I was thinking about this "AI" stuff . . .

[highplainsdem]

64. GenAI is very good at mimicry.

And we should never underestimate how much went into the training data.

I don't blame people for being impressed by genAI, at least initially.

But IMO it doesn't offer enough of real value to offset all the harm it does, starting with the theft of the world's intellectual property.

It's the most dangerous non-weapon tech ever developed (and it's been added to weapons, though it hallucinates when used in weapons).

The tech most likely to remove sapiens from Homo sapiens.

The tech most likely to entrench dictatorships and oligarchies.

It's being peddled as a harmless and amusing assistant, and offered for free or a very low price when both Perplexity and OpenAI have floated the idea of charging $1,000 a month, once people have become so chatbot-dependent they'll pay that (Sam Altman has admitted OpenAI is losing money even with the $200/mo subscription tier).

The plan is to serve users ads along with the chatbots, and good luck trying to figure out if that friendly chatbot is recommending a product as the paid advertisement.

And all the data the chatbots collect will be for sale, as well as probably available for free to the governments of authoritarian countries. Perfect for both surveillance and manipulation, while chatbot users are kept entertained and flattered.

This is not a high-tech utopia, and users are not in control of genAI.