Continuing the topic of how the IT landscape is finally becoming overgrown with concrete fences of geopolitics.

The other day, deputy Anton Gorelkin expressed outrage over a fresh report from the Microsoft AI Economy Institute. The gist is that in the report on the level of global AI penetration, Russia ended up on par with Gambia and Kenya with a figure of 9.5%. The deputy accused MS of espionage, claimed that Windows' share in Russia had collapsed to 10-15%, and that we are generally in the global top 3.

Do I want to say that it's normal for a politician to attack made-up theses? And that this is all a good excuse for ongoing blockades that primarily hit their own citizens? No, I don't. I want to look into it objectively.

Let's try. 👇

1️⃣ Is Gorelkin right?
He's fuming that MS is not measuring the economic impact (GDP, investments), but simply counting users.
And yes, that's true.
But surprise: the report is called Global AI Diffusion. It measures exclusively the B2C segment, not the macroeconomic effect (and they clearly state this in the methodology) — the share of people aged 15-64 who have touched generative AI at least once in a quarter. A schoolboy generating an essay in ChatGPT is a plus in the stats. Evaluating the power of a country's AI economy based on this parameter is nonsense, but Microsoft never claimed that. MS explicitly writes in the PDF: "no metric is perfect, we are measuring consumer adoption".

2️⃣ The strange thing about 10% Windows
Quote:
Microsoft is mistaken in its assessment of AI adoption. By leaving Russia, the company voluntarily reduced its presence to the level of African countries (where Windows share traditionally never rose above 10-15%).
Sorry, but that's complete bullshit. The vendor's departure means a stop to B2B licenses and official sales. No one remotely uninstalled Windows from Russian PCs. On desktops in Russia, Windows still has a total monopoly. Telemetry flows to MS servers in terabytes, unless it's blocked by corporate firewalls. And MS isn't stupid: the report states that raw data is normalized with an adjustment for the OS share in a particular market. Why simplify and distort so much?

3️⃣ What the report really doesn't account for? And what Gorelkin himself didn't mention.
MS telemetry is truly blind to Russian ecosystems. It doesn't see how users interact with GigaChat in Sber's app, chat with Alice on smart speakers, or generate images in the mobile Shedevrum. All this massive base of domestic B2C consumption flies under their radar. And those who use Western LLMs are behind circumvention tools, statistically "smearing" Russian traffic across servers in the Netherlands, Germany, and other countries.

4️⃣ Are we in the top 3?
We indeed have a strong engineering school and sovereign models. That's great.
But top 3 in the world? To compete for top 3, you need not only software but also hardware — access to computing clusters on an industrial scale and venture investments of a completely different order. Do we have that?

Yes, Gorelkin is conceptually right — the research methodology does not reflect the GREATNESS OF RUSSIA. And when I see phrases like:
In any objective ranking that takes into account real indicators, our country will be in the top three leaders.

I have only one question: are these objective studies in the same room with us right now?
Not a single global index (Stanford AI Index, Tortoise Global AI Index, Oxford Insights) places Russia in the top three (there it's the US, China, and a fight between the UK, UAE, and Singapore). Enemy propaganda?

Why not immediately refute false theses with facts? I would really like to see these objective rankings, read the studies, compare, evaluate, and rejoice in our successes. Can you share links?