1,573 karma · joined August 27, 2022
On the other hand, Google was openly collecting data with a view towards building AI.
And I honestly have no idea what Tao wanted to happen. How could you ensure that the AI "empowers" scientists and doesn't replace them. Intelligence is intelligence. If it's powerful enough, it will replace other intelligence.
(Btw, your comment filled me with warmth, thanks. I also recommend Scott Aaronson's "On Self-Delusion and Bounded Rationality" [1].)
[1] https://www.scottaaronson.com/writings/selfdelusion.html
I mean the thing you imagine I don't mean.
https://en.wikipedia.org/wiki/The_Inventor:_Out_for_Blood_in...
A lot of people I've met believed that it was a separate operating system. For example, I remember asking a friend "Did your computer had DOS in the beginning?" and he replied "Even worse, Norton."
My mom found me once aimlessly using my cousin's computer and she was shocked to discover I had opened Norton Commander (unintentionally of course) and told me "you know if you delete that, you'll delete all programs in the system" (passing over the absurdity of the claim itself, even then I didn't understand how opening it will make it more probable to delete it).
Not being Anglophone, even the name "Norton" sounded a bit dangerous, like something radioactive I guess.
Anyway afterwards when we got our own computer I pieced together what Norton actually was and the aura disappeared.
https://en.wikipedia.org/wiki/Dave_Mustaine
who left Metallica to found Megadeth.
They were saying one cannot replace programmers because any such attempt has led to something as complicated as code, and thus needed programmers.
I was telling them that the same way managers give them vague instructions and they turn out with code, managers could give LLMs vague instructions and they output code.
I am also sad, but my sadness is, in a way, priced in.
One thing which would have worked better (which is briefly mentioned) is Cardano's casus irreducibilis, but that would have needed more work.
For example, look at matrices [1]: no one asks themselves whether they are "real" or not. We know that we have defined operations on arrays of numbers which are useful to some degree.
If complex numbers were just called "complexes", a lot of confusion would just wither away.
Also, interestingly, quaternions and octonions aren't called "numbers".
[1] Complex numbers may, of course, be seen as special kinds of matrices, but this is beside my point here.
Er, no. That's the thing about general intelligence, it's general.
Yudkowsky wrote about AI since the 1990s, his first realization of the AI danger was circa 2001, his "Sequences" were 2006-2009.
Bostrom also wrote much earlier about AI existential risk, see e.g. this paper of him from 2002:
(To my credit, I have warned them since more than a year ago that we will reach this point.)