The rise of AI is unquestionably upon us, early days, all happening sooner than expected. Given how big of a phenomenon it is, it's suprising how few people foresaw it, as recent as 2019.
Been curious, why there were so few predictions a few years back. So figured, would just ask ChatGTP, if anybody predicted in 2019 that AI would be where it is now.
It says a few people did make predictions that were close, but most forecasters underestimated the eventual progress, especially the combination of improvements in scaling laws and faster compute power, by relying too much on historical rates of progress.
The few who came close, include Sam Altman, Deepmind & OpenAI researchers and futurist Ray Kurzweil.
AI Will Be The Most Powerful Technology of Our Lifetimes
Of course, now, it's crystal clear to all of us that AI will be the main technology story for decades. The following quote from the below talk Eric Schmidt (former CEO of Google and all around wiz) recently gave captures how pivotal of a technology AI is, "This is the most powerful technology that will be invented in our lifetimes. It’s as important as things like electricity, heating, air conditioning, automobile, things like that. Maybe more so."
Geopolitical Implications: The AI Race Between US and China
What was long the geeklore of sci-fi books and movies is now the real deal, led by crazily outsized investments from the largest tech companies in the world, along with an emerging AI race between US and China focused on leadership. A recent Brookings Institute report predicts that, "Whoever leads in artificial intelligence in 2030 will rule the world until 2100."
It's largely a battle between the private sector in the U.S., home to the largest technology companies in the world. With the top 10 or so companies having valuations that are larger than the economies of all but ten-or-so countries in the world. Verse China's centrally planned and financed system of innovation. Each of which offer advantages and disadvantages.
The capitalist U.S. system directly rewards advancement, while the Chinese system can better scale the infrastructure (including the energy infrastructure that AI Data Centers require) and also has the advantage of a government focused on the long-term securitization of critical mineral resources.
Massive Energy Demands of AI Datacenters Require Acceleration of New Energy Technologies
The rapid growth of artificial intelligence is directly linked to a surge in energy consumption. Data Centers, already significant energy consumers, are paled by the demands of AI Data Centers, which require substantially more power than their traditional counterparts.
This increased energy demand is particularly visible during data-intensive training runs, the critical phase in which new learning models are developed. The peak power consumption of AI Data Centers is not only significantly greater than that of traditional data centers, but also far less predictable, posing new challenges for energy infrastructure and resource management.
The escalating electricity demand needed for AI presents a critical challenge, but also a massive opportunity to accelerate:
Microgrids: localized electrical networks that can operate independently from the main power grid.
SMRs: small nuclear reactors which are are built in factories and shipped to sites for installation.
Offshore Wind Farms: massive deep at-sea energy generation, which can provide a greater energy capacity than other energy solutions.
AI will require advancement and capacity from each of these solutions, in order to meet the ever-growing energy demands that AI data centers require.