Marc Brooker:
Building systems remains hard. Can I assume you’re familiar with Amdahl’s Law? That’s what’s going on: a massive speed up on a portion of the problem, but as that portion speeds up it becomes less and less of a contributor to the overall speedup. Lowering the costs of the rest of the problem is work that remains to be done. It’s going to take a long time, because the real world is fully of sticky problems, surprising feedback loops, human stubbornness, and the occasional adversary.
There’s also going to be great value in ideas. Integration and translation are solved problems. Simple analysis, and small scale synthesis are too. But new ideas, real transformative new ideas, remain hard to come by. And, as the lever gets longer, more and more valuable.
Software’s first act is over. The second act won’t go like anybody expects, but I can bet that it’ll be more interesting, more economically valuable, and more mentally stimulating than we can imagine right now.
The overall performance improvement gained by optimizing a single part of a system is limited by the fraction of time that the improved part is actually used.
Amdahl’s law, Reddy, Martin (2011). API Design for C++. Burlington, Massachusetts: Morgan Kaufmann Publishers. p. 210
About an hour ago, Opus 4.6 created VAPID Web Push notifications for a personal web app in 17 minutes while I was spending time with family. It also designed and created its app icon.
The hard part is everything else: Transformative new ideas. Taste and agency. And anything outside an automated feedback loop.
The easy part is converting business logic to code — that cost is quickly approaching zero.
OpenClaw gives a glimpse of the future, but my bottleneck right now is time I can spend in front of a computer for a very good reason. I have yet to sit in front of my Mac long enough to bootstrap it, so much of its power goes unused.
Claude does a marginal job making ESPHome LVGL UI because it’s rare but even more so because it cannot yet see the outcome and enter a feedback loop. Same for low level Vite internals.
But when there’s a clear outcome and benchmark, amazing results are possible like building a C Compiler with a team of agents.
These barriers are coming down slowly which will produce more speedup.
It’s an incredible time to be a developer.
Leave a Reply