Early Adoption Isn't Always an Advantage

I think it’s misguided in any new technology cycle to think that early adoption is always an advantage. You might see the most benefit by taking a cautious approach and leaning in once the maturity is there.

There’s an argument to be made for this with AI. There is enormous hype and FOMO around what is “technically possible”, ignoring the lag between what’s technically possible and practically possible. If you force adoption at the point of “technically possible” you’ve probably committed too early. Think about the gap between AutoGPT and Claude Code, for example. The former could “technically” solve a problem, but in practice the models, ecosystem, social conditions and maturity needed to move forward a bit. If you had tried to adopt AutoGPT like you adopt Claude Code, you’d probably have such a mess on your hands that you are actually prevented from adopting fully at the right time!