No GenAI Economic Boost without Serious Digital Investment but Who Pays?
Economists analyze GenAI and take on the hype.
The adults are in the GenAI room and pose a fundamental question that needs to be asked: Who will pay for the digital investment required for the GenAI boom?
This is a great question that counterbalances the never-ending hype surrounding GenAI, which is paid for by the deep pockets of big tech and consultants.
In this case, the adults are ING economists who simply have the temerity to point out that much money needs to be spent before the significant GenAI economic boost is realized.
Did any of the consulting company reports we’ve seen ever mention this?
👉TAKEAWAYS
GenAI obstacles:
🔹 Big Tech is dumping money into building infrastructure (see graph above), but this isn’t enough to boost the economy, given that it has to be rolled out everywhere.
🔹 Data centers require electricity and fresh water, which requires further investment by public utilities, which are already struggling to keep up with societal demands.
🔹 Human capital costs in reskilling employees to match new AI-based jobs.
🔹 Where will the chips come from? The current supply is limited.
🔹 Data issues with copyrights and bias remain a problem and will delay roll-out.
🔹 Regulations such as the European AI Act will also impact GenAI as data and model transparency will become far more critical.
🔹 GenAI software is expensive, which will slow the spread of the technology, until the cost drops appreciably.
🔹 Hallucinations will reduce GenAI safety.
👊STRAIGHT TALK👊
The point of this post isn’t to say that the GenAI economic boost won’t materialize; it most certainly will, but let’s be real about it: It isn’t coming overnight.
GenAI is considered “General Purpose Technology,” which puts it at the level of the steam engine in its importance and ability to increase productivity.
General-purpose technologies have three key characteristics: capable of rapid improvement, pervasiveness, and complementary innovation.
GenAI is intended to be ubiquitous as a general-purpose technology. This means that big tech and corporate users must provide computing power for everyone.
Now consider this: we’re currently buying 90% of our AI computing power from one company, NVIDIA! How does that contribute to making GenAI “pervasive?”
Meanwhile, to put AI’s electric use in perspective, four Dall-E drawings or 300 ChatGPT questions consume the electricity of one cell phone battery.
So even if we buy the chips, can we spare the energy for societal-wide GenAI distribution?
GenAI is fantastic tech, and I love it!
That said, I won’t surrender to the hype and think these economists’ message of “Who pays?” is important.
Thoughts?
Thank you to all my subscribers who have been sharing Cashless! You are now the No. 1 driver of subscriptions, and prove that word of mouth still works better than X!
Join the community by subscribing! You’ll be glad you did!
