The world is slowing down. America is not. Huang's law is here to save the day.

Last generation the US was coasting off of Moore's law and the Internet.

The Internet bubble burst in the 90's but the top companies still scaled up consistently that were involved with the internet.

Similarly this generation is running on AI and Huang's law, which declares that GPU capabilities will triple every 2 years. Which is 1.5 times faster than Moore's law.

We are definitely in an AI bubble, the good news however is that it just means that out of the 1000 companies that are advertising themselves as AI focused companies, 900 will probably fail, 90 will get acquisitioned by the top 10 and the top 10 will come out consistently growing further in the years and decades ahead. Just like what happened with the internet bubble.

The good news with Huang's law is that hardware scaling, AI development, and GPU scaling have formed a sort of loop where improvements in one lead to improvements in the other two which loops back to improvements in another one leading to improvements in two so on and so forth.

With the US leading in space launches as a viable business, improvement in robotics, AI/GPUs/ Electronic hardware looped together and consistently scaling up, and the slow scaling up of the US consumer product industry, we can expect the US to keep growing at a rate between 2-2.5% year on year for the next few decades.

Chinese GPUs being 10 years behind US GPUs means they are 243 times weaker than US GPUs, which means Chinese GPUs and semiconductors only have a market in the middle income and developing economies, similar to Samsung phones having a market primarily in the middle income world.

Unless the Chinese figure out fusion energy, the Chinese economy is going to stagnate and decline long before the US economy.

AI in the US will probably double in scale every two years. It just won't scale up at the rate that Chatgpt did in the starting years. Expect AI to keep scaling up at a rate where you don't feel like AI is replacing jobs but AI is indeed replacing jobs.

The next stage of AI development isn't scaling up the amount of data, but improving AI data filtering capabilities, then scaling up those more compressed models to the maximum level, then fine tuning that and maximizing it again, and doing this repeatedly in a loop. ( Personal opinion, based on how people managed to make something as functional as ChatGPT2/3 with far fewer parameters, logically you should then be able to scale up that new more efficient model for even better results then again improve efficiency down the line ).

Huang's law should last for 2 generations, taking into account how long Moore's law lasted for.

The primary challenge for the US:

It has reached a point where only the US can afford its own products when it comes to the top tier medical, technological, robotic equipment that it makes. The US is thus divided into two markets, one internal where only Americans can afford American products, and one external where Americans try to use their superior tech know how to out compete middle income nation states in producing the same products but cheaper through new techniques of production ( Examples - vertical farms, synthetic food alternatives, lab grown meat, etc ).

Conclusion:

The US is growing at a scale that is separating the US from the rest of the world permanently. I do not know the repercussions of this for the world.