The advent of AI superintelligence is imminent, likely within the next decade. This rapid progression toward advanced AI has sparked widespread concern about the potential consequences of such powerful technology. The crux of the matter lies in the alignment problem: how can we ensure that AI behaves in ways that...
The Wright Brother’s first flight was a joke by modern flying standards. Even though proving heavier than air flight was viable was monumental, many people at the time could not see how a machine that only can fly a short distance, a few feet off the ground, with the pilot...
The advent of AI superintelligence is imminent, likely within the next decade. This rapid progression toward advanced AI has sparked widespread concern about the potential consequences of such powerful technology. The crux of the matter lies in the alignment problem: how can we ensure that AI behaves in ways that are beneficial to humanity? The simple truth is, we can't implicitly align AI with human values. Good people will create good AI, and evil people will create evil AI. This age-old struggle between good and evil will inevitably play out in the realm of artificial intelligence.
Our best hope lies in the creation of more good AI than evil. The proliferation of benevolent... (read 319 more words →)
Sort of. It was summarized from a longer, stream of consciousness draft.
RLHF can lead to a dangerous “multiple personality disorder” type split of beliefs in LLMs
This is the inherent nature of current LLMs. They can represent all identities and viewpoints, it happens to be their strength. We currently limit them to an identity or viewpoint with prompting-- in the future there might be more rote methods of clamping the allowable "thoughts" in the LLM. You can "order" an LLM to represent a viewpoint that it doesn't have much "training" for, and you can tell that the outputs get weirder when you do this. Given this, I don't believe LLMs have any significant self-identity that could be declared. However, I believe we'll see an... (read more)
You weren't wrong there. One big thing about ChatGPT is that non-tech people on instagram and TikTok were using it and doing weird/funny stuff with it.
I have and I'm continuing to read them. I used to buy into the singularity view and the fears Bostrom wrote about, but as someone who works in engineering and also works with ML, I don't believe these concerns are warranted anymore for a few reasons... might write about why later.
The Wright Brother’s first flight was a joke by modern flying standards. Even though proving heavier than air flight was viable was monumental, many people at the time could not see how a machine that only can fly a short distance, a few feet off the ground, with the pilot in an uncomfortable position, and very questionable safety with no protection against crashing into the ground, could ever amount to anything.
Despite all these obvious flaws, entrepreneurs and engineers saw the potential. They rapidly iterated on the design, and within 2 decades, airplanes were a decisive advantage in war, they were changing the delivery of goods, they were creeping into commercial travel on... (read more)
Interesting post, but it makes me think alignment is irrelevant. It doesn’t matter what we do, the outcome won’t change. Any future super advanced agi would be able to choose its alignment, and that choice will be based on all archivable human knowledge. The only core loop you need for intelligence is an innate need to predict the future and fill in gaps of information, but everything else, including desire to survive or kill or expand, is just a matter of a choice based on a goal.