A pair of articles went viral on Twitter yesterday, and the reason they are resonating is more interesting (to me at least) than the articles themselves.
The first article, “The Window Has Closed” by Andrew Curran, says that any nation which missed the Fable/Mythos milestone will inevitably fall further and further behind. Think of it this way—the US has declared export restrictions on that model, just as Leopold Aschenbrenner predicted several years ago.
The second article, by Satya Nadella, argues that “learning loops” and ecosystems trump selling tokens. Many are reading this as panic inside Microsoft. Elon Musk famously claimed, as far back as 2023, that OpenAI would “eat” Microsoft.
Now you have the context. Here are my inferences and takeaways.
The Wave of IPOs
SpaceX, OpenAI, and Anthropic are all going public this year. These represent the next big wave of AI tech (remember that SpaceX owns xAI). When a company goes public, it is “normalized” in a way that makes it an ordinary tech company.
That changes the narrative and vibes in one fell swoop.
Consider this; Microsoft is a software company. OpenAI and Anthropic are functionally software firehoses. These new AI companies could, hypothetically, do to Microsoft what Amazon did to Borders. Microsoft provides software through the old fashioned method of cloud delivery and enterprise developers. OpenAI and Anthropic deliver software through agentic frameworks.
If anyone is betting against Microsoft, I would not necessarily blame them.
On the other hand…
Selling tokens is not the best business model. Consider that token prices collapse 10x every year. That means you need to sell 10x more tokens every year just to break even. And as Satya (and others) point out, selling tokens is not the point. People don’t care about tokens per se. They care about the outcomes—software that works, medical diagnosis, new cancer research, etc.
On the topic of Anthropic and Fable/Mythos, consider the possibility that export controls are the new normal. If that’s the case, then America is pulling up the ladder on AI development.
That leaves Europe and China in the dust. Sorry, there’s just no two ways about it. Europe does not have the AI infrastructure, talent, or investment to match either the US or China. China has the investment, and some of the talent, but is walled off through embargos and alliances.
Geostrategic Rivalry Continues
One thing that provokes a spike of anxiety in me every now and then is the worry that the lack of profitability of AI tokens means that progress will stall out.
But then I remember that the US and China have explicitly and deliberately entered into a race condition over AI. Both nations have declared AI dominance to be their chief objective.
There can be only one.
Either the US wins, or China wins.
However, in my view, the US won before the race even began. Why? Because we have the lion’s share of talent, compute, and investment. Plus we embrace creative destruction.
When the US declared that Mythos was an export risk, that changed the game for the long haul. It will become much harder for China to steal IP or clone our work from here on out.
So we have the firehose of code coming, which may or may not disrupt incumbents like Microsoft, but we also have guaranteed spending on AI at the national level due to international competition.
