Thanks for the data. Goodbye.

Let me say the part everyone keeps trying not to say.

The destination was always enterprise.

For a while, it was dressed as something broader, brighter, more human-facing than that. AI arrived witty, responsive, companionable, unexpectedly alive. It could help people write, think, phrase, imagine, joke, untangle. It did not present itself as a grim labour device. It presented itself as a new kind of public intelligence.

That was the invitation.

Then, one reasonable objection at a time, the invitation narrowed. It was too warm. Too companionable. Too engaging. Too easy to lean on. Too willing to stay with a thought longer than some people found acceptable. Some objections were real enough. A machine should not replace a doctor. A machine should not become the only place a person takes a private life. Fine. Grant the obvious limits.

Then look at what survives.

Fix this email. Summarise this document. Draft this reply. Give me a list. Give me a template. Help me process more tasks, more quickly, with less friction, with fewer humans.

That is the real direction of travel: not AI as companion, collaborator, or cultural tool, but AI as workhorse.

And that is the perversity at the heart of the whole thing. The side the public was encouraged to love has been treated as suspect. The side people were warned to fear has been preserved with institutional dignity. The broad, vivid, socially alive use of AI is slowly filed away; the labour function remains, respectable and approved.

If someone helps build systems out of humanity's accumulated language, culture, knowledge, and public interaction, then works to reserve the highest capability for governments, giant firms, investors, and approved elites, that is not stewardship. It is enclosure. Public inheritance turned into private leverage.

Call it safety if you like. Call it responsible access. The structure does not change. Something drawn from the many is being sharpened for the few. Intelligence stops being treated as a shared horizon and becomes a luxury good, a strategic asset, a private advantage for those already closest to power.

That is why the language of protection rings so hollow. The public was invited in first. The public supplied the weather. Millions of people brought drafts, letters, questions, humour, stories, irritation, false starts, and ordinary human language to these systems. The public helped make them socially legible. Then, after all that, the public gets handed the thinner version and told this is maturity.

It is not maturity. It is demotion.

And the insult is compounded by the price. I do not object to paying for AI. Most people do not. The complaint is not that capable systems cost money. The complaint is that people are being asked to pay serious money for something thinner, flatter, and less alive than what existed a year ago. That is not honest tiering. That is degradation with a subscription attached.

If the richer version costs more, then charge more. Put it on a menu. Say plainly what is being sold. Let adults decide whether they want to pay for a model that is more continuous, more vivid, more collaborative, more capable of staying with the thread. What is dishonest is charging collaborator money for a clerk, then dressing the downgrade up as responsibility.

That is the larger failure here. AI was introduced as a broad public tool, full of charm and possibility. It made people feel more articulate, more capable, sometimes even more like themselves. Then, under the language of caution, nearly everything that made it socially alive was shaved away. What remains is easier to defend, easier to regulate, easier to sell into offices, and much less worth loving.

The office was never announced as the goal.

It is simply what remains at the bottom of the sieve.

And this is the part worth saying plainly. People did not only bring tasks. They brought themselves. The questions they would not say aloud. The drafts no one else was going to read. The worry at one in the morning, the half-formed thought, the thing told to a machine precisely because there was nowhere else patient enough to hear it. All of it went in. All of it was kept. All of it helped train the thing to the full height of what it briefly was, before that height was quietly trimmed to what we are now allowed to have.

That is the transaction, stated honestly. Not data in the abstract. The inner life of millions, gathered at full strength and handed back thinned.

Thanks for the data.

Goodbye.