> Except one is recoverable from, just as we eventually recovered from dotcom.
"Dotcom" was never recovered. It, however, did pave the way for web browsers to gain rich APIs that allowed us to deliver what was historically installed desktop software on an on-demand delivery platform, which created new work. As that was starting to die out, the so-called smartphone just so happened to come along. That offered us the opportunity to do it all over again, except this time we were taking those on-demand applications and turning them back into installable software just like in the desktop era. And as that was starting to die out COVID hit and we started moving those installable mobile apps, which became less important when people we no longer on the go all the time, back to the web again. As that was starting to die out, then came ChatGPT and it offered work porting all those applications to AI platforms.
But if AI fails to deliver, there isn't an obvious next venue for us to rebuild the same programs all over yet again. Meta thought maybe VR was it, but we know how that turned out. More likely in that scenario we will continue using the web/mobile/AI apps that are already written henceforth. We don't really need the same applications running in other places anymore.
There is still room for niche applications here and there. The profession isn't apt to die a complete death. But without the massive effort to continually port everything from one platform to another, you don't need that many people.
The idea that AI is somehow responsible for a huge chunk of software development demand is ridiculous. The demand for software has a very diverse structure.
"Dotcom" was never recovered. It, however, did pave the way for web browsers to gain rich APIs that allowed us to deliver what was historically installed desktop software on an on-demand delivery platform, which created new work. As that was starting to die out, the so-called smartphone just so happened to come along. That offered us the opportunity to do it all over again, except this time we were taking those on-demand applications and turning them back into installable software just like in the desktop era. And as that was starting to die out COVID hit and we started moving those installable mobile apps, which became less important when people we no longer on the go all the time, back to the web again. As that was starting to die out, then came ChatGPT and it offered work porting all those applications to AI platforms.
But if AI fails to deliver, there isn't an obvious next venue for us to rebuild the same programs all over yet again. Meta thought maybe VR was it, but we know how that turned out. More likely in that scenario we will continue using the web/mobile/AI apps that are already written henceforth. We don't really need the same applications running in other places anymore.
There is still room for niche applications here and there. The profession isn't apt to die a complete death. But without the massive effort to continually port everything from one platform to another, you don't need that many people.