The only codebase an agentic system could refactor in 15 minutes would be almost trivially small. I still couldn’t do it in 15 mins, but give me a couple hours and I’ll make much more meaningful improvements
I think you’re missing the forest for the trees here because the point is, you’re capable of doing the task, just not doing it in the same amount of time as a computer.
You chose a poor analogy to explain your POV. I’m pointing out the flaw in it.
And I could manually relocate all the contents of a palette, too. Just not anywhere near as quickly and easily as I can with a forklift. The analogy is still apt.
Ok. Look at it the other way. The person who can lift the heavy thing may not be able to continue to lift the heavy thing if they use the forklift all the time and don’t ever train their muscles. Which is what the article is pointing out. Doing the task by hand re-enforces knowledge and skill. Over-reliance on a tool is a well known phenomenon.
And yet the person with the forklift is moving more stuff than the guy who did it by hand could manage. The “over” in “over-reliance” is a subjective value judgment and I just don’t agree.
I’m not seeing the problem here. Technology is developed specifically for this purpose, to remove unnecessary burden from humans and enhance their capabilities. There’s nothing noble about laboring unnecessarily hard to accomplish goals in a suboptimal manner. I could write programs in assembly language but instead I use high-level languages and compilers. Does that result in over-reliance on compilers?
John Henry died in the process of “beating” the steam hammer and then got replaced anyway. Nowadays it’d be considered foolish to do that work by hand.
Ok, there are definitely a lot of trades where things are still taught by hand in the event that you have to do them by hand some day. Doing it by hand does more than just re-enforce knowledge. It also teaches you new things and allows a process, and the space to re-evaluate and innovate. We improve by doing those things by hand. That is very often worth the cost. You very often don’t get things quickly, cheaply, and with quality. The AI will degrade if we don’t provide it with quality information to work with. It is nothing without our skills. We won’t have those skills if we don’t use them.
You talk like a businessman rather than an artisan or a creator of things, so perhaps your mindset is different but what happens when the AI breaks something and nobody can fix it because they lack the ability to think about the problem constructively or understand what the problem actually is.
As I said, I can write programs in assembly language. I have actually done so, small trivial ones. I’m not a businessman, I’m a programmer. But I use compilers basically all the time because it would be ridiculous not to.
If an AI is able to break something in a way that no human can fix then I suppose that’s a sign that AI has exceeded human capabilities. Do you think it’s there yet?
I physically can’t refactor a codebase in 15 minutes.
The only codebase an agentic system could refactor in 15 minutes would be almost trivially small. I still couldn’t do it in 15 mins, but give me a couple hours and I’ll make much more meaningful improvements
I think you’re missing the forest for the trees here because the point is, you’re capable of doing the task, just not doing it in the same amount of time as a computer.
You chose a poor analogy to explain your POV. I’m pointing out the flaw in it.
And I could manually relocate all the contents of a palette, too. Just not anywhere near as quickly and easily as I can with a forklift. The analogy is still apt.
Ok. Look at it the other way. The person who can lift the heavy thing may not be able to continue to lift the heavy thing if they use the forklift all the time and don’t ever train their muscles. Which is what the article is pointing out. Doing the task by hand re-enforces knowledge and skill. Over-reliance on a tool is a well known phenomenon.
And yet the person with the forklift is moving more stuff than the guy who did it by hand could manage. The “over” in “over-reliance” is a subjective value judgment and I just don’t agree.
I’m not seeing the problem here. Technology is developed specifically for this purpose, to remove unnecessary burden from humans and enhance their capabilities. There’s nothing noble about laboring unnecessarily hard to accomplish goals in a suboptimal manner. I could write programs in assembly language but instead I use high-level languages and compilers. Does that result in over-reliance on compilers?
John Henry died in the process of “beating” the steam hammer and then got replaced anyway. Nowadays it’d be considered foolish to do that work by hand.
Ok, there are definitely a lot of trades where things are still taught by hand in the event that you have to do them by hand some day. Doing it by hand does more than just re-enforce knowledge. It also teaches you new things and allows a process, and the space to re-evaluate and innovate. We improve by doing those things by hand. That is very often worth the cost. You very often don’t get things quickly, cheaply, and with quality. The AI will degrade if we don’t provide it with quality information to work with. It is nothing without our skills. We won’t have those skills if we don’t use them.
You talk like a businessman rather than an artisan or a creator of things, so perhaps your mindset is different but what happens when the AI breaks something and nobody can fix it because they lack the ability to think about the problem constructively or understand what the problem actually is.
As I said, I can write programs in assembly language. I have actually done so, small trivial ones. I’m not a businessman, I’m a programmer. But I use compilers basically all the time because it would be ridiculous not to.
If an AI is able to break something in a way that no human can fix then I suppose that’s a sign that AI has exceeded human capabilities. Do you think it’s there yet?