I have been a software engineer for the past 12 years now. I recently caught the AI bug and have decided to go all in on AI.

While the thought of making this shift is intimidating at times, it has been a few years in the coming. At this point I am more okay with learning AI and falling flat on my face than continuing to wonder how wonderful it would be if I somehow became an expert magically.

The current landscape is dominated by researchers and PhD scholars who have spent more than a decade studying and pondering over the minutiae of AI. How can someone like me become an expert in AI without going through the same process.

There are 2 pieces of advice that resonate with me the most.

  • Jeremy Howard of Fast.AI says that the best way to learn is to “Train lots of models”.
  • Andrej Karpathy says that 10,000 hours of dedicated efforts is a sure shot way of becoming an expert.

Now, 10,000 hours = 1,200 days = 60 months = 5 years. That means it would take 5 years of almost full time work to become an expert. While this might work for a uni student, it’s not very feasible for someone in my position.

My plan is to use the same approach to learning AI as the one I used to learn software engineering. And to use the mistakes made over the years to optimize the approach.

What helped me the most in the early days was building personal projects. Projects for my self, projects for my father, projects for my mother, projects for my friends.

Projects, projects, projects.

Over time, I dropped this approach as I became complacent with my learning, but this was probably the one thing that taught me the most.

So the first rule of AI club is to always be working on a project.

One mistake I made while learning SWE is that I rarely shared the things that I learned online. That is something I will change this time.

So the second rule of AI club is to reguraly share the learnings online.

Another mistake I made while learning SWE is that I always used (or abused) open source software but never really contributed anything. I know for a fact that I would have learned way more if I had made a concious effort to become an active contributer. That brings us to the third rule.

Find open source projects that are interesting to you and contribute to them.

And just so that we are on the same page, contributing doesnt mean going to the readme page and fixing a typo here and a punctutation there.

To summarize, the 3 rules of AI club are,

  1. Learn by building projects
  2. Share what you learn everyday
  3. Contribute to open source projects.