Why we started an anonymous desk
AI is being narrated by two crowds — the salespeople and the doom-merchants — and both are wrong in the same way. This is who we are, why we hide our names, and what we promise to do instead.
Start with the thing nobody says out loud: “I can’t tell anymore what’s real about AI and what’s someone trying to sell me something.”
If you’ve felt that, you’re not slow. You’re paying attention. The signal-to-noise ratio on the single most important shift of our lifetime has collapsed — and it collapsed on purpose.
So we built a desk. This is the first thing it files.
The two crowds
Right now, almost everything you read about artificial intelligence is written by one of two crowds, and they hate each other, and they are wrong in exactly the same way.
The first crowd is selling. Founders, funds, consultants, the LinkedIn oracle with a course to move. To them every model release is a revolution, every tool is the future, and the only mistake you can make is not buying in fast enough. Their AI is always about to change everything — because their business depends on you believing it.
The second crowd is selling too — they’re just selling fear. Doom scales beautifully. “AI will take your job” is a better headline than “AI will change part of your job in ways that depend heavily on your sector,” even though the second one is true and the first one is a slot machine. Panic gets the click. The click is the product.
Here’s what the two crowds have in common: neither of them opens the study. They react to the press release, the demo, the rival’s tweet. The actual paper — the payroll data, the central-bank series, the peer-reviewed number — sits unread, because reading it is slow and slow doesn’t trend.
That gap is where we live.
What we actually do
The premise of this desk is almost embarrassingly simple. AI is not really a technology story. It’s a story about human lives — your work, your money, your kids, the quiet decisions you’re making right now about a future nobody can see clearly.
So we keep the lens narrow and the subject human. The lens is always the same: what is AI actually doing to real people? The subjects are endless: the job market, the classroom, the trades, the wage, the thing your government is about to regulate badly. We don’t write “about AI.” We write about the life AI is rearranging, and we use AI as the lens to see it.
And we have one rule that decides everything: we go to the primary source.
Not the article about the study. The study. When we tell you entry-level hiring in AI-exposed roles fell, you’ll get the payroll dataset and the economists’ names. When we bring you a number from Europe, it comes from Eurostat or Destatis or the central bank that published it — not from a blog that read a blog that misremembered it. We read the boring PDF so you don’t have to, and we tell you exactly where every figure came from. If a number reached us secondhand, we say so.
We label everything. When it's a measured fact, we'll show you the source. When it's our interpretation, we'll say "our read" — out loud, every time. When it's a guess about where things are heading, we'll flag it as a guess, never dressed up as a certainty. You should always know whether you're looking at the data or at our opinion of the data. Most writing on this topic blurs that line on purpose. We think the line is the whole job.
We cover Europe and the United States together, always. Most of what you read is written from inside one American zip code and quietly assumes the rest of the world is a footnote. The shift isn’t American. The data lags differently on each continent, but the pattern crosses the ocean — and if you only watch one side, you miss half the picture.
Why we don’t tell you our names
You’ll notice there’s no byline here. No founder’s face, no “as a former engineer at—,” no credentials to trade on. That’s deliberate, and it’s worth explaining, because in a world obsessed with personal brands it looks like a bug.
It isn’t. It’s the point.
The moment this becomes a person, it becomes about the person — their reputation, their reach, their need to be right in public. We don’t want authority that comes from who we are. We want authority that comes from what we can prove. A cited number doesn’t get more true because someone impressive said it, and it doesn’t get less true because we’re anonymous. The source is the credential. Strip away the name and all that’s left is the work — which is exactly the test we want to be judged on.
So we write as we. A collective, not a personality. And to be clear about what anonymity is not: it is not a license to invent. We hide who we are; we never fake what we know. No made-up experts, no fabricated stories, no borrowed authority. The names are hidden. Nothing else is.
What we’re not
We’re not neutral about nonsense, and we’re not here to make you feel better. If the data is grim, you’ll get it grim. But we made one promise to ourselves before we filed a single report, and it’s the one we’ll break last:
We never end on doom.
Fear is easy and it’s cheap and it’s everywhere, and it does something quietly corrosive — it makes you passive. If the future is already decided and it’s coming for you, why act? That’s the trap, and it’s the trap both crowds want you in, because a frightened reader is an obedient one.
The truth is more useful and less comfortable: most of what AI is doing to your life is still, substantially, a set of choices — yours, your employer’s, your government’s, your kid’s. The number tells you where the ground is shifting. What you do while standing on it is not written yet. Every report we file will show you the shift honestly, and then show you where the ground is still solid enough to stand and move.
That’s the deal. No hype, because hype is a sales pitch. No doom, because doom is a different sales pitch. Just the real research, read carefully, brought back to you in plain language — with the receipts, and with the one thing the two crowds keep forgetting to mention: that you still get to decide what to do about it.
Welcome to the desk.
Separating signal from noise.