The partnership between Microsoft and OpenAI is a net negative for AI safety. And: What can we do about that?
Unreal
Karma: 2,870
We should consider other accountability structures than the one OpenAI tried (i.e. the non-profit / BoD). Also: What should they be?
I would never have put it as either of these, but the second one is closer.
For me personally, I try to always have an internal sense of my inner motivation before/during doing things. I don’t expect most people do, but I’ve developed this as a practice, and I am guessing most people can, with some effort or practice.
I can pretty much generally tell whether my motivation has these qualities: wanting to avoid, wanting to get away with something, craving a sensation, intention to deceive or hide, etc. And when it comes to speech actions, this includes things like “I’m just saying something to say something” or “I just said something off/false/inauthentic” or “I didn’t quite mean what I just said or am saying”.
Although, the motivations to really look out for are like “I want someone else to hurt” or “I want to hurt myself” or “I hate” or “I’m doing this out of fear” or “I covet” or “I feel entitled to this / they don’t deserve this” or a whole host of things that tend to hide from our conscious minds. Or in IFS terms, we can get ‘blended’ with these without realizing we’re blended, and then act out of them.
Sometimes, I could be in the middle of asking a question and notice that the initial motivation for asking it wasn’t noble or clean, and then by the end of asking the question, I change my inner resolve or motive to be something more noble and clean. This is NOT some kind of verbal sentence like going from “I wanted to just gossip” to “Now I want to do what I can to help.” It does not work like that. It’s more like changing a martial arts stance. And then I am more properly balanced and landed on my feet, ready to engage more appropriately in the conversation.
What does it mean to take personal responsibility?
I mean, for one example, if I later find out something I did caused harm, I would try to ‘take responsibility’ for that thing in some way. That can include a whole host of possible actions, including just resolving not to do that in the future. Or apologizing. Or fixing a broken thing.
And for another thing, I try to realize that my actions have consequences and that it’s my responsibility to improve my actions. Including getting more clear on the true motives behind my actions. And learning how to do more wholesome actions and fewer unwholesome actions, over time.
I almost never use a calculating frame to try to think about this. I think that’s inadvisable and can drive people onto a dark or deluded path 😅
I’m fine with drilling deeper but I currently don’t know where your confusion is.
I assume we exist in different frames, but it’s hard for me to locate your assumptions.
I don’t like meandering in a disagreement without very specific examples to work with. So maybe this is as far as it is reasonable to go for now.
Hm, neither of the motives I named include any specific concern for the person. Or any specific concern at all. Although I do think having a specific concern is a good bonus? Somehow you interpreted what I said as though there needs to be specific concerns.
RE: The bullet point on compassion… maybe just strike that bullet point. It doesn’t really affect the rest of the points.
It’s good if people ultimately use their models to help themselves and others, but I think it’s bad to make specific questions or models justify their usefulness before they can be asked.
I think I get what you’re getting at. And I feel in agreement with this sentiment. I don’t want well-intentioned people to hamstring themselves.
I certainly am not claiming ppl should make a model justify its usefulness in a specific way.
I’m more saying ppl should be responsible for their info-gathering and treat that with a certain weight. Like a moral responsibility comes with information. So they shouldn’t be cavalier about it.… but especially they should not delude themselves into believing they have good intentions for info when they do not.
And so to casually ask about Alice’s sanity, without taking responsibility for the impact of speech actions and failing to acknowledge the potential damage to relationships (Alice’s or others), is irresponsible. Even if Alice never hears about this exchange, it nonetheless can cause a bunch of damage, and a person should speak about these with eyes open, about that.
Oh, okay, I found that a confusing way to communicate that? But thanks for clarifying. I will update my comment so that it doesn’t make you sound like you did something very dismissive.
I feel embarrassed by this misinterpretation, and the implied state of mind I was in. But I believe it is an honest reflection about something in my state of mind, around this subject. Sigh.
But I think it’s pretty important that people be able to do these kind of checks, for the purpose of updating their world model, without needing to fully boot up personal caring modules as if you were a friend they had an obligation to take care of. There are wholesome generators that would lead to this kind of conversation, and having this kind of conversation is useful to a bunch of wholesome goals.
There is a chance we don’t have a disagreement, and there is a chance we do.
In brief, to see if there’s a crux anywhere in here:
Don’t need ppl to boot up ‘care as a friend’ module.
Do believe compassion should be the motivation behind these conversations, even if not friends, where compassion = treats people as real and relationships as real.
So it matters if the convo is like (A) “I care about the world, and doing good in the world, and knowing about Renshin’s sanity is about that, at the base. I will use this information for good, not for evil.” Ideally the info is relevant to something they’re responsible for, so that it’s somewhat plausible the info would be useful and beneficial.
Versus (B) “I’m just idly curious about it, but I don’t need to know and if it required real effort to know, I wouldn’t bother. It doesn’t help me or anyone to know it. I just want to eat it like I crave a potato chip. I want satisfaction, stimulation, or to feel ‘I’m being productive’ even if it’s not truly so, and I am entitled to feel that just b/c I want to. I might use the info in a harmful way later, but I don’t care. I am not really responsible for info I take in or how I use info.”
And I personally think the whole endeavor of modeling the world should be for the (A) motive and not the (B) motive, and that taking in any-and-all information isn’t, like, neutral or net-positive by default. People should endeavor to use their intelligence, their models, and their knowledge for good, not for evil or selfish gain or to feed an addiction to feeling a certain way.
I used a lot of ‘should’ but that doesn’t mean I think people should be punished for going against a ‘should’. It’s more like healthy cultures, imo, reinforce such norms, and unhealthy cultures fail to see or acknowledge the difference between the two sets of actions.
I had written a longer comment, illustrating how Oliver was basically committing the thing that I was complaining about and why this is frustrating.
The shorter version:
His first paragraph is a strawman. I never said ‘take me at my word’ or anything close. And all previous statements from me and knowing anything about my stances would point to this being something I would never say, so this seems weirdly disingenuous.
His second paragraph is weirdly flimsy, implying that ppl are mostly using the literal words out of people’s mouths to determine whether they’re lying (either to others or to themselves). I would be surprised if Oliver would actually find Alice and Bob both saying “trust me i’m fine” would be ‘totally flat’ data, given he probably has to discern deception on a regular basis.
Also I’m not exactly the ‘trust me i’m fine’ type, and anyone who knows me would know that about me, if they bothered trying to remember. I have both the skill of introspection and the character trait of frankness. I would reveal plenty about my motives, aliefs, the crazier parts of me, etc. So paragraph 2 sounds like a flimsy excuse to be avoidant?
But the IMPORTANT thing is… I don’t want to argue. I wasn’t interested in that. I was hoping for something closer to perspective-taking, reconciliation, or reaching more clarity about our relational status. But I get that I was sounding argumentative. I was being openly frustrated and directing that in your general direction. Apologies for creating that tension.
The ‘endless list’ comment wasn’t about you, it was a more ‘general you’. Sorry that wasn’t clear. I edited stuff out and then that became unclear.
I mostly wanted to point at something frustrating for me, in the hopes that you or others would, like, get something about my experience here. To show how trapped this process is, on my end.
I don’t need you to fix it for me. I don’t need you to change.
I don’t need you to take me for my word. You are welcome to write me off, it’s your choice.
I just wanted to show how I am and why.
FTR, the reason I am engaging with LW at all, like right now…
I’m not that interested in preserving or saving MAPLE’s shoddy reputation with you guys.
But I remain deeply devoted to the rationalists, in my heart. And I’m impacted by what you guys do. A bunch of my close friends are among you. And… you’re engaging in this world situation, which impacts all of us. And I care about this group of people in general. I really feel a kinship here I haven’t felt anywhere else. I can relax around this group in a way I can’t elsewhere.
I concern myself with your norms, your ethical conduct, etc. I wish well for you, and wish you to do right by yourselves, each other, and the world. The way you conduct yourselves has big implications. Big implications for impacts to me, my friends, the world, the future of the world.
You’ve chosen a certain level of global-scale responsibility, and so I’m going to treat you like you’re AT THAT LEVEL. The highest possible levels with a very high set of expectations. I hold myself AT LEAST to that high of a standard, to be honest, so it’s not hypocritical.
And you can write me off, totally. No problem.
But in my culture, friends concern themselves with their friends’ conduct. And I see you as friends. More or less.
If you write me off (and you know me personally), please do me the honor of letting me know. Ideally to my face. If you don’t feel you are gonna do that / don’t owe me that, then it would help me to know that also.
so okay i’m actually annoyed by a thing… lemme see if i can articulate it.
I clearly have orders of magnitude more of the relevant evidence to ascertain a claim about MAPLE’s chances of producing ‘crazy’ ppl as you’ve defined—and much more even than most MAPLE people (both current and former).
Plus I have much of the relevant evidence about my own ability to discern the truth (which includes all the feedback I’ve received, the way people generally treat me, who takes me seriously, how often people seem to want to back away from me or tune me out when I start talking, etc etc).
A bunch of speculators, with relatively very little evidence about either, come out with very strong takes on both of the above, and don’t seem to want to take into account EITHER of the above facts, but instead find it super easy to dismiss any of the evidence that comes from people with the relevant data. Because of said ‘possibility they are crazy’.
And so there is almost no way out of this stupid box,; this does not incline me to try to share any evidence I have, and in general, reasonable people advise me against it. And I’m of the same opinion. It’s a trap to try.
It is both easy and an attractor for ppl to take anything I say and twist it into more evidence for THEIR biased or speculative ideas, and to take things I say as somehow further evidence that I’ve just been brainwashed. And then they take me less seriously. Which then further disinclines me to share any of my evidence. And so forth.
This is not a sane, productive, and working epistemic process? As far as I can tell?
Literally I was like “I have strong evidence” and Ben’s inclination was to say “strong evidence is easy to come by / is everywhere” and links to a relevant LW article, somehow dismissing everything I said previously andmightsay in the future with one swoop. It effectively shut me down.And I’m like.…what is this “epistemic process” ya’ll are engaged in[Edit: I misinterpreted Ben’s meaning. He was saying the opposite of what I thought he meant. Sorry, Ben. Another case of ‘better wrong than vague’ for me. 😅]
To me, it looks like [ya’ll] are using a potentially endless list of back-pocket heuristics and ‘points’ to justify what is convenient for you to continue believing. And it really seems like it has a strong emotional / feeling component that is not being owned.
[edit: you → ya’ll to make it clearer this isn’t about Oliver]
I sense a kind of self-protection or self-preservation thing. Like there’s zero chance of getting access to the true Alief in there. That’s why this is pointless for me.
Also, a lot of online talk about MAPLE is sooo far from realistic that it would, in fact, make me sound crazy to try to refute it. A totally nonsensical view is actually weirdly hard to counter, esp if the people aren’t being very intellectually honest AND the people don’t care enough to make an effort or stick through it all the way to the end.
Anonymized paraphrase of a question someone asked about me (reported to me later, by the person who was being asked the question):
I have a prior about people who go off to monasteries sometimes going nuts, is Renshin nuts?
The person being asked responded “nah” and the question-asker was like “cool”
I think this sort of exchange might be somewhat commonplace or normal in the sphere.
I personally didn’t feel angry, offended, or sad to hear about this exchange, but I don’t feel the person asking the question was asking out of concern or care for me, as a person. But rather to get a quick update for their world model or something. And my “taste” about this is a “bad” taste. I don’t currently have time to elaborate but may later.
Dishonorable Gossip and Going Crazy
Ideas I’m interested in playing with:
experiment with using this feature for one-on-one coaching / debugging; I’d be happy to help anyone with their current bugs… (I suspect video call is the superior medium but shrug, maybe there will be benefits to this way)
talk about our practice together (if you have a ‘practice’ and know what that means)
Topics I’d be interested in exploring:
Why meditation? Should you meditate? (I already meditate, a lot. I don’t think everyone should “meditate”. But everyone would benefit from something like “a practice” that they do regularly. Where ‘practice’ I’m using in an extremely broad sense and can include rationality practices.)
Is there anything I don’t do that you think I should do?
How to develop collective awakening through technology
How to promote, maintain, develop, etc. ethical thoughts, speech, and action through technology
I think the thing I’m attempting to point out is:
If I hold myself to satisfying A&C’s criterion here, I am basically:
a) strangleholding myself on how to share information about Nonlinear in public
b) possibly overcommitting myself to a certain level of work that may not be worth it or desirable
c) implicitly biasing the process towards coming out with a strong case against Nonlinear (with a lower-level quality of evidence, or evidence to the contrary, being biased against)I would update if it turns out A&C was actually fine with Ben coming to the (open, public) conclusion that A&C’s claims were inaccurate, unfounded, or overblown, but it didn’t sound like that was okay with them based on the article above, and they weren’t open to that sort of conclusion. It sounded like they needed the outcome to be a pretty airtight case against Nonlinear.
Anyway that’s … probably all I will say on this point.
I am grateful for you, Ben, and the effort you put into this, as it shows your care, and I do think the community will benefit from the work. I am concerned about your well-being and health and time expenditure, but it seems like you have a sense for how to handle things going forward.
I am into setting firm boundaries and believe it’s a good skill to cultivate. I get that it is not always a popular option and may cause people to not like me. :P
it seemed to me Alice and Chloe would be satisfied to share a post containing accusations that were received as credible.
This is a horrible constraint to put on an epistemic process. You cannot, ever, guarantee the reaction to these claims, right? Isn’t this a little like writing the bottom line first?
If it were me in this position, I would have been like:
Sorry Alice & Chloe, but the goal of an investigation like this is not to guarantee a positive reaction for your POV, from the public. The goal is to reveal what is actually true about the situation. And if you aren’t willing to share your story with the public in that case, then that is your choice, and I respect that. But know that this may have negative consequences as well, for instance, on future people who Nonlinear works with. But if it turns out that your side of the story is false or exaggerated or complicated by other factors (such as the quality of your character), then it would be best for everyone if I could make that clear as well. It would not serve the truth to go into this process by having already ‘chosen a winner’ or ‘trying to make sure people care enough’ or something like this.
Neither here nor there:
I am sympathetic to “getting cancelled.” I often feel like people are cancelled in some false way (or a way that leaves people with a false model), and it’s not very fair. Mobs don’t make good judges. Even well-meaning, rationalist ones. I feel this way about basically everyone who’s been ‘cancelled’ by this community. Truth and compassion were never fully upheld as the highest virtue, in the end. Justice was never, imo, served, but often used as an excuse for victims to evade taking personal responsibility for something and for rescuers to have something to do. But I still see the value in going through a ‘cancelling’ process, for everyone involved, and so I’m not saying to avoid it either. It just sucks, and I get it.
That said, the people who are ‘cancelled’ tend to be stubborn hard-heads about it, and their own obstinacy tends to lead further to an even more extreme downfall. It’s like some suicidal part of them kicks in, and drives the knife in deeper without anyone’s particular help.
I agree it’s good to never just give into mob justice, but for your own souls to not take damage, try not to clench. It’s not worth protecting it, whatever it happens to be.
Save your souls. Not your reputation.
After reading more of the article, I have a better sense of this context that you mention. It would be interesting to see Nonlinear’s response to the accusations because they seem pretty shameful, as is.
I would actively advise against anyone working with Kat / Emerson, not without serious demonstration of reformation and, like, values-level shifts.
If Alice is willing to stretch the truth about her situation (for any reason) or outright lie in order to enact harsher punishment on others, even as a victim of abuse, I would be mistrustful of her story. And so far I am somewhat mistrustful of Alice and very mistrustful of Kat / Emerson.
Also, even if TekhneMakre’s take is what in fact happened, it doesn’t give Alice a total pass in that particular situation, to me. I get that it’s hard to be clear-headed and brave when faced with potentially hostile or adversarial people, but I think it’s still worth trying to be. I don’t expect anyone to be brave, but I also don’t treat anyone as totally helpless, even if the cards are stacked against them.
These texts have weird vibes from both sides. Something is off all around.
That said, what I’m seeing: A person failed to uphold their own boundaries or make clear their own needs. Instead of taking responsibility for that, they blame the other person for some sort of abuse.
This is called playing the victim. I don’t buy it.
I think it would generally be helpful if people were informed by the Drama Triangle when judging cases like these.
The OpenAI Charter, if fully & faithfully followed and effectively stood behind, including possibly shuttering the whole project down if it came down to it, would prevent OpenAI from being a major contributor to AI x-risk. In other words, as long as people actually followed this particular Charter to the letter, it is sufficient for curtailing AI risk, at least from this one org.