Once the hype around AI settles and the true cost of using AI tools is fully understood, an interesting question emerges:⁣⁣
⁣⁣
What does the steady-state impact of AI look like - especially for the technology industry?⁣⁣
⁣⁣
I’ve been thinking about this for a while, and the most compelling lens I’ve found is an analogy with industrialisation.⁣⁣
⁣⁣
Industrialisation dramatically reduced the cost of production, while the cost of human effort steadily increased. Over time, replacing a product became cheaper than fixing it. Cars, phones, laptops - almost everything today is designed with a short lifecycle. Manufacturers actively phase out legacy versions within 3-5 years, making repair or long-term maintenance economically irrational.⁣⁣
⁣⁣
Product lifecycles that once spanned 15-20 years are now closer to 3-5.⁣⁣
⁣⁣
This raises an uncomfortable but important question for software:⁣⁣
⁣⁣
Will AI make rebuilding software cheaper than fixing or upgrading existing systems?⁣⁣
⁣⁣
If that turns out to be the trajectory, the implications are profound. Software systems may become increasingly disposable - rewritten rather than refactored, replaced rather than evolved.⁣⁣
⁣⁣
In such a world, the most critical architectural layer will not be the application, the UI or even the business logic.⁣⁣
⁣⁣
It will be the data.⁣⁣
⁣⁣
Think about how businesses once maintained physical ledgers. The format was simple, durable and readable by any qualified professional, regardless of who created it.⁣⁣
⁣⁣
In the future, data will serve the same role - the only enduring business context.⁣⁣
⁣⁣
Everything else may be transient.