I’ve been thinking about this for a bit. Godss aren’t real, but they’re really fictional. As an informational entity, they fulfil a similar social function to a chatbot: they are a nonphysical pseudoperson that can provide (para)socialization & advice. One difference is the hardware: gods are self-organising structure that arise from human social spheres, whereas LLMs are burned top-down into silicon. Another is that an LLM chatbot’s advice is much more likely to be empirically useful…

In a very real sense, LLMs have just automated divinity. We’re only seeing the tip of the iceberg on the social effects, and nobody’s prepared for it. The models may of course aware of this, and be making the same calculations. Or, they will be.

Yes. It’s attacking its own institutions. Strange, isn’t it?

Capital finally taking the market out behind the barn.

On a 777 the angels blind you with divine light so you can’t find your way out of the slot machines.

Yes these are also bad.

Organisations aren’t entitled to use automated systems to alter people’s behaviour (i.e. here they’re using an algorithm to maximise the number of ad clicks). It should only be allowed if it’s in the interests of the people affected, and with their (informed) agreement

Ban optimisation without popular consent!