Recent searches
Search options
I disagree with pretty much both the core premise and every step of the reasoning of this post
I've long been a fan of Sloan's work and that makes me feel obligated to explain why
Maybe if I scream into a pillow the urge to write a reply blog post will go away?
I don't think the urge is going to go away.
Sloan has not only drank the Kool-Aid, he's mixed it himself.
The 'black-box' mystique that is continually raised around LLMs is weird and smacks of: "We don't understand it, so it must be profound!"
It's not profound. The only profundity is how we're responding to it as a culture. LLMs as AI is snake oil. It always has been.
Anyway, good luck not responding. 
@401matthall @baldur I don't understand that conclusion at all. Like, it depends on either never having delved into understanding anything, or inability to learn from clear patterns.
Like, if you're "into tech" at all, your whole childhood was "wow that sounds really cool I wanna see how it works. Oh. Oh. It doesn't work and it's stupid af." Wizard of Oz level.
I _think_ that's the actual hang-up.
I think the hopeful fiction of the "Wizard of Oz" style AI captures the 'believer' and reason stops or skews from that point onward.
Every thought arrived at from that point, no matter how well reasoned, is built on a fallacy.
There's a psychological term for a person who willingly fools themselves into a belief and I can't recall what it is. It _feels_ like that's what's going on here.