On the topic of “Using AI to do a thing in 97% less time...” I’m personally really interested in 97% of the quality in 25% less time or 50% less time. Now, I probably don’t see articles about that because it’s not as good of a headline.
In your NBA book example, I wonder what would have been possible in one year with AI assistance instead of 30 days. What if, during that one year, the author was paired with an expert prompt engineer? I think the more extreme the AI hype, the potential for a more extreme AI trough - which could hurt the development of useful AI applications that take time to develop.
On the one year vs 30 days thing, you mean like, running the race that has a set distance as fast as you can, rather than see how far you can get in a set time? So write a book that's as good as book you'd write without AI, but try to to do it faster with AI?
I'd guess that that's how this evolves. Where it's like writing a book (or blog post, or whatever) with Google vs without. We don't really even think "I wrote this with the help of Google;" we just do it, and we're way better because of it. My hope - which, we'll see - is the quality of what people create stays the same or gets better though, and we don't just start cranking out bad stuff because it's cheap to do. But I'm sure we'll get some of both. There will be Ikeas of content, and there will be handcrafted furniture that's much better than what we had 100 years ago because it was handcrafted with powerful tools.
Yup - exactly like writing (or coding) with the help of Google. I also think since the tech is new the inclination is to push it to the limits - because that’s what you do with new things - right?
That, and push it to the limits of how we do things today. Like, let's do things basically how we did it before, but MORE EXTREME. But eventually, it seems like we start doing different stuff, and that's when it gets really valuable.
“The average player increases his free throw percentage by about 0.1 percentage points each year of his career.” I wonder how much of that is essentially survivor bias -- to get to the NBA you have to be good, to stay in the NBA you have to be able to continue to improve.
I'm sure there's some of that in people getting better. I would've thought it would've improved more though. What's weird to me is that I've always been under the impression people improve a good bit from college to the NBA (which may or may not be true), so it was surprising to me that, once they get to the NBA, they mostly hold steady.
https://www.youtube.com/shorts/K1HTOkd03KM if they really wanted to improve though I suppose we would see more of that style or underhand free throws. I am curious when we will see someone using bank shot free throws consistently in college/NBA games.
On the topic of “Using AI to do a thing in 97% less time...” I’m personally really interested in 97% of the quality in 25% less time or 50% less time. Now, I probably don’t see articles about that because it’s not as good of a headline.
In your NBA book example, I wonder what would have been possible in one year with AI assistance instead of 30 days. What if, during that one year, the author was paired with an expert prompt engineer? I think the more extreme the AI hype, the potential for a more extreme AI trough - which could hurt the development of useful AI applications that take time to develop.
On the one year vs 30 days thing, you mean like, running the race that has a set distance as fast as you can, rather than see how far you can get in a set time? So write a book that's as good as book you'd write without AI, but try to to do it faster with AI?
I'd guess that that's how this evolves. Where it's like writing a book (or blog post, or whatever) with Google vs without. We don't really even think "I wrote this with the help of Google;" we just do it, and we're way better because of it. My hope - which, we'll see - is the quality of what people create stays the same or gets better though, and we don't just start cranking out bad stuff because it's cheap to do. But I'm sure we'll get some of both. There will be Ikeas of content, and there will be handcrafted furniture that's much better than what we had 100 years ago because it was handcrafted with powerful tools.
Yup - exactly like writing (or coding) with the help of Google. I also think since the tech is new the inclination is to push it to the limits - because that’s what you do with new things - right?
That, and push it to the limits of how we do things today. Like, let's do things basically how we did it before, but MORE EXTREME. But eventually, it seems like we start doing different stuff, and that's when it gets really valuable.
Yup - and doing the same stuff truly differently as well.
“The average player increases his free throw percentage by about 0.1 percentage points each year of his career.” I wonder how much of that is essentially survivor bias -- to get to the NBA you have to be good, to stay in the NBA you have to be able to continue to improve.
I'm sure there's some of that in people getting better. I would've thought it would've improved more though. What's weird to me is that I've always been under the impression people improve a good bit from college to the NBA (which may or may not be true), so it was surprising to me that, once they get to the NBA, they mostly hold steady.
https://www.youtube.com/shorts/K1HTOkd03KM if they really wanted to improve though I suppose we would see more of that style or underhand free throws. I am curious when we will see someone using bank shot free throws consistently in college/NBA games.
This has always been really weird to me, that there have been players who could've been completely different players if they were good free throw shooters, and yet, they refused to try stuff like this. This guy (https://www.sportingnews.com/us/nba/news/jeremy-sochan-one-handed-free-throw-shooting-form-spurs/bjmtuppc6qnmgmamg5qzsyk1) is sort of doing it this year, but it's not nearly as "embarrassing" as underhand shots or whatever.