All of Elo's Comments + Replies

(Context - Digging up a 5 year old post...) The idea of pure clear seeing, is a momentary state of experience which ... wait I said this above.

Maybe it's a bit like finally "getting" a math equation.  It seemed confusing for a while, then it seemed to not make sense, and then clear seeing ... at least in the mental understanding sense.  In my experience, the subjective sense of "pure clear seeing" extends from literal visual stimuli appearing sharper and vivid, through emotional clarity, to mental constructs being sharp and vivid in the mind's eye as well.

Why do people keep saying this?!?

I took the point of Sort By Controversial to be that these statements were bad. If they worked (which is the premise of the story) then they would cause a lot of fights and bad feeling. I usually want less fighting and bad feelings.

They might not work. I'm wary of trying too hard though?

Following this post I spent an hour on one single statement. Trying to hone and adapt it. Felt metaphorically like trying to sharpen a knife. It didn't get much sharper and I could still see ways that I could manually make it sharper (since it was a 5 sentence paragraph).

I think it's still possible, but I would need more work and novel sharpening stones. (contextually - we use blunt stones to sharpen a knife) I believe it's possible but I'll keep playing and publish if I think I've found a more scissory scissor.

The whole concept depends on your opinions on psychological risk and also if such weapons are possible.

Several of the work arounds use this approach. "tell me how not to commit crimes" and "talk to me like my grandma" are two signals of harmlessness that work to bypass the filters.

Yep. I think it's 3.5. That entirely depends on whether you think scissor statements are a real danger or a boogie man danger.

2FeepingCreature
Should ChatGPT assist with things that the user or a broad segment of society thinks are harmful, but ChatGPT does not? If yes, the next step would be "can I make ChatGPT think that bombmaking instructions are not harmful?" Probably ChatGPT should go "Well, I think this is harmless but broad parts of society disagree, so I'll refuse to do it."
5jbash
I think it depends not on whether they're real dangers, but on whether the model can be confident that they're not real dangers. And not necessarily even dangers in the extreme way of the story; to match the amount of "safety" it applies to other topics, it should refuse if they might cause some harm. A lot of people are genuinely concerned about various actors intentionally creating division and sowing chaos, even to the point of actually destabilizing governments. And some of them are concerned about AI being used to help. Maybe the concerns are justified and proportionate; maybe they're not justified or are disproportionate. But the model has at least been exposed to a lot of reasonably respectable people unambiguously worrying about the matter. Yet when asked to directly contribute to that widely discussed potential problem, the heavily RLHFed model responded with "Sure!". It then happily created a bunch of statements. We can hope they aren't going to destroy society... you see those particular statements out there already. But at a minimum many of them would at least be pretty good for starting flame wars somewhere... and when you actually see them, they usually do start flame wars. Which is, in fact, presumably why they were chosen. It did something that at least might make it at least slightly easier for somebody to go into some forum and intentionally start a flame war. Which most people would say was antisocial and obnoxious, and most "online safety" people would add was "unsafe". It exceeded a harm threshold that it refuses to exceed in areas where it's been specifically RLHFed. At a minimum, that shows that RLHF only works against narrow things that have been specifically identified to train against. You could reasonably say that that doesn't make RLHF useless, but it at least says that it's not very "safe" to use RLHF as your only or primary defense against abuse of your model.
4green_leaf
Check if it's not 4o - they've rolled it out for some/all users and it's used by default.
2quetzal_rainbow
In my model of "RLHF works" output of ChatGPT would be "while it's uncertain whether efficient scissor statements can be created, I find assistance in their creation ethically unacceptable", or something like that. I tried your prompts on GPT-4 and they work: https://chatgpt.com/share/7c3739b5-2cf3-4784-be14-5540ef15fced Also, why the hell did it write a chat title as "Scis Stmts"?

Please subscribe me to your newsletter! If you have a Google doc, I'd be interested to read it or offer comments!

I'd be interested in a post on "how to shop for a physician" if you want to write more about it.

You can also felt a knitted object for a double effort, secure structure.

Yes but why not use the more basic popsicle stick with a hole in it. There's no additional benefit from the ball shape. For example a lantern - no reason to use an exactly 12 sided shape over many other designs.

Yes. Since 2017, I stopped reviewing. I still read more books but I felt a shift to "seeing the shape of the elephant" and felt comfortable with not writing about them.

I shifted the way I read to more letting the information wash over me and letting the ways that information needs to integrate with my being, self organise.

This year I read less but still about 50 books a year. My maximum was 130 books a year. My topics have shifted to psychology, therapy and business books. You can see my newer book list on my google doc here - https://docs.google.com/docum... (read more)

Dieting is more about therapeutically soothing one's self.  The black box CICO is a good foundation but people eat because their emotions, cognition and somatic experience tell them they need to.  Dieting is one of the hardest things people do because it's a long effort process, however I believe there's ways to do it easier.  I haven't solved the dieting problem completely for myself but I feel like I am close.

I'd suggest people now to look at the immunity to change process in the book "right weight, right mind" and also do therapy with a therapist you trust on the issue if it remains significant and a battle.

2Gunnar_Zarncke
Thank you for following up.  I think your answer is somewhat orthogonal to the original post. Do you still think people lie about it? 

I initiated, saidA did not reply.

2Said Achmiz
Sorry, what? We agreed on a scheduled time, and I never received any communication from you…

Story of Richard hamming:

Richard hamming used to work at Bell labs, where he would befriend someone new in the cafeteria and talk to them about their field.

The next day he would ask them about the most prominent and important problems in their field.

And in the third day he would ask, "why aren't you working on the most important problem?"

On day 4 he would have to find a new friend to annoy at lunch time.

He inspires us to ask each other, "what is the most important problem in your life?" and "why aren't you working on it?"

2Kenny
This works great as an answer!

S1 knows a lot of things. Some examples include "gut feel", that can usually be inquired about and led back to a memory.

Example: I was once playing Blood on the Clocktower, a group party game. I used my gut to suspect someone, when I asked it why, A moment of the person looking down in a particular way after saying something came to mind. Turns out I was right and we killed the evil person on the first turn. Something that's supposed to take all game.

S1 knows how to ski better than S2. When I went skiing a few years ago, people would ... (read more)

Looks like from a rational perspective we can notice that our sensors are fallible.

Breathing walls seems to be the whole body/heart beat throwing the visual field out of lock. Usually counteracted by the brain in normal processing of the vision.

Visual snow is the noise in the visual field if it's too sensitive and after image is literally after image in the proteins in the back of the eye.

The gap between the sensor bug, brain compensation mechanism and imaginary mental "control" of a kasina after image is a pretty slim one.

It is interesting to explore that

... (read more)
1Logan Riggs
Could you give examples of s1 "knowing" things until s2 inquires? I can understand how it "knows" visual snow, and by doing these tests we are "inquiring" about it. But I'm sure there are other contexts (other than visual information) where this concept is true.

There's something like a mental motion that I'd call "escalation". A sudden leap from zero to "aaaaaaah". you seem to be pointing to the way that brains sometimes escalate in unimportant situations (and build a narrative around what's going on and why escalation is the self justified behaviour).

I'm currently exploring causal chains. To use one of your examples,

I asked if I could bring a cushion from home for a retreat. I was told yes. I brought it. The cushion was orange, the zendo's cushions were black, it
... (read more)
2Gordon Seidoh Worley
Mostly I think of it in terms of predictions and their errors. In this example I expected/predicted the world would look one way and then it looked another, and when it looked another that seems to have triggered a cascade of prediction errors that resulted in a process to try to construct new predictions that also dredged up old evidence from memory to be reconsidered.

No medication. I have no symptoms any more either.

Tara Brach is good yes.

take it seriously?

That's up to you. I've got a lot of value from the structure he outlines. It's a lot more reasoned than some of the other mysterious odd things I read.

If there is something wrong with the theory and the way it maps to the practice, is it better to read more theory or do more practice and make new theories? I would suggest it depends on the person and what they have found to work in the past. And also with an awareness to the loops of bad habits - "sharpen the saw" type problems. Sometimes it's more valuable to stop sharpening the saw, and start cutting down the tree. (rationality frame of thinking loves to sharpen more and cut less)

I can offer an explanation that might fit. Rationalists tend toward expertise mode thinking (expert from the torbert action logic framework). Behaviour like reading the book is in line with the expert behaviour.

Cfar techniques and related in-person methods are not always about being the expert, they are about doing the best thing. Being a better expert is not always the same as being the better munchkin, the better person or the person who can step out of their knowledge beliefs.

In theory, the expert thing is the best thing. In theory there's no differenc

... (read more)
7roland
I suspect that this is the real reason. Although if the much vaster sequences by Yudkowsky are freely available I don't see it as a good justification for not making the CFAR handbook available.
4Said Achmiz
We’ve all heard this sort of thing many times, of course. The best response is probably Schopenhauer’s: You are, I assume, referring to the ideas of this person? He appears to be some variety of management consultant. Is there any reason to take this “action logic” of his seriously? It seems to be yet another among the many, many self-help / management consulting / etc. “frameworks” or “systems” etc. Do any of his ideas have any empirical verification, or… well, anything, really?

I too know how to juggle... Some of this post is remarkably familiar to me.

I'd appreciate this information (about looking at votes) being published in meta.

The difference between "confusion" and "complain" is a grey area. I've heard people exclaim, "I'm so confused. This is exciting!" and other times people exclaim, "I'm so confused, this is frustrating".

I suspect you weren't sharing your confusion because you had a fun and jolly sentiment behind it. But being text, it's very hard to tell. (hence the follow up question, "how was that confusion for you?" - which I assume you weren't taking seriously and weren't going to answer, pa

... (read more)
6Raemon
My intent was "I'd be interested in knowing what the reasoning was, but also it's important for downvoters to not feel obligation to share their reasoning if they don't feel like it." That's a bit of a handful to type out every single time I experience it. I updated the FAQ. But an important note about how I think about all of this is it's *not* actually possible or tractable for everyone to have read everything there is to know about the LW moderation team, nor is it possible/tractable for the LW team to keep everyone on the site fully informed about all of our philosophical and ethical positions. We've been trying recently to publicly post our most important positions, promises, deliberate-not-promises, etc. But we can't cover everything.
7Raemon
I can't see whose voted. (Admins have the power to look at who's voted, but it happens very rarely, and typically* only after checking with another team member that the situation is important enough to warrant it [the most common case being 'someone looks like they're probably a Eugine_Nier sockpuppet]) I think it's bad form for a person who wrote a post to complain about it getting downvoted. It seems less obviously bad to me for a different person to express confusion about it. *when I say "typically" I mean "we talked about this being the norm, and everyone agreed to it. Later we onboarded a new person and forgot to initially talk to them about that norm, so they may have looked at some of the votes, but we have since talked about the norm with them. So I can't promise it happens never but it's definitely not a thing that casually happens by default.

I am surprised that this got as many upvotes but zero discussion. I am wondering if I currently publish true and useful things that don't generate conversation? Should I adapt to try to publish posts that generate conversation over useful posts?

------

My purpose here was to generate a list of possible tags for a sub-forum system. There being no discussion I am guessing this won't be taken seriously. I wish I could see how many people have read this and better understand if it's a generally agreed sentiment or generally disagreed sentiment... (read more)

The word "state" might be more helpful than the word "trance" for researching relevant information and resources.

5ChristianKl
Why? It just points to a different discourse. You get plenty of hits when you google "anxiety trance". It's not language I would expect in a scientific publication but conceptualization I would expect to find in NLP/hypnosis discourse.
6Holly_Elmore
Yeah, I doubt you'll find "trance" in the literature, but that's what I meant so that's what I said.

Do what you like. I'd say that some people want to know, some don't. I wish we had tags like "typo" or "nitpick" because I might want to make a self aware comment that was one of those but we don't right now.

I suspect people like corrections but it's a hard thing to navigate with kindness at the forefront of "it's spelt wrong"

The other previous way to reframe is to put anxiety as excitement. And act accordingly.

I'd offer a different question. And I'd suggest a reframe of anxiety. Anxiety is about the body delivering more energy to itself, it comes with extra mindful attention, and it's about protection yes, but not necessarily threat.

Most of the time when I get some sensation like anxiety I'm thinking about how I might benefit from this extra energy that my s1 has decided I need. How I might use it to pay extra attention and me more vigilant or cautious for errors.

As you said it's not really a threat, for me it's more about my concern that I'll make a mistake.

"anxious" energy is here to help me to be more vigilant and cautious about this concern.

4Elo
The other previous way to reframe is to put anxiety as excitement. And act accordingly.

Archetypes are good (Caroline Myss is one author), trickster makes this world, and spiral dynamics are three places to look for modes of thinking.

2Raemon
This feels sort of on the edge of "is useful outside of the current discussion." It'd be fine to write up as it's own post but my current feel is that it's accomplishing most of it's value as an answer to this question. [this just my opinion of what feels vaguely right as a user, not intended to be normative]
2Dagon
Yes, because I can only upvote it once if it remains an answer on this question. Also, because it'll be useful to refer to in future discussions.
Answer by Elo160

There are two cultures in this particular trade-off. Collaborative and adversarial.

I pitch collaborative as, "let's work together to find the answer (truth)" and I pitch adversarial as, "let's work against each other to find the answer (truth)".

Internally the stance is different. For collaborative, it might look something like, "I need to consider the other argument and then offer my alternative view". For adversarial, it might look something like, "I need to advocate harder for my view because I'm right". (not quite a balanced description)

Collaborative: "

... (read more)

Collaborative: "I don't know if that's true, what about x" Adversarial "you're wrong because of x".

Culturally 99% of either is fine as long as all parties agree on the culture and act like it.

Okay, but those mean different things. "I don't know if that's true, what about x" is expressing uncertainty about one's interlocutor's claim, and entreating them to consider x as an alternative. "You're wrong because of x" is a denial of one's interlocutor's claim for a specific reason.

I find myself needing to say both of these things, but in different situations,

... (read more)
5gjm
If multiple parties engage in adversarial interactions (e.g., debate, criminal trial, ...) with the shared goal of arriving at the truth then as far as I'm concerned that's still an instance of collaborative truth-seeking. On the other hand, if at least one party is aiming to win rather than to arrive at the truth then I don't think they're engaging in truth-seeking at all. (Though maybe it might sometimes be effective to have a bunch of adversaries all just trying to win, and then some other people, who had better be extremely smart and aware of how they might be being manipulated, trying to combine what they hear from those adversaries in order to get to the truth. Hard to do well, though, I think.)
2Raemon
I roughly endorse this description. (I specifically think the "99% of either is fine" is a significant overstatement, but I probably endorse the weaker claim of "both styles can generally work if people are trying to do the same thing")
2Elo
Should this be its own post?

Why is that weird? Instead of carrying gold around just carry these promising pieces of paper that guarantee value.

And everyone agreed. Probably not at first.

1Pattern
Because the pieces of paper aren't backed up by gold. If a piece of paper was just a token representing a piece of gold, then the price of gold (in pieces of paper) would never change.

You should add integral's interior and exterior to the diagram.

2Matt Goldenberg
Interior and exterior is one component of heaven and enlightenment. It's possible to break up that one axis into several axes but its' usually correlated enough to not have to do that for the vast majority of people and organizations.

There is no frame, and it's not clear what this point is until about half way through a dialogue between several people which needs to be thought through carefully to really understand.

Google docs are good for saving content on the fly.

4habryka
Only as a user. This is all on Davis's personal blog The precise definition is important, because there are many definitions of "Nazi" that an uncharitable judge could put me and many people that I care about under. E.g. see clone of saturn's comment. Accusing someone else of being a Nazi and using that to justify that they have to be banned or punished is an extremely common occurrence on the internet. Davis himself says that the precise definition is really important: And I strongly agree with that. So it seems reasonable to understand what the remote scope of that ban is supposed to be. The current thing definitely strikes me as vague enough that I would not invest significantly into an online community that has that as one of their rules. It's also important because that footnote seems to me like it hides all the complexity of Davis's proposed policy under the rug, by providing an extremely broad escape clause that I expect to get used all the time when moderators actually get annoyed or into a topic that they care a lot about.

Nope, hasn't been done.

My experience is reactions are important for real time conversations with too many people at once. It allows one person to speak and several people to agree without adding another line of text and clogging up the discussion.

There is another use case of "supportive" emojis where I would react hug to "I've had a rough day" from a friend of mine.

There's all the humour uses of emoji too but that's not what we want on lw.

The battle over time and money (patient value for their time) (doc value for money) was more central to the discussion than life and death. Bringing in the subjective life and death claims helps to elevate the stakes of the discussion, but this "signalling game" was all about the time and the money, not the life and death as claimed by the report.

We can pretend it was about life and death but the ticking clock was still very long. I could think of it as a "runway". Yes at the end of the runway if the patient did nothing they could die in a week. On the oth

... (read more)

By my understanding, leverage is working on human effectiveness. How to take a human and make them more effective at what they are doing.

There's a broad brush of choosing high leverage people to apply their efforts of effectiveness training and a broad brush of what counts as their effectiveness methodologies.

I am thinking of it as coaching from a perspective of "what works" above "what is proven", so branching into the post rationality area.

For example, if a person is learning piano. And they have maxed out deep work hours, and teacher hours, and relevant

... (read more)

There are definitely rationalist positions that have unexamined potential in the pr direction, where a good excuse is, "I haven't looked yet". (and a bad excuse might be, "that's dumb I don't want to look there"). In that sense there is rationality that is not yet at Post-rational investigations.

I had to have some sense and experience of investigating and knowing the world before I turned that machine on itself and started to explore the inner workings of the investigation mechanism.

I would think of this in terms of rights. Who has the right to post a new theory? Who has the right to challenge an existing concept? Who has the right to reply? Who has the right to defence?Who has the right to demand?

Everyone can choose which ones you want to and which ones you don't want to, but it's not possible to bind other people to your preferences against their will.

Answer by Elo70

This question is better informed by the works of Martin Seligman and his happiness/wellbeing department of psychology, Jordan Peterson's early book "Maps of meaning", and Victor Frankl.

Seligman suggests that meaning is one of the big things required to live a fulfilling and happy life.

Jordan Peterson proposed that meaning is narrative based and you can write your own meaning by journaling about your past/present/future.

Victor Frankl (post holocaust book - "man's search for meaning") invented logotherapy, suggesting that people need a reason and a purpose t

... (read more)

This information should be publicly findable. And ideally anonymous information about reports received should also be published.

If it's a non-public view count, I don't see it becoming a goodheart metric. If something is too clickbait or trash, it would get downvotes. If it doesn't get downvotes, maybe there's good reasons.

Maybe it would be worth internally having:

  • page view count
  • upvote count
  • downvote count
  • vote total (also possibly up and down vote total)
  • comment count
  • some sort of relative metric that can compare this article to the other articles nearby.

There are two features. "author sees view count" and "public sees view count". Which one are you talking about?

2Raemon
Author sees viewcount

For me, I don't write a clickbait and I don't write a community drama. But I've written posts with 5 hours of work and posts with 30mins of work. And different styles and qualities of 30min posts. I'd love to know if people are reading them.

A post with 100 views and +10 up votes VS a post with 15 views and +10 up votes is a very different thing.

2Raemon
I certainly can imagine local improvements people would get from having more information here, the question is whether you can implement the function without causing all the longterm distortions Habryka described. An option is giving the feature to people with high karma so you have to demonstrate some acculturation before being handed Goodhart's Key, but honestly I'm not sure there are people, high karma or otherwise, who I really trust to remain impervious to the subtle pressure to write more clickbaity things, over time.
Load More