming language better than a tutorial, given that you recognize its mistakes when they happen. Also, it cannot build upon itself, it doesnt really understand instructions. In the industry, it's about as helpful as a helpful dog. Maybe it will bring you the slippers and take out the garbage, but it's cute and eats a lot.