Amjad's Law: the return on learning to code is doubling every 6 months.

Some people naively assume that the better AI gets at coding, the less value there is in knowing how to code. It's actually the opposite: as the AI gets better, the people who know how to code get more and more leverage.

From our recent podcast with @amasad:

coding is the new babysitting

For someone with 5-10 hours available per week, what's the most effective approach to learning how to code?

I see an opportunity for a team to come into an org and help that org save loads of money by helping internal IT teams learn to leverage AI tools to write custom solutions, reducing the over-dependence on cloud and really niche SaaS solutions. I have an example considering a community college. They pay thousands of dollars a year for a SaaS based tool for staff to clock in and clock out. I used ChatGPT and Claude to write an app with the same functionality in a couple hours on a Saturday out of curiosity. Its super empowering and disruptive. Theres still really only a select few that realize the power of these AI tools to crush the learning gap for coding/engineering.

As someone who started today and already was able to build a scheduled reminder app I agree that as AI improves at coding, the value of knowing how to code increases, giving coders greater leverage.

This was awesome to hear

@garrytan What exactly do we mean by learning to code, though? Since it’s not a binary outcome, the real question is, what are the minimal/essential coding skills to benefit from Amjad’s Law? @amasad

Agree with this

amjad’s law is true. i’ve seen so many narratives saying learning to code is not useful anymore. and i definitely don’t think so.