DonyChristie

Money can be thrown at my Patreon here: https://www.patreon.com/reflectivealtruist

Wiki Contributions

Comments

Discussion with Eliezer Yudkowsky on AGI interventions

"...What do you do with this impossible challenge?

First, we assume that you don't actually say "That's impossible!" and give up a la Luke Skywalker.  You haven't run away.

Why not?  Maybe you've learned to override the reflex of running away.  Or maybe they're going to shoot your daughter if you fail.  We suppose that you want to win, not try—that something is at stake that matters to you, even if it's just your own pride.  (Pride is an underrated sin.)

Will you call upon the virtue of tsuyoku naritai?  But even if you become stronger day by day, growing instead of fading, you may not be strong enough to do the impossible.  You could go into the AI Box experiment once, and then do it again, and try to do better the second time.  Will that get you to the point of winning?  Not for a long time, maybe; and sometimes a single failure isn't acceptable.

(Though even to say this much—to visualize yourself doing better on a second try—is to begin to bind yourself to the problem, to do more than just stand in awe of it.  How, specifically, could you do better on one AI-Box Experiment than the previous?—and not by luck, but by skill?)

Will you call upon the virtue isshokenmei?  But a desperate effort may not be enough to win.  Especially if that desperation is only putting more effort into the avenues you already know, the modes of trying you can already imagine.  A problem looks impossible when your brain's query returns no lines of solution leading to it.  What good is a desperate effort along any of those lines?

Make an extraordinary effort?  Leave your comfort zone—try non-default ways of doing things—even, try to think creatively?  But you can imagine the one coming back and saying, "I tried to leave my comfort zone, and I think I succeeded at that!  I brainstormed for five minutes—and came up with all sorts of wacky creative ideas!  But I don't think any of them are good enough.  The other guy can just keep saying 'No', no matter what I do."

And now we finally reply:  "Shut up and do the impossible!"

As we recall from Trying to Try, setting out to make an effort is distinct from setting out to win.  That's the problem with saying, "Make an extraordinary effort."  You can succeed at the goal of "making an extraordinary effort" without succeeding at the goal of getting out of the Box.

"But!" says the one.  "But, SUCCEED is not a primitive action!  Not all challenges are fair—sometimes you just can't win!  How am I supposed to choose to be out of the Box?  The other guy can just keep on saying 'No'!"

True.  Now shut up and do the impossible.

Your goal is not to do better, to try desperately, or even to try extraordinarily.  Your goal is to get out of the box."

 

Could you have stopped Chernobyl?

Noting that the more real, second-order disaster resulting from Chernobyl may have been less usage of nuclear power (assuming that had an influence on antinuclear sentiment). Likewise, I'm guessing the Challenger disaster had a negative influence on the U.S. space program. Covid lockdowns also have this quality of not tracking the cost-benefit of their continuation. Human reactions to disasters can be worse than the disasters themselves, especially if the costs of those reactions are hidden. I don't know how this translates to AI safety but it merits thought. 

TAPs 3: Reductionism

“He has half the deed done who has made a beginning.”
– Horace

Simulacrum 3 As Stag-Hunt Strategy

This happens intergenerationally as parents forget to alert their children to the actual reasons for things. Having observed this happen with millenials, I am scared of what we are all collectively missing because older generations literally just forgot to tell us.

What do you think we are missing?

Matt Levine on "Fraud is no fun without friends."

A crucial consideration for why destroying restaurant business is good: factory farming.

CollAction history and lessons learned

Hey Ron, I am working on my own version of this (inspired by this Sequence), and would love to get your advice! Right now I am focusing on crowdfunding via dominant assurance contracts on Ethereum.

How did you / would you verify that someone did something? What are specific examples of that happening for different actions? What kinds of evidence can be provided? I have a fuzzy sense of what this looks like right now. The closest sites I can think of just off the top of my head that involve verification are CommitLock (which I made a successful $1000 commitment contract on to get myself to do swim lessons) and DietBet, which requires a photo of your scale (it also has that 'split the pot' feature you mentioned, which I am pretty excited for).

Dony's Shortform Feed

I am very interested in practicing steelmanning/Ideological Turing Test with people of any skill level. I have only done it once conversationally and it felt great. I'm sure we can find things to disagree about. You can book a call here.

Dony's Shortform Feed
I’ve mentioned previously that I’ve been digging into a pocket of human knowledge in pursuit of explanations for the success of the traditional Chinese businessman. The hope I have is that some of these explanations are directly applicable to my practice.
Here’s my current bet: I think one can get better at trial and error, and that the body of work around instrumental rationality hold some clues as to how you can get better.
I’ve argued that the successful Chinese businessmen are probably the ones who are better at trial and error than the lousier ones; I posited that perhaps they needed less cycles to learn the right lessons to make their businesses work.
I think the body of research around instrumental rationality tell us how they do so. I’m thankful that Jonathan Baron has written a fairly good overview of the field, with his fourth edition of Thinking and Deciding. And I think both Ray Dalio’s and Nicholas Nassem Taleb’s writings have explored the implications of some of these ideas. If I were to summarise the rough thrust of these books:
Don’t do trial and error where error is catastrophic.
Don’t repeat the same trials over and over again (aka don’t repeat the same mistakes over and over again).
Increase the number of trials you can do in your life. Decrease the length and cost of each trial.
In fields with optionality (i.e. your downside is capped but your upside is large) the more trials you take, and the more cheap each trial costs, the more likely you’ll eventually win. Or, as Taleb says: “randomness is good when you have optionality.”
Write down your lessons and approaches from your previous successful trials, so you may generalise them to more situations (Principles, chapter 5)
Systematically identify the factor that gives positive evidence, and vary that to maximise the expected size of the impact (Thinking and Deciding, chapter 7)
Actively look for disconfirming evidence when you’ve found an approach that seems to work. (Thinking and Deciding, chapter 7, Principles, chapter 3).

https://commoncog.com/blog/chinese-businessmen-superstition-doesnt-count/


Don’t do trial and error where error is catastrophic.

Wearing a mask in a pandemic. Not putting ALL of your money on a roulette wheel. Not balancing on a tightrope without a net between two skyscrapers unless you have extensive training. Not posting about controversial things without much upside. Not posting photos of meat you cooked to Instagram if you want to have good acclaim in 200 years when eating meat is outlawed. Not building AI because it's cool. Falling in love with people who don't reciprocate.

The unknown unknown risk that hasn't been considered yet. Not having enough slack dedicated to detecting this.

Don’t repeat the same trials over and over again (aka don’t repeat the same mistakes over and over again).

If you've gone on OkCupid for the past 7 years and still haven't got a date from it, maybe try a different strategy. If messaging potential tenants on a 3rd-party site doesn't work, try texting them. If asking questions on Yahoo Answers doesn't get good answers, try a different site.

Increase the number of trials you can do in your life. Decrease the length and cost of each trial.

Talk to 10x the number of people; message using templates and/or simple one-liners. Invest with Other People's Money if asymmetric upside. Write something for 5 minutes using Most Dangerous Writing App then post to 5 subreddits. Posting ideas on Twitter instead of Facebook, rationality content on LessWrong Shortform instead of longform. Yoda Timers. If running for the purpose of a runner's high mood boost, try running 5 times that day as fast as possible. Optimizing standard processes for speed.

In fields with optionality (i.e. your downside is capped but your upside is large) the more trials you take, and the more cheap each trial costs, the more likely you’ll eventually win. Or, as Taleb says: “randomness is good when you have optionality.”

Posting content to 10x the people 10x faster generally has huge upside (YMMV). Programming open-source something useful and sharing it.

Write down your lessons and approaches from your previous successful trials, so you may generalise them to more situations (Principles, chapter 5)

Roam is good for this, perhaps SuperMemo. Posting things to social media and coming up with examples of the rules is also a good way of learning content. cough

Systematically identify the factor that gives positive evidence, and vary that to maximise the expected size of the impact (Thinking and Deciding, chapter 7

Did messaging or posting to X different places work? Try 2X, 5X, etc. 1 to N after successfully going 0 to 1.

Actively look for disconfirming evidence when you’ve found an approach that seems to work. (Thinking and Deciding, chapter 7, Principles, chapter 3).

Stating assumptions strongly and clearly so they are disconfirmable, then setting a Yoda Timer to seek counter-examples of the generalization.

Load More