So, this week, a Beijing startup called Moonshot released Kimi K3, 2.8 trillion parameters, the largest open-weight model (in other words, free) ever built, benchmarks within touching distance of the very top American systems.

Kimi's weights go public at the end of the month.

Meaning: in about a week, anyone on the planet can download a near-frontier AI, modify it, and run it.

Good stuff.

Or is it?

For some, not necessarily. The markets did a DeepSeek dance. Chip shares are down. Analysts on television are explaining "geopolitical shifts in AI." 

Panic across… because the Western AI industry is essentially a messianic project in which the models, the infrastructure, and the chips are assigned two jobs at once.

Job one: save humanity, give it purpose, deliver the techno-optimist paradise (I wrote about this new corporate faith last week).

Job two: maintain the global economic status quo of the Western world. Where the core assets, the means of production, the financial power, and other goodies stay geographically located exactly where they are located now.

This is important because the narrative is the collateral that secures the $1.8 trillion valuations at 100 times revenue that I wrote about in Fomonomics. 

VOODOO plus FOMO equals money.

But if VOODOO is open source… It's not much of a VOODOO, is it?

So any news that challenges the core narrative is, essentially, blasphemy in techno-optimist corporate religion.

Which brings me to my favourite quote of the week. OpenAI's head of strategy responded to Kimi by warning that a probable outcome of an open-weight-dominant world is "full AI communism."

Communism. The man said communism.

Are Llama and Mistral also secretly communist? They have been open-weight for years… sleeper agents, presumably. OpenAI itself released open-weight models last summer, after its CEO admitted the company had been "on the wrong side of history."

So the doctrine, as I understand it, is: open weights are communism, except when we do it, which is freedom.

Now, we could indulge in commenting on corporate hypocrisy, but I'm afraid there isn't enough time in the universe, afterall even the sun apparently has only a few trillion years left. 

What's much more relevant to what we do here at Meme Guide to Management is to draw lessons.

What is happening with Kimi used to be a healthy technological competition. Company shipping a better, cheaper product. This used to be the foundation of the happy globalization these same people claimed to believe in.

Markets. Competition. May the best product win. All the good stuff…

But we no longer live in a normal economic model. We are adepts of a semi-religious one. So competition is now wrong, and it gets rebranded in near-religious terms. 

Give it a quarter, and someone will call open weights the antichrist. I am only half joking.

The ability to see past narrative nonsense is a core managerial skill.

We need to train critical thinking to perfection. Because managers make decisions that affect other people. Real ones, with mortgages and school fees, who bear the consequences of whichever story their leadership swallowed that quarter in boardroom lalaland.

And as I wrote in the ice-cream letter: at the very top, the world is a bunch of self-interested children who want money and power, building conditions to get their version of ice cream.

And this is not just a problem of AI giants.

It is a new culture quietly chipping away at the true competitiveness of major organisations across the world. I see this behavior… narratives stacked on narratives, challenged by narratives everywhere. In organisations big and small these days. 

We were supposed to end up as managers in a world of digital transparency.

We are slowly drowning in a sea of BS. A manager who knows one from another based on smell alone is rare… be that manager.

Nothing is new under the sun. This too shall pass. Twenty-five years ago, during the previous bubble, Microsoft's Steve Ballmer famously called Linux "a cancer."

A cancer.

Free, open-source software, built by volunteers, “A cancer”. Apocalyptic stuff. The stakes were civilisational.

Same nonsense as today.

And then what happened?

Everything settled.

Linux quietly won the server and most of the internet. Microsoft did not die. Today, it runs enormous Linux workloads on Azure and declares its love for Linux in official communications.

The company that cried apocalypse became more valuable than ever, by selling things around the free thing it called Cancer.

That is almost certainly the boring, profitable future here, too.

So the lesson, for us managers, is short.

Be level-headed, and know your history.

When next week's BS narrative arrives at your desk demanding an urgent strategic response, ask two questions.

And a practical tip on the models: watch what the engineers download, not what the strategists preach, and the analysts put in PowerPoint.

The engineers already know how this ends.

Talk soon,
Sultanbek