Papiha Joharapurkar

ai spending and diffusion

Taking a break from AWS for today. I saw this tweet a while back and thought it was interesting and I'll try to put down my thoughts about this today - scrambled as they may be.

Even though we all think AI is so ubiquitous and life changing and it must be applied everywhere and there's nothing left for humans to do ... I really don't think that's the case at all. According to this graph, let's assume that the top 1-10% are the real AI heavy users that are good at their craft of learning how to turn AI into something that's useful/productive. That means there's the remaining rest of the world that has little idea about the potential of AI. The bottom 50%'s AI usage potentially limits to just thinking of ChatGPT as being a better-search tool than Google. In the bottom 50%, there's plenty of opportunities remaining to be solved with AI.

I heard in a podcast that after the camera was invented, it was used initially for filmmaking but only if the camera was stationary as the sets changed, scene after scene. After 10 years of this, only then did technology catch up to the point that they were able to make cameras that moved easily and were not as bulky - which then revolutionized the field of filmmaking creating whole new professions.

With every era of a new technology being introduced, there will be time needed for it to diffuse into the economy. With AI just being introduced, there's still years and years of exploration of AI being applied. In this world of AI, where everything is new and all of us learning to figure out how to work - then it remains that the biggest challenge would be our imaginations and figuring out how AI can be made helpful in every part of the world.

In my work, I think about how even though we have agents today that we have created to mimic or replicate how humans work (similar to how early film captured theatre-style) maybe in the next phase, that will change. Maybe we'll identify new ways for agents to work that are more well-suited for agents and less for humans. There's still so much innovation to be had I think.

All daily logs