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.
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.