Hmm, you would probably be great playing the jailed AI in an AI boxing experiment (can you beat [someone like] EY?), but how successful would you be playing the guard?
The AI box game still seems stacked against the AI roleplayer for any similar skill level. As the AI, I don’t think I could beat someone like EY or myself on the other end, and as the gate keeper I think I would beat someone like EY or myself.
I still wouldn’t consider myself secure against even human takeover in general, especially if I’m not prepared for mental assault.
Also, can you write an AI bot that would have a decent success rate against humans, by finding and exploiting the standard wetware bugs?
Not for any interesting opponent. I can’t even write a ‘real’ chatbot. The only reason I get the results I do is because I immediately force them into a binary yes/no response and then make sure they keep saying yes .
Something here feels off: I’d call the parent a pretty strong claim, effectively “I can cure akrasia (sorta) in the majority of people who ask”. I would have expected someone to have tested this, and reported their results; if positive, I would have expected this to be the sort of thing I would have noticed much sooner than a year and two months later. (In fact, around the time this was posted, I had started reading LessWrong and I’d received an email entitled “Jim’s hypnotherapy” that I ignored for some months).
Basically, my first reaction to this was “Why ain’t ya rich?”
Having said that, I want to build up the courage to PM you for a test, if you’re still doing so; if you’re half as powerful as you claim, then of course I want to benefit from that. ;p
(I’ve been reading your blog and wound up finding this because I typed “hypnotism” into the LW search box.)
effectively “I can cure akrasia (sorta) in the majority of people who ask”.
I wouldn’t quite say that. I meant “yes, akrasia is fixable in this way”. Less “I’m a wizard!” and more “Yes, there’s a solution, it looks like that, so have fun solving the puzzle”
To make a personal claim of competence, I’d have to add some qualifiers. Maybe something like “I expect to be able to cure akrasia (sorta) in the majority of people that commit to solving it with me”, which is a much stricter criteria than “asks”. I’d also have to add the caveat that “curing” might, after reflectively equilibriating, end up with them realizing they don’t want to work as hard as they thought—that’s not my call and it wouldn’t surprise me if a significant number of people went that way to some degree.
I would have expected someone to have tested this, and reported their results; if positive, I would have expected this to be the sort of thing I would have noticed much sooner than a year and two months later.
I’m not sure if you mean within LW in particular. I haven’t yet worked magic on any LWer in this context, but I did offer a couple times.
If you’re counting outside LW, hypnotherapists get results all the time—even “amazing” results. Some people are convinced, some people write it off in one way or another. It doesn’t surprise me all that much given how people get with “skepticism” and not wanting to be made fool about hypnotism.
Basically, my first reaction to this was “Why ain’t ya rich?”
Good question.
The first part of the answer is that I have gotten a ton of value out of these skills, and only expect to gain more.
The second part is that it’s not magic. It’s more of a martial art than a cheat code. Even when it appears to be magic, there’s usually more going on in the background that made it possible. The toughest part is all the meta-level bullshit that people carry around about their problems which makes getting them into “lets solve this” mode the hard part. Once you get someone to congruently say “Yes, I’m going to be a good hypnotic subject and we’re going to fix this”, you’ve done 90% of the work—but everyone focuses on the last 10% which looks like magic and then wonders “why not sprinkle this magic pixie dust on everyone!?!”.
Also, getting “rich”—assuming you mean at a level more than charging a couple hundred dollars per hour like many hypnotherapists do—requires you to be good at finding high leverage applications and working your way into them. That’s a whole new subset of skills and I haven’t yet gotten to that stage—though I plan on working on it.
Having said that, I want to build up the courage to PM you for a test, if you’re still doing so; if you’re half as powerful as you claim, then of course I want to benefit from that. ;p
First of all, I don’t like this “if you’re half as powerful as you claim” thing—especially since you seem to have read it as stronger than intended. When I make “strong claims” I do not expect, in the social obligation sense, to be believed. I’m just trying to be understood—that really is how I honestly see things. Take with as much salt as you please.
It’s important to make this explicit because setting up high expectations for a hoping skeptic is a sure way to fail—It sets up a dynamic of me being responsible for their behavior. While I do take responsibility for their actions internally, the only real way I can do this is through making sure they take responsibility for their own actions.
Also, there should be no courage needed. I can’t just take over your mind. With you, I’m not sure whether I’d pull out hypnosis at all, and (almost) certainly can’t just get into it from the start. Also, I can only push you as hard as you let yourself be pushed. Let’s chat some time and see where it goes.
Your link to your blog is down, but once its back up and if I find this claim plausible upon reading it, I would be very interested in trying this on myself.
Yep :)
Hmm, you would probably be great playing the jailed AI in an AI boxing experiment (can you beat [someone like] EY?), but how successful would you be playing the guard?
The AI box game still seems stacked against the AI roleplayer for any similar skill level. As the AI, I don’t think I could beat someone like EY or myself on the other end, and as the gate keeper I think I would beat someone like EY or myself.
I still wouldn’t consider myself secure against even human takeover in general, especially if I’m not prepared for mental assault.
Would you know what to look for?
Also, can you write an AI bot that would have a decent success rate against humans, by finding and exploiting the standard wetware bugs?
For the most part
Not for any interesting opponent. I can’t even write a ‘real’ chatbot. The only reason I get the results I do is because I immediately force them into a binary yes/no response and then make sure they keep saying yes .
Something here feels off: I’d call the parent a pretty strong claim, effectively “I can cure akrasia (sorta) in the majority of people who ask”. I would have expected someone to have tested this, and reported their results; if positive, I would have expected this to be the sort of thing I would have noticed much sooner than a year and two months later. (In fact, around the time this was posted, I had started reading LessWrong and I’d received an email entitled “Jim’s hypnotherapy” that I ignored for some months).
Basically, my first reaction to this was “Why ain’t ya rich?”
Having said that, I want to build up the courage to PM you for a test, if you’re still doing so; if you’re half as powerful as you claim, then of course I want to benefit from that. ;p
(I’ve been reading your blog and wound up finding this because I typed “hypnotism” into the LW search box.)
I wouldn’t quite say that. I meant “yes, akrasia is fixable in this way”. Less “I’m a wizard!” and more “Yes, there’s a solution, it looks like that, so have fun solving the puzzle”
To make a personal claim of competence, I’d have to add some qualifiers. Maybe something like “I expect to be able to cure akrasia (sorta) in the majority of people that commit to solving it with me”, which is a much stricter criteria than “asks”. I’d also have to add the caveat that “curing” might, after reflectively equilibriating, end up with them realizing they don’t want to work as hard as they thought—that’s not my call and it wouldn’t surprise me if a significant number of people went that way to some degree.
I’m not sure if you mean within LW in particular. I haven’t yet worked magic on any LWer in this context, but I did offer a couple times.
If you’re counting outside LW, hypnotherapists get results all the time—even “amazing” results. Some people are convinced, some people write it off in one way or another. It doesn’t surprise me all that much given how people get with “skepticism” and not wanting to be made fool about hypnotism.
Good question.
The first part of the answer is that I have gotten a ton of value out of these skills, and only expect to gain more.
The second part is that it’s not magic. It’s more of a martial art than a cheat code. Even when it appears to be magic, there’s usually more going on in the background that made it possible. The toughest part is all the meta-level bullshit that people carry around about their problems which makes getting them into “lets solve this” mode the hard part. Once you get someone to congruently say “Yes, I’m going to be a good hypnotic subject and we’re going to fix this”, you’ve done 90% of the work—but everyone focuses on the last 10% which looks like magic and then wonders “why not sprinkle this magic pixie dust on everyone!?!”.
Also, getting “rich”—assuming you mean at a level more than charging a couple hundred dollars per hour like many hypnotherapists do—requires you to be good at finding high leverage applications and working your way into them. That’s a whole new subset of skills and I haven’t yet gotten to that stage—though I plan on working on it.
First of all, I don’t like this “if you’re half as powerful as you claim” thing—especially since you seem to have read it as stronger than intended. When I make “strong claims” I do not expect, in the social obligation sense, to be believed. I’m just trying to be understood—that really is how I honestly see things. Take with as much salt as you please.
It’s important to make this explicit because setting up high expectations for a hoping skeptic is a sure way to fail—It sets up a dynamic of me being responsible for their behavior. While I do take responsibility for their actions internally, the only real way I can do this is through making sure they take responsibility for their own actions.
Also, there should be no courage needed. I can’t just take over your mind. With you, I’m not sure whether I’d pull out hypnosis at all, and (almost) certainly can’t just get into it from the start. Also, I can only push you as hard as you let yourself be pushed. Let’s chat some time and see where it goes.
Your link to your blog is down, but once its back up and if I find this claim plausible upon reading it, I would be very interested in trying this on myself.
EDIT: read the blog, and it looks awesome.