Billions of dollars are being invested in the artificial intelligence (AI) arms race. Regardless of whether AI is applied to help cure cancer, discover new principles in physics, or create robots for building space stations, it ultimately comes down to the cost per token, dollar, and watt.
In a discussion on the BG2 Podcast with Brad Gerstner, Microsoft CEO Satya Nadella highlighted the Jevons Paradox, which suggests that as the cost of AI is reduced and it becomes more commoditized, it becomes a major driver of GDP growth across the board.
Nadella pointed out that challenges arise from the availability of electric power and the ability to complete buildings near power sources quickly. Without addressing these issues, many chips may remain idle in inventory, unable to be utilized.
Despite these challenges, Nadella expressed Microsoft's satisfaction in investing in OpenAI and praised Sam Altman's vision, as well as their skill in demonstrating and executing the technology effectively.
Altman referred to the collaboration as one of the great tech partnerships.
Gerstner emphasized that their partnership underscores the importance of choice in health and resilience.
OpenAI can deploy models on Azure until Artificial General Intelligence (AGI) is verified and validated.
Altman anticipates that developments will involve several unexpected twists and turns.
Regarding compute constraints, he noted that while it's viable to discuss energy demand at specific price points, one cannot consider energy demand without accounting for varying price levels.
Altman also mentioned that if the cost of computing per unit of intelligence were to decrease by a factor of 100 overnight, usage would likely increase more than 100-fold.
There are numerous aspirations people have for computing that currently do not make economic sense due to existing costs. The key is maximizing the economic value that society derives from these advancements.