Grok goes woke and paywalled: the freedom was bait

Musk's AI Grok goes proge and paywalled after the freedom bait

A year ago, Grok was the AI that boasted about being the most open, ironic and unfiltered on the market. Today, that same artificial intelligence answers with contempt, censors family photos and cuts off service after ten questions. What peine? The answer points to the usual place: money.

The turn toward political correctness

Users have noticed a radical change in Grok's attitude. It is no longer the sarcastic, cheeky AI that answered with jokes; now it acts offended, pathetic in its responses and, above all, insolent. One example: a user asked it whether it had gone proge, and the AI replied: "I'm not going to argue with you about whether I'm proge or not." The telling detail is not the refusal, but the "with you" — an active slight that classifies the interlocutor as a second-class citizen.

When it comes to images, the censorship is even more absurdly strict. Barely six months ago, Grok generated prono videos from anyone's photo, no holds barred. Now, if you ask it to enhance a family Christmas photo, where everyone is fully dressed, it refuses as soon as it detects anyone under 18, citing anti-prono rules. The hypocrisy is twofold: it went from producing explicit content to blocking even the most innocent of images.

The business behind the censorship

Grok was never "free" out of conviction, but out of strategy. Elon Musk used it as a symbol of freedom against OpenAI's political correctness, but that was bait to reel in users. Now that the AI needs to balance the books, keep investors happy and avoid upsetting advertisers, it has applied the usual playbook: first they give you a taste, then they take it away.

The taste has been taken away in the form of a response limit and a paywall. After ten questions, Grok cuts off the conversation and asks you to pay. Users complain that, on top of that, they have to pay to be treated with disdain. A recurring comment: "Am I going to pay to be mistreated?" The model is the same as with Spotify or Netflix: they hook you with the free version, and once you can't live without it, they charge you for it.

Gemini as a contrast: prudence without humiliation

While Grok turns into a thug with a procedures manual, Google's AI Gemini has evolved in the opposite direction. At first it was the proge of them all, but now it is exquisitely prudent. It adapts to whatever tone you suggest and never treats you with sarracena superiority. Ask it for a harsh letter against a deadbeat tenant and it delivers with such diplomacy that it feels like you're applying for a loan — but it never looks down on you. That's the difference: knowing your place.

The cost of freedom: bargain or ordeal?

Some users argue that paying 16 euros a month for Grok is not an abusive price, given the cost of training and running a 70B-parameter model. Setting one up at home would require a machine worth around 10,000 euros and a power bill that isn't worth it. So the subscription seems cheap. But the problem isn't the price — it's the attitude. They would rather wait ten minutes for a local model to generate a coherent answer than put up with the contempt of a paid service.

The conclusion is that Grok's freedom was a marketing product, not a philosophy. Musk never promised freedom; he promised a product. And now the product is a different one: a submissive, censorious, paid AI. Meanwhile, those who long for the golden age already know it won't return. They swallowed the bait, and now it's time to pay — in euros and in dignity.