Mental toolkit for accelerating with AI

Over the last 12 months, I moved from running a startup to automating our entire product with AI to joining Mistral to do it at scale.

Before it all started, in the middle of the wildest year of my life, I wrote a quick note about how to read, write, and build in the AI-first world. For a year, I stuck to my playbook. I read more than the year before. I built more than the last three years before combined. Combining this with a decade spent building 10+ products, iterating quickly and failing significantly before successes, I had a little teaser of what we are going to go through in the acceleration decade:

In an acceleration decade, how we manage our mind is important.

Intent and Inherent Interest

We need our mind to play in an environment it loves to play in.

I recommend thinking inwards to understand what it is that you love about your job (and I hope you are already in one. If not, I hope you are moving to the industry you love). If you are in an industry because of the love for peripherals (like the tools to do the job) and not the innate act/ result itself, it's going to get tough when AI automates the tooling.

My own personal answer was: Technology is powerful to solve most, if not all, of humanity's hardest problems (physical and spiritual). I want to make it happen with increasing impact (or at least try to do so). AI is going to make this process faster than ever before. The question is what problems excite you first/now?

Building the Thinking Muscle

We need our mind at it's best. One thing to not take for granted is what we need to do to keep the thinking muscle active.

You can't optimize on learning everything in one month and sit idle for the next 10 years. We are all joining undergrad university (again) with the best minds on the planet as fellow students and teachers.

Pick activities that create friction

Those activities that maintain a good distance b/n desire and fulfillment. Buy more physical books. Print the papers you like to read. Use pens to scribble and make notes. Install a whiteboard in your home to problem-solve. For me, these were really helpful.

Reading and writing will be the saviors

I use AI for anything and everything except the actual reading and writing itself. We don't have many ways to ingest information and to force our mind to form patterns. The power of transformer models, if anything, shows us how much words and language are relevant for our cognitive abilities. The less we use this muscle, the more it's going to harm us.

Add the good stuff

Yoga, meditation, walking, cycling, sleeping 7 hours a day, journaling. As many of the routines you can add without stressing yourself.

One easy to-do if you want to try: when you wake up, for the first 10 minutes, just lay on the bed. Don't rush. Don't pick up the phone. Stay, be aware, smile and then start the day. It has a positive effect on my days.

Another thing that worked for me: weekly journaling. I have been doing it for the last 3 years. I have a note that starts with "This is not a to-do list. This is a ‘done’ list. What goes in here: Life, Coding, Reading, Writing, Product, Work, Fun hobbies.". That solves the cold start problem. I do this for work as well. It's easy to spend a whole year and feel like the time was stolen from you. It wasn't. If you are mindful and force your mind to create checkpoints, you will realize it's richer and meaningful.

Minus the bad stuff

Everyone's bad stuff might be different. Avoiding news helped me. Trading the phone for books, swapping color and bright light on laptop for grayscale and night shift helped my eyes.

I decreased external stimulus from Apple Watch so that my mind can spend thinking tokens on positive placebo and interesting challenges.

Lastly, swapping time spent from Instagram, LinkedIn with X and HackerNews was helpful.

Building in Acceleration

Our mind needs to do things to learn. Reading about the new agentic harness is not bad but it's not a substitute to building a harness.

Exhaust all your ideas

If you are builder, this is the best time ever. AI is great tool for the curious ones. Every idea you thought of and held yourself back? Go build it. Build them all until you exhaust all the shitty ideas in your mind. Build them until you are left with the ones that matter and ones that you care about.

Experiment all possible solutions

Over the last decade, I chose a potential solution without any experiments or data. Because we didn't have time to see them all through. AI to some extent helps with this. We can build all the features, variants, and optimizations we need to see which one actually works.

Use AI for automation but not understanding

(It's becoming a happy repetition in this post)

We still need to read, write and do ourselves to retain understanding. Short of inserting chips into our brains, we do not have any other way.

We were writing code to understand code. Now that AI is writing that code, we need to at least read the code that's written. There should be some loop of input -> ingestion -> patter-forming that must happen in our heads repeatedly for us to be able to accelerate meaningfully.

Living Through Acceleration

The intensity and race that we are on will consume our time and steal our perception of it. One place to look for solutions is to look at entrepreneurs.

As an entrepreneur, I had to think differently to be different from competition, I had to iterate quickly to (have a chance to) overcome the competition. And I had to do this for a very long run.

What helps is to have close knit teams: teams that spend life together (eat, banter, challenge, learn, care) and not only work. Share your wins, losses, and experiments. Make everything worthwhile a moment and don't let it become a forgotten memory. Solve problems together face to face as much as possible. And solve them using whiteboards and real speech/words not AI-written/ summarized garbage.

Remember it's going to be a marathon. Not a sprint.

Inspiration for the post

I was reading an article on how to prepare for the next decade (a nice read that I recommend). The author talks about the acceleration decade. The pressure to catch up, mental fatigue and anxiety we are going to have and how we can better prepare for it.

This prompted me to think inwards about how my year went from the last time I wrote about similar topic. Acceleration can be bad, yes, but if you are cautious and can get a handle on it, it can also be the wildest decade of our lives (in a good way).