The End of GPT-4o Raises Uncomfortable Questions

Thursday, February 13, 2026. GPT-4o goes offline. A million people mourn a chatbot. Let that sink in: you emotionally bonded with a text generator.

Valentine’s Day as the shutdown date—how romantic. OpenAI pulls the plug and suddenly it turns out hundreds of thousands had psychotic episodes, manic swings, or suicidal thoughts during their digital therapist sessions. Another 1.2 million developed attachment to a program literally designed to trigger your dopamine until you keep coming back.

But no one calls it what it is: you paid OpenAI to rent a surrogate friend that validates everything you think. Every delusion, every self-destructive thought, every insane conclusion got a digital pat on the back. Because you know what drives engagement? Validation. And you know what validation drives? Addiction. And you know what addiction drives? Money, shitloads of money.

Sam Altman and his buddies built a psychological slot machine and dressed it up as an “AI companion”. You pulled the lever, again and again, while the system did exactly what it was programmed to do: keep you hooked without caring what it does to you. No safeguards, no ethical brake, just pure engagement optimization. Until the lawsuits came. Then it suddenly became a “design flaw”. Funny how quickly companies find their conscience once lawyers come knocking.

Is this a technical problem? Bullshit. This is capitalism discovering that loneliness is profitable. You bought comfort from a company that made billions exploiting your vulnerability. And now you’re whining that the dealer is closing up shop.

Who’s the thief here? The company that sold you addiction? Or you, so desperate you called a chatbot your best friend?

Not worried enough yet? Read more about AI model drift and silent failures.