Money is like a drug. With proper dosage it can make you well. Take too much and you overdose.
Silicon Valley began overdosing in May 2017.
That is when Masayoshi Son and the Saudis’ Public Investment Fund announced the “Vision Fund,” a $100 billion effort to dominate the tech future. It backed a host of losing propositions, most notably WeWork. It conned itself with a belief in visionary founders, who naturally turned out to be con men.
The stupidity was at the fund’s core, the idea that markets could be built and bought if a start-up founder had enough capital. This is the way things work in the oil business. It’s not how it works in the tech business.
This caused the current AI crash.
The idiocy behind the Vision Fund and garbage like WeWork has since infected the entire Valley. When you’re funding startups at a valuation of $9 billion, something is wrong.
Startups must be about mission, not money. Resource constraints are essential to keep the team’s eyes on the ball. Managing the team through the start-up phase should also inform the VC on how to proceed. None of this has been happening.
Worse, this confusion over money, founders, and market control is now global. Many analysts now believe Deepseek’s success means Alibaba and Baidu are certain to control AI.
Nothing could be further from the truth.
A Proper Startup
The whole idea of a software startup is that a small team works intently toward a product that could, with proper capitalization, turn into something wonderful. Deepseek’s success was based on its being committed to the goal, and being resource constrained, not its backing from the Chinese government.
The model used by China in developing its Electric Vehicle market, and the model used by Silicon Valley before the Vision Fund, are models of what should be done.
The LLM race shows precisely what not to do, and how a money overdose can occur.
OpenAI was founded around a mission that at the time looked impossible. That’s why it was an open source project.
Nvidia showed that running a Large Language Model, capable of flexible output from an enormous database, was possible. Most of the money OpenAI raised from Microsoft went right back to Microsoft, in the form of Azure credits. In this way it was a little like America’s aid to Ukraine, in that what look like a big payout was mostly money flowing back to where it came from.
The launch of ChatGPT in late 2022 was based on this pattern. Resources were constrained, far more than reporters admitted. Microsoft used an order for capacity it created to build the data centers needed for the job. The money Nvidia got for its chips was real. The money Microsoft gave ChatGPT was not.
Strange Days Indeed
It’s what Altman and the VCs did after this that was irrational. ChatGPT pioneered a new concept. It wasn’t the destination, but the starting gun. Rather than looking to build on its work or improve upon it, Cloud Czars played the Great Game, plowing billions into copying ChatGPT and building infrastructure around the copies. Rather than fund 100 startups, VCs clustered around Altman’s old team, turning each one of them into a mogul.
The result is we don’t know if the Nvidia infrastructure the Cloud Czars are building will find a market yet. This is the uncertainty behind the recent fall in “Big Tech.” Small wonder that Son and Altman went into business together at the dawn of the new Trump era.
LLMs aren’t magic. They’re a method. They’re also limited. They don’t get better as they get bigger. On a cost vs. value basis, they get worse.
Big Tech needs to go back to the garage with AI. It needs to fund 100 startups, including 50 failures, each with a clearly defined goal, a minimal budget, and people with experience watching their every move.
But this won’t be done in Silicon Valley. That well has been poisoned by too many Vision Fund deals, in AI and elsewhere. This is a game anyone can play, and one that everyone will. There’s no assurance that China will control the game, nor that the Cloud Czars will. There’s every opportunity here for European, Asian, and even African start-ups. I think Marco Fioretti has it just right.
It’s just that some stocks are going to fall before Silicon Valley calls AA over its overdose of money.
