I'm all for folks expounding on the ways in which "current AI" sucks, but I have little patience for people who think that what it is now is all it will ever be

I'm all for folks expounding on the ways in which "current AI" sucks, but I have little patience for people who think that what it is now is all it will ever be. Sure, it could be 6 months or 6 years before we see real improvements, but no one knows. No one knew that transformers were coming in the months before they arrived. One thing that I think is encouraging is how fast computing power is increasing, along with languages, frameworks and models which allow you to make use of that power. I think that alone, even without AI, will result in world-changing advancements across many fields. The power to simulate large systems is growing rapidly and will enable prediction and discovery to achieve great societal benefits.