I don't think any of this makes much sense. The problem that this article ignores is that people build different things with AI, not the same things but faster. They build more superficial things, more throwaway things, things that are tailor-made for chasing the buck, and it just doesn't feel genuine or interesting.
It was much the same with search. Google search allowed us to find things faster at first...but over time websites responded to that and started doing SEO crap. Now almost every search results in a billion websites that say the same thing, but shallowly. Yes you can still find some gems but was it worth it?
AI makes things faster, but now it means we can target superificial needs very quickly...and so a new breed of crap software will be built with it. As long as there's a dollar sign at the end of the road.
People respond to incentives: Why do a hard thing when I can do an easy thing?
During the hype cycle, making "x for y" is easiest to get funded and exit.
Once the hype crashes and we're in the trough of disillusionment, then people are bearish. So founders either move to the next hype cycle or figure out a more durable business from the ashes
To be honest I don’t believe this is the case. Admittedly, I don’t have any experience with the best of the best models, only up to clause Opus. AI accelerates coding for myself a hell of a lot, but frankly the amount of time I spend coding complex programs feels about the same. I think there’s still a bottleneck, and it has been reiterated here before: judgement, testing, and understanding. If AI really accelerated coding, no amount of cultural inertia would be able to hold it back. It’s just not.
I was about to say something similar. The article’s premise is solid as long as we accept the assumption that AI is actually fundamentally changing the velocity and methodology of coding.
To me, so far it seems like it is and it isn’t simultaneously. Some things are going at insane paces. I see things like reverse engineering and emulation breaking barriers. But we don’t see things like the quality and complexity of highly architected systems getting much better, maybe even getting worse as management pushes velocity onto employees.
I think what this shows is that AI is great at tedium and brute forcing, but it’s not actually good at making substantially better software with less of that “rationed” expertise. At least, not yet.
... by improving model efficiency so AI tokens are "too cheap to meter" or alternately not that much more expensive to operate than conventional sorts of software?
It's a different view of scaling. I mean a lot of people have this absurd idea of "super intelligence" which is like the cover of this book
i mean what does "intelligence quotient 10,000" even mean? On the other hand it's very easy to imagine a world where we have systems like the best ones we have now but we can really afford to use them.
I don't think any of this makes much sense. The problem that this article ignores is that people build different things with AI, not the same things but faster. They build more superficial things, more throwaway things, things that are tailor-made for chasing the buck, and it just doesn't feel genuine or interesting.
It was much the same with search. Google search allowed us to find things faster at first...but over time websites responded to that and started doing SEO crap. Now almost every search results in a billion websites that say the same thing, but shallowly. Yes you can still find some gems but was it worth it?
AI makes things faster, but now it means we can target superificial needs very quickly...and so a new breed of crap software will be built with it. As long as there's a dollar sign at the end of the road.
Before we were spending 10:1 on features:quality . I worry that ratio will remain, but at 100x or 1000x the velocity.
There's no reason to think that companies are suddenly going to be spending 1:1 features : quality because both are faster / cheaper .
So expect the low quality to continue, and naturally get worse, because complexity is going up 1000x
People respond to incentives: Why do a hard thing when I can do an easy thing?
During the hype cycle, making "x for y" is easiest to get funded and exit.
Once the hype crashes and we're in the trough of disillusionment, then people are bearish. So founders either move to the next hype cycle or figure out a more durable business from the ashes
To be honest I don’t believe this is the case. Admittedly, I don’t have any experience with the best of the best models, only up to clause Opus. AI accelerates coding for myself a hell of a lot, but frankly the amount of time I spend coding complex programs feels about the same. I think there’s still a bottleneck, and it has been reiterated here before: judgement, testing, and understanding. If AI really accelerated coding, no amount of cultural inertia would be able to hold it back. It’s just not.
I was about to say something similar. The article’s premise is solid as long as we accept the assumption that AI is actually fundamentally changing the velocity and methodology of coding.
To me, so far it seems like it is and it isn’t simultaneously. Some things are going at insane paces. I see things like reverse engineering and emulation breaking barriers. But we don’t see things like the quality and complexity of highly architected systems getting much better, maybe even getting worse as management pushes velocity onto employees.
I think what this shows is that AI is great at tedium and brute forcing, but it’s not actually good at making substantially better software with less of that “rationed” expertise. At least, not yet.
The free holiday will last forever. Make your plans accordingly.
... by improving model efficiency so AI tokens are "too cheap to meter" or alternately not that much more expensive to operate than conventional sorts of software?
It's a different view of scaling. I mean a lot of people have this absurd idea of "super intelligence" which is like the cover of this book
https://en.wikipedia.org/wiki/Supermind_(novel)
i mean what does "intelligence quotient 10,000" even mean? On the other hand it's very easy to imagine a world where we have systems like the best ones we have now but we can really afford to use them.