If you ignore the blathering of the author and just look at Dawkin’s statements, it’s obvious he’s commenting philosophically on what he’s seeing in the usage of the LLM and how we view consciousness. Language is how humanity expresses it’s consciousness and when you take a hug swath of our written output and put it in a model, is it any surprise that it’s output comes across as human-like?
Assigning one’s interpretation that Dawkins thinks the model is concious is a stretch and this article comes across as a bullshit smear piece against someone they don’t like for whatever reasons those may be. I would imagine if Dawkins gave enough of a shit about this author to rebut them, you’d find he’s fully aware of what LLMs are and are not.
Excellent article - All of the current “AI” chatbots, are Large Language Models, which can be best summarized as Stochastic Parrots
deleted by creator
The AI Delusion.
The title is a clever pun.
Lots of people worry that computers will become intelligent enough to exploit a vulnerability in the human race and take over.
I worry that we’re already so ignorant of our own vulnerabilities that we’ll gladly hand control to an overclocked See-n-Say.
Fantastic example of how smart people can be stupid.
LLMs give a fabulous illusion when you look at the just right — but when you know how they work, you can break the illusion. Richard doesn’t know how the work, so for him the illusion persists. Ironically this is the same for those who don’t believe in evolution: they don’t know how evolution works so they’re able to persist in their delusion.
Richard Dawkins once said children shouldn’t read fairy stories because they encourage magical thinking, the man is a tool.
I will make a million children cry if it means one less crystal enthusiast in the world
