None of the previous stabs at AI were more than a parlour trick, modern AI are capable of not only full and natural conversations but have the unique ability to turn that into completing tasks based on how well the human operator can describe the problem and explain the proposed solution.

It’s not always perfect, but it gets close enough for the professional to make use of it by cutting out the research phase of any given project. Or by getting the bulk of the work done without the hours it would have taken to do it. Refining the solution might take ten to fifteen minutes but you don’t have to be a math genius to see the benefits. Plus the models we have now are exploding in niche use-cases. We have image generation, voice generation, code generation, all at near human standards. I’ve had it walk me through how to deploy python scripts via VSC, then I had it walk me through setting up a Git repository, then I asked it to take me through a DnD/Choose your own adventure scenario with specific choices having consequences down the line. It was a little basic but I gave it a preestablished universe and the general premise, it researched the rest on its own and used the data to fill in the gaps in a way I hadn’t even suggested based on what it found of the universe.

That last one isn’t a productive use case, sure. The point is that what we have now isn’t just some one off computer like a chess bot or a Smash Bros CPU set to its highest level, it’s a seed for every future version of machine learning algorithm that will be used to specifically design models for special scenarios. It’s become ingrained in our society now, and it’s unlikely to just disappear like the rest of what you’re describing.