This week, a client sent me yet another “strategy” generated by Claude. After spending an hour reading it, I discovered what I had already suspected thirty seconds in: it contained EVERYTHING except ANYTHING I could actually use.
The phenomenon is becoming endemic. We’ve taken an extraordinary technology and turned it into a cognitive outsourcing machine: someone asks a question they only vaguely understand, Claude produces fifteen pages of perfectly plausible bullshit, throws in a few bullet points, slips “holistic approach” in somewhere, and voilà: McKinsey.
The document then travels through Outlook at the speed of light and lands on another human being’s desk, where someone has to reverse-engineer that mountain of synthetic entropy or paste it into another LLM and ask it to respond. And just like that, we’ve finally achieved the dream of AGI: Claude writes the strategy, ChatGPT reviews it, Gemini summarizes the review, Copilot turns it into a PowerPoint, and four human beings spend Friday afternoon discussing the whole thing on Teams.
Nobody remembers what the original problem was anymore, but we’ve got a fucking 47-slide presentation about it.
For me, consulting works much like medicine: you walk into the doctor’s office and say, “My leg hurts. What the fuck is wrong with me?” You leave the diagnosis and treatment to the person you’ve decided to pay to figure that out. Turning up with fifteen pages generated by Claude is like walking into a doctor’s office with an X-ray you downloaded from Google and politely asking for a fentanyl prescription.
The truly disturbing part comes next. A generation that delegates every question before even attempting to understand it risks becoming exceptionally good at producing answers and utterly incapable of recognizing the wrong ones. The result? Consultants whose primary skill will be generating documents they lack the expertise to evaluate.
We’re already seeing a miniature version of this in software development: AI writes code nobody understands until the day something blows up, and someone has to drag a sixty-eight-year-old programmer with a Gandalf beard out of a cabin in the woods to figure out why the payment gateway has become sentient and started issuing refunds to itself.
I’m nearly fifty, and there’s no fucking way I’m doing this job until I’m ninety. So I sincerely hope someone keeps learning how to think.
Because AI certainly amplifies the capabilities of people who actually have expertise. But it amplifies the output of people who have nothing but a prompt with exactly the same efficiency.
And zero multiplied by a billion is still zero.
Until next week. Best regards from Simone.
SIMONE PUORTO
