I was wrong and I underestimated the technology. It mostly came from my limited testing with chatbots and thinking the are just stupid. They used to claim oposite of what is true quite frequently. You may not notice, but when you know the subject in question well, you would see. Then I saw some stupid people using it badly. For example, as if it is an oracle for truth. People using "chatgpt says so!" as an argument from authority still deserve to be smacked is the face with a wet fish.
The mistake was to think it will not impove much. Yet the models got massive improvements since then. Then smart people learned how to work around the limitations that still exist, and they found tasks at which the tech is mega useful.
Now I predict there will be robot wars. A ship landing with thousands of drones coming out, able to ocupy an etire city with 15 minutes. Ping me in 10 years to tell me I was wrong, again.