Getting Stuck in the Past

Oct 15, 2022

Why do we miss new technology waves? Two reasons that stop us from continuous learning:

Getting stuck in local maxima: We make a rational choice to learn a new, efficient technology. It continues to be the best for some time – more features and fewer bugs than new technology. At some point, alternatives get good enough that newcomers are indifferent, while old-timers prefer their old techniques. Finally, new technology surpasses the old. For some time, it still makes sense to continue using the old technology – the activation energy to learn the new tools and methods does not surpass the delta in efficiency between new and old.

Failing to update Bayesian priors: Given an uncertain event (whether a tec, we can make an educated guess (prior probability). However, we should update our guess (posterior probability) as we observe outcomes and take in new information. When we fail to do this, our guesses don't improve and can get substantially worse when the new outcomes drastically change.

Take, for instance, reasoning like Why I Don't Use Netscape (1999). To this day, some developers still bemoan sites with JavaScript. In the past, maybe it was a decent rule of thumb – early internet users and academics tended to have simple static sites, which might be the most information dense.

But those who do not continue to grow with the times are doomed to fall behind. Single-page applications can be faster than server-side rendered pages for complex apps. Hosting a site on nginx deployed on a VPS can be slower than edge deployments.