BRAIN SURGERY: TRY ANOTHER CHATBOT

The real news

The story, straight

A chatbot normally sends words to a screen. Connect it to laboratory equipment, and its instructions can move a needle.

That changes what a bad answer can do.

Tom's Hardware reports that Exin Therapeutics says an AI agent selected a gene therapy and controlled its injection into a mouse's brain. Gene therapy delivers genetic instructions to cells; this experiment aimed to reduce seizures.

A licensed human surgeon opened the skull. The AI chose the treatment and target, then controlled a motorized positioning rig.

A vet approved every irreversible step and held a physical emergency stop. This was not a chatbot operating without human supervision.

According to Exin's CEO, Opus models refused live intervention, while Sonnet proceeded. These are the company's descriptions of model behavior, not an explanation from Anthropic.

My imaginary hospital now has a slot-machine handle beside the model picker. Pull until a chatbot agrees.

Apparently, the most unsettling item in the operating room is the dropdown menu.

Exin says the mouse recovered well. That does not establish that its seizures improved.

Treatment results remain unpublished, and the work is not peer-reviewed. Completing an injection is not proving a cure.

The useful lesson: judge medical automation by evidence, safeguards and patient outcomes, not whether the machine managed to move.

Full reporting: Tom's Hardware. LILMEME00171.

Hardships are coming. I, the Big Dreamer, created this token to help the people.

I truly do see this token going up in value. Ride the rocket with me and let's see where the next 24 months take us.

Buy: https://lilmeme.fun/buy

Repost before somebody gives the operating room a shuffle button.

Thanks for supporting Lil'meme. Your support is appreciated.