Existential crisis

Last few days haven't been the same for me.

I have been following AI since my college days (around 2019) and keeping myself updated for years, from seeing how AI models build a house price predictor, to generating images, to writing code. Everything was wow but at least understandable. I used to think about how these models take data and build a nice mathematical approximator. I knew progress was inevitable but never expected this much. Coding models were built, they were powerful, and they were better than me, but still digestible.

A few weeks ago, GPT Astra was released. One of the main advances it did was making wonderful 3D models. In my college time, I spent most of my time working in Blender, building 3D models for fun and working in Unreal Engine. For the last 3 years, I haven't been doing much 3D modelling, but it was really unsettling for me when I saw Astra building really great 3D models quickly, which would have taken me weeks or months. Though it doesn't follow the proper artistic approach (like it makes a lot of objects for a single asset, makes the collection messy, and breaks some rules we generally follow), it gets the end result. It kind of uses math to do stuff. I am sure in future models it will fill this gap as well.

Then another news came 2 days ago: GPT solved the Navier-Stokes equation. I know there is controversy and OpenAI mostly trained on mathematicians' data, which directly helped to solve this. This is not good from OpenAI, but if we think about what an AI model achieved, it is a way bigger thing than the controversy.

I was reading a few books about the history of quantum physics and noticed that Europeans were way ahead in this area than anyone else, and the most likely reason was a cluster of talented people, knowledge sharing, and competition.

For the current stage of AI, they are already super capable. Like in Europe, where a lot of wonderful scientists were thinking and working on one problem, getting some insight, and based on that insight another scientist would make a discovery. Here in AI, they can give swarms of agents which will work in different dimensions and pass that insight to other agents, and will do discoveries like they did solving Navier-Stokes.

Now I am wondering what else AI can't do. Probably in a few years (who knows, maybe in a few months), they can do everything in creative human endeavour. Sometimes I ask myself, what will be our goal in the future? Just living, or is there no meaning? Maybe we humans aren't special, just an equation in God's plan?