Generative Artificial Intelligence (GenAI) has transformed the tech landscape in less than 18 months.
Since OpenAI launched ChatGPT, there has been an arms race to upgrade data centers and support it. Cloud Czars are buying Nvidia chips with both hands. This has made Nvidia the 3rd most valuable company in the world, behind only Apple and Microsoft. Microsoft, meanwhile, has surpassed Apple in value, despite being much smaller. It’s seen as the only clear winner in this race.
But Microsoft put $22.7 billion into capital spending during the first two quarters of this year. It’s on pace to blow away last year’s record capital spending of $31.9 billion. It had 2023 revenues of $257 billion. It’s turning over $1 of every $10 it makes into cloud data centers.
Along with the hardware cost come software costs. An AI programmer can earn 50% more than their more conventional cousin. These costs are why even the Cloud Czars have begun laying off good programmers, by the thousands. They are not saving money, they’re reinvesting it.
The Hidden Costs
There are hidden costs to all this, costs we have not yet begun to pay.
The most obvious is what users must pay to access the good stuff. The Linux analyst Nixcraft calculates you’ll pay $20/month to use any of the most advanced models out there. This includes some like Claude and Perplexity you may never have heard of.
In their rush to afford the costs of AI, other Cloud Czars are killing the free web. This is especially obvious in the case of Google. Its search engine is becoming ever-more worthless thanks to layoffs. Google is also becoming a less attractive place to work. Major media is delighting in the comedown of “Big Tech” programmers, but that’s not the problem.
The problem is that tech is developing the same class system that troubles the general economy. There are the Gods who bind and loose and are worth billions, handed to them by investors who often don’t know what they are buying. There is a growing middle class of programmers and professionals choosing which Cloud Czar to align with. (It’s a little like choosing among streaming services.) Then there are the great majority of Web users, who are seeing the Internet resource made worthless by the rush to AI.
Finally, there’s AI itself. Geoffrey Hinton, who pioneered the chatbot technology at Google, has now quit to warn of AI’s dangers full-time. Like Robert Oppenheimer before him, he says his creation may do more harm than good, giving himself “the normal excuse,” that someone else would have come up with this if he hadn’t.
Thus, we have a stark contrast between the 1990s boom I participated in and today’s GenAI boom. It’s not just bigger, with bigger numbers and bigger stakes. Its players will also be the world’s largest companies, and governments, many of which see the people they rule as mere objects.
It’s a dark tunnel we’re going into, with no idea what’s on the other side. This is not 1995.