Where the Work Goes
Short term pain. Long term expansion.
If 1 engineer can do what 50 did, what happens to everyone else?
Short term: displacement. That’s real and it’s happening.
Long term: we’ll need more engineers than before.
The Bottleneck Was Never Ideas
It was cost.
There are two levels of unrealized software:
Level 1: Ideas people had but couldn’t afford to build. The “would be nice” backlog that got deprioritized forever. Every product manager has a list of features that never shipped because the engineering cost was too high relative to the value.
Level 2: Ideas that were never even seriously thought about. Whole new categories of software that weren’t in the conversation because they seemed impossible. Nobody was even imagining these.
Level 2 is bigger.
What Gets Built
The cost of building just dropped 50x. That doesn’t mean we build the same things cheaper. It means we build things we never considered before.
Personal files. You’ll have a file that software reads to customize to you. Not your preferences in each app. A portable identity that all software can access. Everything adapts to who you are. This is infrastructure, not a feature.
Trained personalities. Creators train AI versions of themselves. Tim Pool trains an AI that knows his takes, his style, his voice. That AI creates content across video, audio, text, social. All at once. While he sleeps. More efficient. More consistent. Lets the human have a life.
Dynamic content. You can’t personalize static pre-rendered content. If you want a movie that adapts to you, the scenes have to be generated based on who you are. Generation isn’t just cheaper. It’s the only way to make content truly personal.
Talent multiplicity. Brad Pitt can star in 8 movies simultaneously. The licensed personality becomes the asset. Talent isn’t bottlenecked by physical presence anymore. The economics of entertainment flip.
Why This Is Inevitable
Hyper-personalized software is more complex than one-size-fits-all.
Every user gets a different experience. The software has to understand identity, adapt in real-time, generate content dynamically. That’s harder than shipping the same thing to everyone.
Companies will differentiate by how well they personalize. The efficiency gains from AI development get reinvested into more ambitious software.
The Conclusion
Short term pain. Long term expansion.
Once people see what’s possible, they want it. Once companies see competitors personalizing, they have to match it.
More engineers needed. Just building different things.