Why AI will increase the size of the software market

My conviction is that there will be more AI code generated by more people without any training or experience designing high-optionality software.

I think this is a good thing.

If only 1% of vibe coded projects generate any income, but now everyone can be a vibe-coder, we’ll still see a huge expansion in the market of software.

One of the hardest problems when building software, “will anyone buy this / is this even worth building?”, can now effectively be validated at far cheaper and faster than ever before thanks to AI.

I think we will see more people with apps and working software in the hands of users that can make money today, but will be incapable of making 10 dollars tomorrow. To evolve to that level will require someone who knows what they’re doing when it comes to increasing the optionality of that software.

Notice, I’m not talking about someone who can write code. Because that is no longer the bottleneck.

It’s understanding the code, the systems it’s deployed upon, and crucially the business context and the new future state that the software is trying to achieve.

AI will aid software engineers in achieving this, but that ability to envisage a new future and work creatively will remain human.

As regular reader Peter, aka my Dad, writes in:

Humans make choices - using a judgement, intuition and with real world existential skin in the game - because humans are always evaluating the mortal costs of their actions but instinctively open to improbable possibilities of the future.

AI makes decisions - using replication, logic, inference and with a virtually infinite capacity to backtrack - because AI is the specifically designed to overcome the mortal costs of its actions and algorithmically programmed to focus on past probabilities.

So optionality is human.

Couldn’t put it better myself.