What happens to Artificial Intelligence (AI) and how it will deeply affect our lives has been a well-discussed question. Even how much it will evolve has now become a central part of the conversation. Now, the US President Donald Trump’s White House AI and crypto Czar, David Sacks, has put his insights into the evolution of AI in the future. He predicted that AI will evolve 1,000,000x in four years, meaning it will 1 million times! Isn’t that a shocking probability?   

Speaking on a podcast by The All-In Podcast, “I would say the rate of progress is exponential right now on at least three key dimensions”, with this statement, he explains how three significant dimensions, such as models, chips, and compute, define the progress of AI. 

The models, the chips, the compute

Starting with the dimension models, David Sacks finds that the algorithms used in AI are progressing with speed, accuracy, and usefulness each year. At first, AI used basic language models — LLM chatbots — to process conversations, typically asking questions and generating answers. 

Next came reasoning models, which provide more problem-solving answers rather than the basic question-and-answer style. The recent evolution is AI Agent, an autonomous AI model using less human input, designed to automate workflows and tasks. 

Meanwhile, each generation of chips is perhaps “3 – 4x better than the last”. Engineers are now figuring out how many chips can work together to bring better output rather than focusing on the performance of a single chip. He pointed out the example of the NVL72 chip, which creates better performance by using many chips.       

Coming to the third dimension, compute, Sacks hinted at the number of Graphics Processing Units (GPUs) used in datacentres, specifically mentioning the Grok feature on X, which might have had 100K GPUs. He expects that the Grok now holds 300K GPUs!

Essentially, the podcast platform emphasized how the crypto Czar calculated the 1 million exponential growth in AI.

 In his opinion, the three dimensions are “improving or scaling at a rate of 3-4x a year.” That said, the expected progress is 10x in two years. 

“Where people don’t understand exponential progress is that if you’re getting better at 10x every two years, that doesn’t mean you’ll be at 20x in four years. It means you’ll be at a 100x,” stated Sacks.

Multiplying the three dimensions: 100x models,  100x chips, 100x compute results in 1,000,000X AI, which means a 1 million increase!

Sacks ultimately stressed the point that the impact of AI is massive, however, many people fail to appreciate or recognize this reality.