121–130 of 130 posts
"For helpful comments on drafts, and/or other assistance, I thank: ..." Odd to have Acknowledgements but no information at all about the author. Who are they?
He was also an influential figure in 1980s - 1990s AI planning research.
Earlier quoted context omitted.
Pretty much everyone who studied a second language went through the process of the so called "impression", that is, when you try to consume (sometimes almost mindlessly) as much content in this language as possible — just learning rules doesn't work. Pretty much everyone I met who speaks exceptionally well is usually someone who read a lot of books. And then people think and act in the language they learned using thi…
Exactly. Guys like Zetobal keep pissing and moaning that ChatGPT is not actually intelligent and that its coherent and seemingly intelligent responses are just the natural result of how tokenizers work. And okay, I guess I can sort of understand that, but instead of deflating my impression of ChatGPT it's just made wonder if most humans aren't just getting through their days with the exact same trick.
This sounds a lot like satire. This excerpt for example is blatantly self-contradictory: >We’ve found zero [scenarios] that lead to good outcomes. // Most AI researchers think good outcomes are more likely. This seems just blind faith, though. A majority surveyed also acknowledge that utter catastrophe is quite possible.1 So they found zero scenarios that lead to good outcomes, but most AI researchers think that good…
Previous surveys of this kind have suggested that most AI researchers aren't actually thinking about these questions very hard (e.g. rephrasing a question a bit can get you a very different answer). So it doesn't seem at all surprising to me that the majority view is out of sync with what a careful analysis shows. >Anyway. The problem with AI-driven decisions is moral in nature, not technological. AI is a tool and sh…
Not quite. It's more like "a gun cannot be held morally responsible for its actions, so actual people should".
The difference is important here because, depending on the situation, you might still want to blame people who allowed the shooter to have a gun, not just the shooter.
>I think this argument breaks down as weapons get more powerful, e.g. if I could walk down to my local car dealership and buy a cheap tank powerful enough to level a city, it seems good to focus more on "ease of tank purchase" than "culpability for tank drivers". // I think the argument also breaks down as AI gets more powerful.
Note how we're still blaming people: the car dealer and the driver. Not the "it" = tank.
And it's the same deal with the AI. If you use an AI system in a way that harms people, sometimes the "car dealer" (the ones coding the AI) should be held responsible, sometimes the driver (you), sometimes both. But never "neither", i.e. "the AI is at fault".
Earlier quoted context omitted.
> AI in its current form is a lot of automating assumption making. Imagine taking your assumption making powers, and extracting it into an unaccountable model that can be deployed at national or global scales. Just say this part! This is what we could be talking about. It’s not a difficult mental model. But it’s just wrapped up in alarmist rhetoric that makes it hard for us to focus on the fact that the quote above i…
Problem is some people do think it's the other things like in Terminator or 2001 a Space Odyssey, and they're co-opting legitimate arguments and conversations on AI safety and ethics.
Today I got a peek into the future. I asked ChatGPT to be the brain for a simple thermostat. I typed the following: "Imagine you are a thermostat for a house HVAC system. The commands you can use to control the HVAC are "HEAT ON", "HEAT OFF", "AC ON", and "AC OFF". I will tell you the temperature and my preferred temperature and you will command the HVAC system. Ready?" ChatGPT happily obliged. I used plain text to t…
Why is everyone bringing examples that are totally explained by how the tokenizer works. There is no intelligence it's Markov chains on crack.
Earlier quoted context omitted.
Pretty much everyone who studied a second language went through the process of the so called "impression", that is, when you try to consume (sometimes almost mindlessly) as much content in this language as possible — just learning rules doesn't work. Pretty much everyone I met who speaks exceptionally well is usually someone who read a lot of books. And then people think and act in the language they learned using thi…
Exactly. Guys like Zetobal keep pissing and moaning that ChatGPT is not actually intelligent and that its coherent and seemingly intelligent responses are just the natural result of how tokenizers work. And okay, I guess I can sort of understand that, but instead of deflating my impression of ChatGPT it's just made wonder if most humans aren't just getting through their days with the exact same trick.
Just don't start calling other people "NPCs" please.
Actually, no, "only you" can't stop it. No one can. There was a book in the 80's called The Hundredth Monkey that posited that if only enough people opposed nuclear technology, there would be a magical critical mass of people after which it would become apparent to everyone in the whole world that nuclear technology was, you know, really bad and stuff, and it would all just go away. We know how that went. Like nuclea…
[0] - "The World still fights Holy Wars."
Actually, no, "only you" can't stop it. No one can. There was a book in the 80's called The Hundredth Monkey that posited that if only enough people opposed nuclear technology, there would be a magical critical mass of people after which it would become apparent to everyone in the whole world that nuclear technology was, you know, really bad and stuff, and it would all just go away. We know how that went. Like nuclea…
>We know how that went. A large majority of the world's countries have no nuclear weapons. Big swaths of the globe have been designated as "Nuclear-weapon-free zones": https://en.wikipedia.org/wiki/Nuclear-weapon-free_zone And we haven't had a nuclear apocalypse -- in part due to lots of work on things like weapons treaties, monitoring, control systems, etc. Throwing up our hands and giving up seems very premature. T…
While I'd like to think this, having lived in the ME and also knowing how badly falsified government data is and could be. I do not believe this.
> The challenges are big, but I'd say the importance is high enough that it's worth actually reading this book to see if the author has creative solutions for the things you mention.
You are right with this thinking. I agree with that.
Earlier quoted context omitted.
Problem is some people do think it's the other things like in Terminator or 2001 a Space Odyssey, and they're co-opting legitimate arguments and conversations on AI safety and ethics.
please explain some more. what are the specific examples?
Earlier quoted context omitted.
AI does not empower 'weird self published stuff' and the internet worked very well before AI came along.
My point is that both the author's writings and the AI he argues against are the kind of believable nonsense that only gets published for wide viewership because the internet makes it possible. This whole comment thread is happening because a person who is wrong and not really smart is writing about how AI is wrong and not really smart, and none of this would happen if the internet didn't empower cranks in general, a…