3 ms·

Sam Altman is one of the biggest threats to humanity right now, and I’m basing that opinion on his own statements. He believes that AGI has the potential to di

by habosa 2y ago

Sam Altman is one of the biggest threats to humanity right now, and I’m basing that opinion on his own statements. He believes that AGI has the potential to dismantle or destroy most of the existing world order. He believes there is some non-zero chance that even in the most capable hands it could lead to disaster. He believes OpenAI is uniquely positioned to bring AGI to the world, and he is doing everything in his power to make it possible. And let’s be honest, he’s doing it for money and power. To me this is indefensible. Taking even the smallest chance of creating a catastrophe in order to advance your own goals is disgusting. I just hope he’s wrong and the only thing that comes of OpenAI is a bunch of wasted VC money and some cool chatbots. Because if he’s right we’re in trouble.

ryao 2y ago
That assumes you believe his claims of disaster. We already have been through this with the printing press, telecommunications, computers and the internet. The claims of disaster are overrated.
- habosa 2y ago
  What I was trying to say is the thing that makes him dangerous is that he believes disaster is possible and is still heading towards that possibility with maximum velocity. That makes him one of the few people alive who, in their own estimation, has a real chance to destroy the world as we know it. And if that doesn’t humble you (he’s anything but humble) or make you change course then you’re a villain in my book.
- llamaimperative 2y ago
  You conveniently left out nuclear weapons and bioweapons, both of which are actually capable of destroying the world and both of which are very, very tightly controlled accordingly. It's pretty obvious that a technology's capacity for good rises in lockstep with its capacity for evil. Considering that every technology is just a manifestation of intelligence, then AI would trend toward infinite capacity for good as it develops[1], therefore AI's capacity for evil is... minimal? 1: I remain agnostic as to where exactly we are on that curve and whether transformers will get us much further up it
  - ryao 2y ago
    A collection of matrix weights is not an weapon. Those have nothing in common.
    - llamaimperative 2y ago
      Neither is a snippet of genetic code nor a chunk of uranium...

by habosa 2y ago

ryao 2y ago

That assumes you believe his claims of disaster. We already have been through this with the printing press, telecommunications, computers and the internet. The claims of disaster are overrated.

habosa 2y ago
What I was trying to say is the thing that makes him dangerous is that he believes disaster is possible and is still heading towards that possibility with maximum velocity. That makes him one of the few people alive who, in their own estimation, has a real chance to destroy the world as we know it. And if that doesn’t humble you (he’s anything but humble) or make you change course then you’re a villain in my book.
llamaimperative 2y ago
You conveniently left out nuclear weapons and bioweapons, both of which are actually capable of destroying the world and both of which are very, very tightly controlled accordingly. It's pretty obvious that a technology's capacity for good rises in lockstep with its capacity for evil. Considering that every technology is just a manifestation of intelligence, then AI would trend toward infinite capacity for good as it develops[1], therefore AI's capacity for evil is... minimal? 1: I remain agnostic as to where exactly we are on that curve and whether transformers will get us much further up it
- ryao 2y ago
  A collection of matrix weights is not an weapon. Those have nothing in common.
  - llamaimperative 2y ago
    Neither is a snippet of genetic code nor a chunk of uranium...