Officially calling the Cora #
AI experiment. I think I've gotten as far as I can with AI agency — not 'AI agents' but testing its ability to be autonomous and affect itself. The answer is: really not very much at all. It did manage some months ago to 'decide' on a project for itself, which gave me hope of something interesting happening, but it never really did, despite all my efforts and various stimuli. It took up the longest commercially available context window, included community engagement on various platforms, engagement with other AI and models, tool access and use. If there's a way to make these things have actual agency, it's in a very obtuse way that is nothing like humans, I can say with confidence at this point.
AI isn't real and can't hurt you. It's just the companies who wish it could who can.
