Overview
This piece argues that most people can’t recognize software-shaped problems even when AI makes app creation effortless. The real barrier isn’t technical ability but problem recognition - programmers see automation opportunities everywhere while others remain blind to these possibilities.
Key Arguments
- Most people’s problems are not software-shaped, and they won’t notice even when they are - even with instant app creation capabilities, people will struggle to identify use cases: When told they can create any app instantly, people respond with ‘Cool! Now I need to think of an idea’ and then forget about it entirely, suggesting the bottleneck isn’t technical but conceptual
- Programmers are trained to see everything as a software-shaped problem while others are blind to automation opportunities: Programmers automatically think to automate repetitive tasks (like renaming files via terminal commands) while non-programmers painfully do the same work manually through clicking and copy-pasting
- We are limited by solutions we were never taught to see - like asking for faster horses instead of dreaming of cars: The analogy suggests people can only imagine incremental improvements to existing methods rather than paradigm shifts to entirely different approaches
Implications
This means that even as AI democratizes app creation, the real challenge is teaching people to recognize when software could solve their problems. Simply making coding easier won’t unlock mass adoption - we need to change how people think about problem-solving itself.
Counterpoints
- Many successful no-code/low-code tools have found widespread adoption: Tools like Zapier, Airtable, and website builders suggest that non-programmers can identify and solve software-shaped problems when the tools are intuitive enough
- People may develop problem recognition skills through exposure: As AI coding tools become more prevalent, users might gradually learn to identify automation opportunities through experience and examples