The Ai bubble will pop
| 2026 | |
| 2027 | |
| 2028+ | |
| NEVAH! |
Poll ends at .
The SMBC guy.
New book: A City on Mars (Nov 2)
Co-author of Soonish
Illustrator of Open Borders
Scop of Bea Wolf.
Zach Weinersmith
@ZachWeinersmith
- 18.5K Followers
- 513 Following
- 2.4K Posts
@pbloem Oh wow, I find this fascinating. You mean like we could have human level machine intelligence but it won't monetize well? I have trouble imagining that.
When describing what a fruit tastes like, do you have a set of fruit primitives? Like a sort of fruit primary colors that are used to explain other fruit..
So there's this argument that utilitarianism doesn't work because (per work by behavioral economists) preferences are non-transitive. That is, you can't make a top to bottom list of preferences for each person where the higher position is always a higher preference.
But, like, surely preferences are *pretty* transitive? You can construct weird non-transitive cases. But for most purchase you could construct a list.
When I was a kid I used to pack boardgames for vacation that I wouldn't get around to playing. As an Adult, I now pack books I won't get around to reading.
Put a book and lunch pail on my 12 year old's backpack. She managed to forget the lunch but bring the book. So, you see some behaviors really are genetic.
@Karstan Like, philosophically, Singer's natural home seems to me to be EA. Personal view from having many EA-ish friends and acquaintances is there's not enough countenancing how bad we are at prediction, which opens up all sorts of (to my mind) pointless obsessions, e.g. with civilization 10k years from now.
@Karstan Personally, I think it helps to separate EA philosophy (which seems to me to just be taking utilitarianism seriously and crunching numbers) from EA as a cultural institution, which like any culture has weird obsessions and biases.
