More AI Silliness – Magpies Talk to Humans

 More AI Silliness – Magpies Talk to Humans

Large language Models often construct laughable sentences. I enjoy mocking them, even if they have no mind to actually mock. Below is another good example of AI silliness. At least I hope it was done by and AI LLM; it would be extremely silly if a human wrote this.

Do Magpies Talk to Humans?

Magpies are one of the most interesting birds in the world. Not only are they known for their beautiful plumage, but they are also known for their intelligence. It is believed that magpies can communicate with humans.

In some cases, the magpie will try to mimic the person’s voice. This is thought to be a way of communicating with humans. Magpies are also known to use a variety of sounds and gestures to communicate with people.

It is believed that magpies are attracted to humans because they are curious creatures. They want to know what we are doing and why we do things. Magpies may also be trying to learn more about us by talking to us.

https://animalbehaviorcorner.com/what-do-animals-think-of-humans/

Ok, if this was written by a human, I would assume that it was meant metaphorically or perhaps in some mystical/religious sense. But, much as I like magpies, I have yet to have heard one talking to me. Yes, they do have various meaning-conveying vocalizations, but to “talk to a human” has a specific meaning, that goes a lot further than what magpies do.

Just as magpies don’t actually talk to us, neither do AI large language models. They don’t even hallucinate or make mistakes. They just connect strings of text together based on statistical correlations. From somewhere in the AI’s training corpus, a transformer algorithm connected a number of correlated tokens to churn out a sentence that claims magpies talk to people. It did not “write” these sentences, because that requires a mind, not a mindless algorithm.

I assume that some actual human filled a blog with AI prompted material and didn’t bother editing the AI silliness (yes, I know that by my own logic an AI can’t be silly, it is the human that is being silly). This sort of thing is getting more and more common. It is making good reliable information harder to find on the internet. A version of the saying “bad money drives out good money” seems to apply here.

Sometimes it seems like technological advances in communications just make things worse, by making things “free”. For example, you can know nothing about a subject and do zero actual research about that subject, then blog about it, in the hopes of getting clicks and a bit of ad money, with very little (apparent) work. Eventually this degrades the internet research experience, so that people of average intelligence begin to scoff and ignore blogs and much other content as well.

It is similar to what happened to email – it is free, so scammers have rendered it fairly useless. Same with long distance telephone calls – so cheap, that scammers can use them in fraud schemes across continents. It seems to be a rule, that as something becomes free (or asymptotically approaching free) it becomes used for frauds and scams to the point that it becomes useless and irrelevant.

It is amazing, the amount of effort and intelligence that goes into the AI development business. I predict that if it ever works, the most common monetizable purpose will be AI-scammers replacing human scammers. Eventually the advanced world economy will grind to a halt, since all other economic activity will be crowded out by hyper-intelligent scam-bots attempting to scam other hyper-intelligent scam-bots. This may be the solution to the Fermi Paradox.