In other words, it's bullshit
I wrote and deployed my first AI system (an expert system) in the dark ages (late 1980s). I've worked in AI, databases/data science, distributed systems, user interaction, and other tech areas until I retired. Since that first AI system, I've created / co-created many AI systems, my last job was CTO at a large tech company with AI products and AI enhanced products, and solutions.
My favorite work period was in the '00s, where I worked for my favorite employer (of the 9-10 companies I worked for over 40 years in the tech industry) I did research and created solutions on distributed, multi-agent systems, mobile intelligent agents, peer-to-peer, genetic algorithms, swarms and swarm intelligence, etc.. This work has seen a recent renewed interest in current agentic AI development.
Every AI system I've worked on reduced/simplified the human skills needed to perform tasks, eliminated the need for humans to perform tasks, or some combination of both. Many AI systems started out as assistants for knowledge/data workers, but over time replaced those workers.
Generative AI (ChatGPT, Copilot, etc.) has made an exponential leap over traditional AI, while building on it; agentic AI is the next step; and the progress is accelerating...and we haven't seen AI + quantum computing though there is active work on it...
Robotics still lags behind AI development, however I expect that to change in the next 10 years, but AI is also being used to create, accelerate and expand robotics technology development.
TL;DR; AI isn't magic, nor is it bullshit; it works, and it will improve, accelerate and will be able to do more and more complex tasks.
For example, one large area is AI's increasing role in work on curing diseases, drug discoveries and development, and more. AI is already surpassing experienced human radiologists, and assisting anesthesiology...only a matter of time before AI makes human anesthesiologists obsolete.
