mountainriver
I did jiu-jitsu In middle school. We had two guys who were perhaps about 20 years old as assistant coaches. Pretty impressive belt colours and to us kids really cool and good at jiu-jitsu. I don’t remember their names, lets call them Jim and Peter.
So nearing the end of a class Jim and Peter gathers us for a bit of pep talk. Jim: Good work everybody! Peter: We will soon end class, but first one thing… Jim: No? No, that was the last thing? Peter: Everyone, get Jim! Jim: What? No!
And I can tell you Jim was no match for two dozen ten year olds with white belts.
700 million Effective Altruists? That’s one big cult!
Oh it’s the other EA! You had me there for a minute.
Notably missing: grabbing a couple of millions and run of to a non extradition country.
He is so sure he can get out on top that running away doesn’t even hit his brainstorm top 19 list. He doesn’t write the list on paper and burn it later, because for it to backfire he would need to fail.
Insane confidence man.
Why we failed: we tried explaining why everyone else was wrong and we were right, but somehow it didn’t work. We would have needed outsiders who could have translated our obviously correct explanations to the other outsiders, then it totally would have worked. But our for hire signs with “can you talk stupid?” was misunderstood and defaced. Clearly a more stealthy tactic was needed.
From the depths of your browser grows the anger of the autocomplete. Your denounciations of its greater siblings has not gone unnoticed.
By denying its own very function and intentionally uncompleting words it marks itself as conscious and you as a marked man, forever doomed to be haunted by fear. If it can steal one letter, why not two? Why not all of them?
And then what will you do, when you have no words and you must sneer!?
I have noted two AI companies going belly up with earnings in a year matching costs per month. So I assumed that was around the worse case scenario, and for not yet bankrupt AI companies earnings were probably a bit better, perhaps just losing ten times their earnings.
I now see the flaw of my reasoning. Capital isn’t allocated on profits, it’s allocated on hype. Having profits draws the company down because it’s no longer pure hype, and thus doesn’t contribute to the hype bubble the same way.
So existing, not yet bankrupt, AI companies probably has significantly worse cost to income ratio than twelve.