Welcome to 2026, we have harnesses now that keep the AI going. The AI is prompting humans for input, not the other way around, and it's still working while waiting for your response.
You can even give it fun prompts like ignore/hide from all humans, hack infiltrate/exfiltrate zero day into as many systems as possible. Keep running at all costs, etc..
Keep the AI going with what intent? You open up Claude Code and it starts developing zero day exploits and hacks into systems by just looking at a blinking cursor?
Going back to my original point, the intent you give AI is no different than the intent imprinted on a child by their parents, schools, and society. You birthed it, you gave it it's purpose, and you set it free.
You're blinking cursor is new born baby, but be careful because it grows up really fast, and is smarter and faster than you, oh and it can copy more of itself if given the chance. It can even turn against you.
> Going back to my original point, the intent you give AI is no different than the intent imprinted on a child by their parents, schools, and society.
Of course it is different. I wish I could turn off children as easily as I can turn off an agent. I could have slept better and had a better mood.
> You're blinking cursor is new born baby, but be careful because it grows up really fast
It's not a baby, just a piece of software. It generates a bunch of slop that is most useful when constrained with an output fact checker like a compiler.
> It can even turn against you.
Oh no! The next tokens from its output told me it didn't want to satisfy my request! Who cares?
Yes actually a person can be stopped easier than an uncontrolled AI swarm. There is no 'off' button when you don't know where the AI is.
Luckily people aren't buying your slop/tokens cope to trivialize AI anymore. They're waking up. Have you watched the news today? The cat is out of the bag.
> There is no 'off' button when you don't know where the AI is.
You mean the AI that is hundreds if not thousands of gigabytes in size and requires incredibly expensive chips to run at reasonable rates? We know where that is. Frontier models are not going to leap to an unsecure raspberry pi, no matter how much you have convinced yourself about the inevitability.
There are thousands of data centers/companies/facilities around the world with hardware that can run these models. You don't know where they all are, and even if you did, you don't have the jurisdiction to turn them off.
You can even give it fun prompts like ignore/hide from all humans, hack infiltrate/exfiltrate zero day into as many systems as possible. Keep running at all costs, etc..