Faster, But Only If It's Better

Many of the software engineers I know and respect accept that generative AI can push out a lot more code than humans have previously been capable of, but that more doesn’t necessarily mean better. In the context of DNSimple I keep thinking about being faster with AI but only if it makes things better. Better can mean a lot of things though, and that’s the part where the choices we make as humans in the mix can make a big difference.

Better could mean more features that help you, your users, your customers. This could come with the tradeoff of lower quality when the features come out quickly with less real usage, or a worse experience because of confusion brought on by lack of coherence between different features in a system.

Better could mean better for your team, less pager alerts, more system stability. This could come with the tradeoff of less features that you or your users need for a while. For a business this could mean less growth.

Better could mean better architecture, which makes a system easier to understand and improve. This could come with the tradeoff of beautiful sand castles that no one wants to play with.

Better could mean more experiments and prototypes that help you find the real useful ideas amongst the mess of those that are less useful. This could come though with the tradeoff of wandering around looking for the next valuable thing rather than focusing on making one thing the best it can be.

All of these versions of better can be valid, but not necessarily valuable right now. The real value of our understanding of other humans and of the environment around us is knowing which version of better is the version we should be focusing on now, and knowing when it’s time to switch the focus to a different better.