Trillions will be ploughed into LLMs looking for a moonshot “productivity miracle”, the only thing that could kick the can for the collapsing fiat economy.
AI companies will be awash with venture capital but LLMs will not birth AGI, and these companies will largely produce little of value before that bubble bursts.
AI could well be the peak of fiat. 
I think all of the major improvement and huge benefits for LLMs lie not in making the models better (which will only give marginal benefit) but in figuring out the most clever ways and areas to integrate them.
Heading into a bull market for image and video diffusion, but a bear market in LLMs, imo.
I agree but it’s not how Silicon Valley VC works. It’s not how the Government with a money printer which desperately needs a “productivity miracle”, works.
They are going to firehose money at AI companies and spin a tale of how it’s helping the West outcompete China. It won’t matter how realistic their usecases are, it won’t matter that probably the best case for LLMs is just better integration into existing processes, all the vested interests will convince themselves that with enough money “AI” will actually fix everything and they will triple down.
Like your recent pod “the next big thing will look like a toy” - this capital won’t be where AGI will actually emerge from because the “experts” will be busy chasing the LLM deadend having refused to understand its limitations and the result I’m anticipating is the biggest misallocation of capital in history.
Agree about the fed, but I don't agree that LLMs need to lead us to AGI in order to be useful. They already boost my productivity
