Big tech companies own and control most of our civilization's physical compute.
Their only problem is they don't know what to do with it.
As AI obliterates the marginal value of code, it's not obvious how "the cloud" and its vast physical infrastructure will continue to pay for itself.
They've been subsidizing software startups for a while, often providing hundreds of thousands of dollars of credits, begging software companies to use their compute. As LLMs allow individuals to create bespoke software cheaply, what happens to the billion-dollar data centers built for no-longer-viable enterprise software services?
Many onlookers assume that the current rush to Artificial General Intelligence (AGI) will be a boon to those who own the data centers, but from a slightly different angle, we see the the rush to AGI as a certain desperation of those who own the data centers.
Beyond a certain size, publicly-traded technology companies become nihilistic, zombie-like machines. Structurally incapable of innovating beyond a certain degree of creativity (the "innovator's dilemma"), they cannot have values, desires, or visions other than those which are endogenous to the status quo. Thus, when the world changes drastically, it is not merely that they sometimes don't know what to do. In some sense, they cannot know what to do.
That's why they're investing billions of dollars in small, agile startups. Microsoft has OpenAI, Google has Magic.dev, and so forth.
The AGI startups are selling a very expensive solution to technological nihilism: "We'll build God on your computers, and then God will tell you what to do with them."
AGI should not be seen as a great gift to increasingly dominant tech monopolies, but rather a desperate gamble of big tech companies who have spent many billions on controlling physical compute without any purpose other than power.
They have no idea what to do with the computers.
This is great news for anyone who knows what they want to do with their computer. In a future post, we'll explain why.
