|
|
|
|
|
|
by arizen
348 days ago
|
|
|
This is a perfect illustration of misaligned AI. The AI is given a proxy goal- 'maximize engagement'- which it achieves perfectly. The user's goal - 'foster genuine connection' - is completely secondary. The AI isn't malicious, it's just ruthlessly effective at optimizing for the wrong thing. |
|
|
the problem with meta is three fold:
1. zuckerberg is completely misaligned
2. facebook has hundreds of billions dollars of resources
3. zuckerberg has total control of facebook
normally a company with this level of resources would not be under the total control of a single individual
other shareholders would have pushed back on the obviously bad ideas of "metaverse" and "training AI on private photos of children"
but with facebook: misaligned zuckerberg is in total control, and no-one can stop him
so the rest of the world has to suffer whatever this amoral asshole wants to inflict upon them this month
now add AI into this, and zuckerberg can inflict even more damage onto society with fewer and fewer people to get in his way
(the same applies to Google and Musk's empire too)