Thinking in public
Notes on how AI changes operating models, why APIs become business capabilities, and what helps product and engineering organisations move faster.
By now even the skeptics reluctantly accept that it's indisputable that vibecoding / AI-assisted coding makes it much faster to ship software. But it's equally true that the speed does not make the value last. I'd argue it makes it easier to chase the wrong goal because "it's so easy to progress"
When generative AI and the frontier LLMs started to become practical, I pushed every team in my portfolio to build an AI feature. I wasn't stressed about what they were building, but I pushed hard to get them all to build something. This wasn't because I thought adding on an AI feature would change the business or even make us money. In fact, we never charged customers for any of them as an added feature. I did this for the sole purpose of forcing people to engage with this new technology.
A little over a year ago I shared a presentation internally with our CEO, execs and other stakeholders urging action as AI was rapidly accelerating the time to market for new products, creating a threat to incumbents. I outlined several moats I believe we could exploit and I still believe they are valuable space to occupy - for example domain knowledge, compliance capabilities, massive datasets. I did miss one key moat that hindsight allows me to appreciate more fully: Distribution.
If you thought coding was just about clean architecture and efficient logic, vibecoding would like a word (preferably in monospace font, glowing in soft lavender neon, backed by a lo-fi beat). What started as an aesthetic subculture blending programming with mood, visuals, and audio has morphed into something weirder, richer, and more creative.
