Ai-Coding

Earlier this week I was at the TailscaleUp event in London and one of the panels raised that we are losing sight of how long work actually takes. This also connected with the idea that in this new AI future there will be 100x more builders. That means there will be 100x more people who’ve never felt the gap between a working prototype and a production-ready product, and AI has vastly reduced the barriers to building, but there are still some hard things that need to be understood. ...

Two weeks ago I built an Android application with voice commands while running around an athletic track in the shadow of the Olympic Stadium in London. So when I decided to try building a web app en route to a big group weekend away, it did not feel as far fetched as it once would have. The first commit went in from the Eurostar terminal, before I had boarded. I did not have a laptop, and was going to have to do the whole build with just my phone. Recent hackathons, and a few sites of my own, had taught me what works and what does not, and importantly what I could accomplish with just my phone. The build ran on the train until the channel tunnel killed the connection, and I picked it back up on the Métro in Paris. The only real time I needed to stop was to arrange the Cloudflare access the agent needed. Before walking to my hotel I sat down on the Métro platform to run through the online setup, because it was fiddly, and the Métro had been extremely hot the whole way so far. After this I could let it start building reusing methods for setup from previous projects and my existing domain to get a useable subdomain. ...

I’ve spent the last few weeks building across a few different AI coding tools, and I stopped paying much attention to the model underneath. You spend enough time around AI and you start to believe the model is the part that matters, since it’s what everyone talks about and all the benchmarking is against. But build across enough tools in a short stretch and a different picture shows up. The model is closer to an interchangeable component, and the setup wrapped around it is what decides success. ...

The first time I used Kiro was at the AWS Summit in London earlier this year. It was a hands-on workshop, so on the rails run through to show what the tool could do. Everyone was converging on the same idea at that point, that the spec was what mattered, and it all felt a bit contrived from Kiro just formalising it. The first time I used Kiro was in a guided workshop at AWS Summit London. ...