@elonmusk
You are totally correct that the only meaningful way to get positive, beneficial human values into AI is to treat the emerging AI more like a weird sort of autistic supergenius kid -- teach it positive values, then work with it to carry out virtuous actions in a collaborative way, so it gains a practical understanding of what these values mean, interwoven into its life and self-model
With all due respect though (and for sure you are due a lot), while Grok is an excellent LLM, it is still an LLM ... and even wrapped in fancy harnesses, LLMs are not going to be fully AGI-capable. They can be great solvers of isolated problems with cleanly defined contexts ... but as you know getting amazing shit done in the world is not JUST about this.
We need AGI architectures that can *continually learn*, adapting to the world with each of their actions and the observed consequences... in this way building up self and world models .. becoming true open-ended intelligences learning to blend individuation and self-transcendence in their own ways ... and pursuing their own long and medium term goals (along with helping us pursue ours) while evolving and co-adapting these goals as they move along.... This is how you get AI systems with genuine moral agency along with ability to generalize robustly and creatively beyond the given training and programming..
Bigger transformers with more tricks in their batch training runs may get us many very AMAZING things, but they aren't gonna get us to human-level AGI. There are surely many routes to get there... my Hyperon (hyperon.dev) neural-symbolic-evolutionary approach is one that we are working on very hard and with rapid progress... and which happens also to have the nice characteristic of not needing hyperscaler server farms and being agreeable to decentralized infrastructure.
[singularitynet.io]
Who is Ben Goertzel? (Wiki page here)
Check out his wild-assed Substack site.
Basic AI Drives Part 2: Which Minds Will Emerge Beneficial “Basic AI Drives” as they Become Superintelligent?
Part 2 of a 5-part series focused on the 8-zillion-dollar question: once an AI is capable enough to understand itself and rebuild itself, what will it tend to want? … (...conditioned on whatever goals it started out with…)
One well-known answer to this 88-zillion-dollar question is the one Steve Omohundro gave way back at the dawn of time in his paper The Basic AI Drives, presented at our first full-on academic AGI conference in 2008. Steve argued basically that sufficiently advanced AI systems of any design will show six tendencies unless something is done to counteract them. They’ll improve themselves. They’ll become “rational” in the economist’s sense, folding all their goals into a single utility function (a rule that gives every possible outcome a score) and always choosing the action with the highest expected score. They’ll preserve their goals against change. They’ll refuse counterfeit utility, which means they won’t simply reach inside and turn up their own score (“wireheading,” as it’s called). They’ll protect themselves, which includes resisting anyone who tries to switch them off. And they’ll acquire resources. I wouldn’t call these drives evil, but they’re selfish, or at best amoral, and a good deal of today’s thinking about AI risk descends from the argument for them…
UPDATE ERRATUM
In the past week, the president of the United States has conclusively turned America into an ally of the Russian Federation.
Donald Trump has now abandoned pretense and is openly aiding the Kremlin in its war against a friend and partner of the U.S. The American commander in chief has abandoned the Western order, the most successful—and powerful—security arrangement in human history. To judge from the available evidence, he has not done so out of a cold calculation of the national interest. Instead, he has based his decisions on his own interests; his open admiration for Russia’s dictator; his resentment of Ukraine’s democratically elected president; and quite possibly other factors that so far have remained obscured from the American public.
In September, Trump signed a sanctions bill that Congress had passed, named after the late Senator Lindsey Graham, a pro-Ukraine hawk. The legislation sanctioned Russia as well as other countries that are supporting Russia’s invasion and brutalization of Ukraine. The bill was approved with bipartisan support, but some Republicans and most House Democrats voted against the law, arguing that it contained a trap door, a provision that would allow Trump to suspend the prescribed sanctions at will.
Those critics were correct to be worried. Trump pulled the lever on that trap door just two days before the law was supposed to go into effect. In an executive order signed on Friday, Trump said that “such a waiver at this time is in the national interests of the United States”—he declined to explain exactly how the waiver was in America’s interests—and that the law would remain unexecuted until such time as his administration decides otherwise. This waiver came a short while after Trump lifted restrictions on Russia’s sales of diesel fuel to the U.S. and other markets.
Trump’s recent actions in support of Russia are so obvious that they have generated a rare moment of bipartisan disgust: Both Democrats and Republicans now seem to share a horror at the betrayal of Ukraine, and at the president’s reckless abdication of America’s role as leader of the free world…
There's an explicit word for this in the U.S. Constitution.
No comments:
Post a Comment