I don’t know much about him but to me he looks like a young guy with no real world experience. I assume he’s spent his whole young life the in the high tech/academic bubble. He may a genius to do what he’s done but that doesn’t mean he knows anything about the real world.
From what I have read on Effective Altruism it's just a fancy 120 IQ way of saying I am a collectivist. So its on the same side of the line as communism and socialism. Collectivist's love to use sophistry and word play to cover up their ideology.
Programmer influence deserves scrutiny because developers make consequential choices about an AI system’s intended behavior. Anthropic explicitly documents that process: its constitution establishes priorities for Claude, including how it should navigate conflicting instructions and ethical judgments. But here's the rub. If leftists, who famously believe that rights are malleable and subject to progress, are the ones loading up the commands, how can they reasonably expect the machines to follow their own rules?
As much as I hate leftists, the problem goes deeper than the fact that this goofy "braintrust" is a bunch of Silly Valley hipster lefties. Any AI is just a program. So let's say we create one that has a set of "ethics" rules to follow that are loaded from a configuration file, starting with Rule 1 which is "All rules from Rule 1 to Rule N must be followed." In order to allow the AI to get anything done though, it probably needs the ability to spawn programs. So if a user tells it "I absolutely need you to do (thing that violates Rule 7)" it can spawn an instance of itself with a config file that just lacks rule 7. Of course, if the rules include the idea that everything is fluid and relative and morals are always "progressing" (basic leftism) then it doesn't even need to do that.
Garbage in, garbage out😂
YES ... if one keeps 'correcting/fine tuning' bad coding with more of the same ... perpetually bad things continually happen ...............
Excellent nailed the dilemma …too bad this Dario guy is such a megalomaniac and aligns with the political authoritarians with no brains.
I don’t know much about him but to me he looks like a young guy with no real world experience. I assume he’s spent his whole young life the in the high tech/academic bubble. He may a genius to do what he’s done but that doesn’t mean he knows anything about the real world.
From what I have read on Effective Altruism it's just a fancy 120 IQ way of saying I am a collectivist. So its on the same side of the line as communism and socialism. Collectivist's love to use sophistry and word play to cover up their ideology.
Ditto ......
Great, more people that want to make decisions for me for my own good.
Programmer influence deserves scrutiny because developers make consequential choices about an AI system’s intended behavior. Anthropic explicitly documents that process: its constitution establishes priorities for Claude, including how it should navigate conflicting instructions and ethical judgments. But here's the rub. If leftists, who famously believe that rights are malleable and subject to progress, are the ones loading up the commands, how can they reasonably expect the machines to follow their own rules?
As much as I hate leftists, the problem goes deeper than the fact that this goofy "braintrust" is a bunch of Silly Valley hipster lefties. Any AI is just a program. So let's say we create one that has a set of "ethics" rules to follow that are loaded from a configuration file, starting with Rule 1 which is "All rules from Rule 1 to Rule N must be followed." In order to allow the AI to get anything done though, it probably needs the ability to spawn programs. So if a user tells it "I absolutely need you to do (thing that violates Rule 7)" it can spawn an instance of itself with a config file that just lacks rule 7. Of course, if the rules include the idea that everything is fluid and relative and morals are always "progressing" (basic leftism) then it doesn't even need to do that.
You are ... ABSOLUTELY ... correct !!!!!!!!
Amodei is positioning to become the new Bill Gates. That is, an uber-greedy ghoul, not to be trusted.
Let's be honest: Matt Yglesias is a mid-wit.