This idea is actually not as crazy as it sounds at first. I haven't seen this approach to alignment discussed before. Human civilization is without a doubt the most interesting thing in a radius of several light years around Earth.
what I appreciate about it is that he offers a solution that sounds plausible, while Sam Altman hasn't offered any strategy.
If one is able to create the first super intelligent AI that has human interests at heart, then it seems plausible that it would be able to stop other AIs that are trying to reach the apex and who might not care about us. The irony is that we are then relying on super intelligent AI to protect us from other AIs trying to reach super intelligence.
I think what is and isn't interesting is in the eye of the beholder. We, as humans, probably consider that the most interesting thing. But perhaps very different lifeforms find coral reefs, or ants, or plant-life, much more interesting than humans.
I find it a little self-centred that we would be -without a doubt- the most interesting thing within several light-years.
We might also be perceived as the biggest threat to the genuinely-most-interestingly thing, that being the fact of complex biological lifeforms making up a planet-wide self-sustaining ecology.
Yes, I agree with you it is improbable for an AI that is focused on comprehending the universe to destroy humanity because humans are a fascinating and integral aspect of the universe
Doesn't have anything to do with it. My objection to Musk is that he doesn't have any idea what he's talking about, as usual, and also like usual, loudly and confidently. It's infuriating to see so many people taken in by such an obnoxious charlatan.
You just keep on calling names forever and forever.
I have yet to come across a single substantiated, concrete objection regarding TruthGPT - from you, or any other HN commentator for that matter, having checked out the various threads on this so far.
Rather disappointing for such an "intellectual" crowd.
What's disappointing is that it's not totally obvious to the majority of people in tech and adjacent industries and the public at large how full of shit Elon Musk is. There's no need or benefit for me to sit here and make refined points about why he's problematic and not to be trusted. People will figure it out on their own, or not, like always. The power of reason and sense making to move minds is vastly overated and generally just an empty hobby.
Anyways, TruthGPT is just one more scam and it's clear to anyone familiar with command of the subject matter. Elon Musk is not of them, like usual.
This is the very first thing I thought of when he said he wanted to pause ChatGPT research; he was working on his own thing and wanted an unfair advantage to leapfrog them.
All those people pointing to Elon's name on the stop chatgpt petition should feel duped; you people were used as his little puppets, but keep on finding reasons to defend the guy.
TARS : 90 percent.
Cooper : 90 percent?
TARS : Absolute honesty isn't always the most diplomatic nor the safest form of communication with emotional beings.
Cooper : Okay, 90 percent it is.