Author Archives: Benquo

Guess culture screens for trying to cooperate

My friend Miri (quoted with permission) wrote this on Facebook a while back:

Midwesterners are intolerably passive aggressive. My family is sitting among some grass in the dunes because it's the only shady place and a park ranger drives by and says, "That grass you're sitting in--we try to protect that." I say the only thing that makes sense to say in response, which is, "Thanks for letting me know! We'll be careful with it." And I go back to my reading.

Then I look up and she's still there. I look at her for a few moments and she says, "You need to get out of there." I'm like, ok. Why can't you just say that the first time? Not everyone grew up in your damn convoluted culture. Say what you fucking mean.

In the comments, someone replied:

One of the best parts of NYC is that no one dances around what they mean to say here. On the contrary, once I heard a guy on the subway say, to confused-looking strangers, "Do you need some fucking help or what?”

This particular incident seems like obnoxious behavior on the part of the park ranger, but it got me curious about why this sort of norm seems to win out over more explicit communication in many places. Continue reading

Honesty, magic, and science

A chocolatier friend posted this to Facebook (quoted with permission):

Just turned down an invite to sell chocolate at an event because they were going to advertise it using *free Tarot readings*

Three reasons:

-Do we as a society need more of this nonsense?

-Do I want to deal with customers that naive?

-Do I trust organizers that are either credulous or unethically pandering?

Nope, nope and nope.

I think that this is an excellent example of sticking up for principles in ways that it seems a lot of the people around me find nonobvious: refusing to sanction something you think is deceptive. This is a good practice and needs to be more widespread.

I've previously criticized the practice of crediting "matching donations" drives with gains from controlling others’ behavior, but not the corresponding loss of information they would otherwise have contributed (or the loss from accepting their symmetrical control over you). Similarly, there’s a temptation to count the gains from exploiting an event full of Tarot-credulous customers to sell your actually-high-quality chocolate, but not to count the loss of allowing such an event to exploit you. When you help someone else attract attention to something dishonest, you are imposing costs on others.

That said, I think things like Tarot (and "Magic" in general) are hard to talk about reasonably because people mean such different things when talking about them. Obviously which Tarot cards one draws are determined by a pseudorandom process, and not one meaningfully causally entangled with the future life outcomes of the person for whom the Tarot cards are being read.

However, like many other divination processes, Tarot can serve as a seed around which the reader can signal-boost their own insights about the person being read for. Often we have subtle intuitions about each other that don't make it into consciousness but are fairly insightful. I've done a Tarot reading (once), and while I don't need the cards to weave a story about someone with my intuitions, it's easy for me to imagine someone only having access to that kind of intuition if they're in a headspace where they imagine that the cards are "telling" them the story.

I also wonder whether it's possible to consistently apply this epistemic standard. The replication crisis really happened and we need to update on it - even "science" isn't immune to casual deceptiveness and sloppiness with the facts. Someone giving a TED-style talk on psychology research is also likely to be saying stuff that's intuitive but not based on solid knowledge, and making up a story whereby we "know" these things because an experiment was performed.

(I'm not saying that science isn't real. Science was clearly real at some point in the past, and some forms of science and engineering now seem to be making real progress even to this day. I'm just saying that not ALL contemporary "science" is clearly better than Tarot.)

IF we don't apply this epistemic standard consistently, then what we're actually doing is calling out the out-group for deception, while tolerating in-group hypocrisy. We have cultural cover in our in-group for calling out Tarot as lies, but people would probably look at us funny for refusing to associate with someone giving a talk on power poses for the same reason. This might actually be the right choice, I'm not sure - in practice it's close to what I do - but it seems important to notice when that's what we're doing.

Exploitation as a Turing test

A friend recently told me me that the ghosts that chase Pac-Man in the eponymous arcade game don't vary their behavior based on Pac-Man's position. At first, this surprised me. If, playing Pac-Man, I'm running away from one of the ghosts chasing me, and eat one of the special “energizer” pellets that lets Pac-Man eat the ghosts instead of vice-versa, then the ghost turns and runs away.

My friend responded that the ghosts don't start running away per se when Pac-Man becomes dangerous to them. Instead, they change direction. Pac-Man's own incentives mean that most of the time, while the ghosts are dangerous to Pac-Man, Pac-Man will be running away from them, so that if a ghost is near, it's probably because it's moving towards Pac-Man.

Of course, I had never tried the opposite – eating an energizer pellet near a ghost running away, and seeing whether it changed direction to head towards me. Because it had never occurred to me that the ghosts might not be optimizing at all.

I'd have seen through this immediately if I'd tried to make my beliefs pay rent. If I'd tried to use my belief in the ghosts' intelligence to score more points, I'd have tried to hang out around them until they started chasing me, collect them all, and lead them to an energizer pellet, so that I could eat it and then turn around and eat them. If I'd tried to do this, I'd have noticed very quickly whether the ghosts' movement were affected at all by Pac-Man's position on the map.

(As it happens, the ghosts really do chase Pac-Man – I was right after all, and my friend had been thinking of adversaries in the game Q-Bert – but the point is that I wouldn’t have really known either way.)

This is how to test whether something's intelligent. Try to make use of the hypothesis that it is intelligent, by extracting some advantage from this fact. Continue reading

Automemorial

In early 2014, as I was learning to be motivated by long-run considerations and make important tradeoffs, I started to worry that I was giving up something important about my old self - that some things that had been precious to me, would never quite be worth the price of holding onto, so the parts of my soul that cared for them would gradually wither away, unused, until it wasn’t even tempting to try and reconnect to going to the opera, translating classical Greek, or any of the other things in my life that I chose for their beauty but not their utility.

It turned out that I was right, though not quite in the way I expected.

This is my story. It is an honest report of that story, but that is all it is.

This is the story of how, over the past year and a half, I died and was reborn. In it, you'll find the ways I had to learn to model the world to effect this transformation. I hope that some of them are useful to you. Continue reading

Claim explainer: donor lotteries and returns to scale

Sometimes, new technical developments in the discourse around effective altruism can be difficult to understand if you're not already aware of the underlying principles involved. I'm going to try to explain the connection between one such new development and an important underlying claim. In particular, I'm going to explain the connection between donor lotteries (as recently implemented by Carl Shulman in cooperation with Paul Christiano)1 and returns to scale. (This year I’m making a $100 contribution to this donor lottery, largely for symbolic purposes to support the concept.) Continue reading

References   [ + ]

1. This phrasing was suggested by Paul. Here's how Carl describes their roles: "I came up with the idea and basic method, then asked Paul if he would provide a donor lottery facility. He did so, and has been taking in entrants and solving logistical issues as they come up."

The engineer and the diplomat

I used to think that I had poor social skills. So I worked hard to improve, and learned a lot of specific skills for interacting with people more effectively. My life is a lot better for it. I have deeper friendships, and conversations go interesting places fast. I'm frequently told that I'm an excellent listener and people seek me out for emotional support, and even insight into social conflict. But I'm told that I have poor social skills more often than before.

Not everyone means the same thing by social skills. It's important to distinguish between the social skills that are valued for their own sake – the social skills people identify themselves with – and the social skills that are a means subordinated to some other specific ends. Continue reading

Canons (What are they good for?)

People in the Effective Altruist and Rationalist intellectual communities have been discussing moving discourse back into the public sphere lately. I agree with this goal and want to help. There are reasons to think that we need not only public discourse, but public fora. One reason is that there's value specifically in having a public set of canonical writing that members of an intellectual community are expected to have read. Another is that writers want to be heard, and on fora where people can easily comment, it's easier to tell whether people are listening and benefiting from your writing.

This post begins with a brief review of the case for public discourse. For reasons I hope to make clear in an upcoming post, I encourage people who want to comment on that to click through to the posts I linked to by Sarah Constantin and Anna Salamon. For another perspective you can read my prior post on this topic, Be secretly wrong. The second section explores the case for a community canon, suggesting that there are three distinct desiderata that can be optimized for separately.

This is an essay exploring and introducing a few ideas, not advancing an argument. Continue reading

Be secretly wrong

"I feel like I'm not the sort of person who's allowed to have opinions about the important issues like AI risk."
"What's the bad thing that might happen if you expressed your opinion?"
"It would be wrong in some way I hadn't foreseen, and people would think less of me."
"Do you think less of other people who have wrong opinions?"
"Not if they change their minds when confronted with the evidence."
"Would you do that?"
"Yeah."
"Do you think other people think less of those who do that?"
"No."
"Well, if it's alright for other people to make mistakes, what makes YOU so special?"

A lot of my otherwise very smart and thoughtful friends seem to have a mental block around thinking on certain topics, because they're the sort of topics Important People have Important Opinions around. There seem to be two very different reasons for this sort of block:

  1. Being wrong feels bad.
  2. They might lose the respect of others.

Continue reading

Mic-Ra-finance and the illusion of control

Microfinance charities make small loans to very poor people. The Unit of Caring has a post up answering a reader’s question on microfinance:

intomeans asked: So based on your post about microloans, do you think it's better to give $1000 to one person one time, or to lend it out through microloans and then, as the money's repaid, keep relending it to other people indefinitely? That's the main argument that pushed me to lend through microloans (in addition to giving to charities like AMF), and I don't think Givewell's analysis addresses that.

I think it’s better to give $1000 to one person one time.

The business model of micro loan organization is to loan $1000, take back $1200 if the recipient is able to pay it back, hope the additional $200 covers the money they are spending on identifying recipients and ensuring repayment, and loan $1000 again.

That this constitutes ‘the money doing good indefinitely’ is listed on GiveWell ‘six myths about microfinance’, which also links this really useful article. Basically: there is a lot of overhead involved in selecting and monitoring recipients, such that every time the loan is re-loaned out a significant fraction is lost. Overhead isn’t inherently bad but even if all the loanees repay the loans, it’s misleading to suggest that with some fixed amount of money to start, a microloan charity could make loans indefinitely. And not all the loans are repaid. (And sometimes, charities that report really high repayment rates, higher than American banks achieve, are a sign they have a lot of coercive power to get their money back, not a sign that the program is going brilliantly.)

So, a thousand dollars enables more than one thousand-dollar loan. But almost certainly less than ten, and some of those people repaid at great personal cost and ended up in a worse position than they started in (because they didn’t understand the terms of the loan or similar.)

This seems true as far as it goes - but even if the empirical premise were true, this case for microlending seems pretty weird. This argument for microlending is that each time you make a loan, you help the borrower - and because they typically pay back the loan, you can keep relending the principal, thus continuing to help people.

Let’s think about it disjunctively. For any microloan recipient, either they have a good way to invest the money, or they need the loan for short-run consumption. If they’re financing consumption, then either having to pay back the loan puts them even worse in the hole, or they’re using it for consumption smoothing and what they really need is savings.

If, on the other hand, they have a good way to invest the money, then they might pocket a profit even after paying back the microloan, which can then be lent out to someone else with an investment opportunity, a clear instance of “the money doing good indefinitely.” But what happens if you keep not making them pay you back? If they reinvest that money, then that’s also an instance of the money doing good indefinitely! Reinvestment of earnings is a thing, even in poor places, and so is helping one's neighbors.

When deciding between microloans and cash transfers, you’re not deciding between doing good one time and doing good indefinitely. The only thing that makes microloans feel like the impact is longer-lasting is because you can feel like you’re holding onto control for longer. The charity doesn’t just give the money and go away - at the end of the loan’s term, it gets to decide who gets the money next - and again, and again, and again. [UPDATE: Per Paul's comment below, there are some reasons to think that this kind of control control can be a good thing. My problem is with the assumption that it is.]

You the donor don’t even have the control here. You aren't lending to people you know or have otherwise personally verified can use the money. The only question you get to decide is: should your donation be administered by a big official charity? Or should it be administered by some random person in a poor village who knows the people and situation there? If they end up with a lot of money - and microlending would be a good idea - then wouldn’t the recipient of your cash transfer be motivated to do their own microlending?

The first option, picking a charity to administer your donation, might do better at weeding out obviously irresponsible recipients, but on the other hand, it comes with massive overhead costs that likely outweigh this benefit.

(As usual, the form itself is not the problem. I expect there are cases where microfinance is in fact helping. I expect that most of these are for-profit. The problem is the automatic deference to the form.)

I’m embarrassed not to have noticed this obvious flaw in the argument for microloans earlier. This seems like the sort of pathological thinking Sarah Constantin was trying to describe in Ra. Long-run wealth accumulation due to cash transfers doesn’t count because it’s in the hands of some specific individual as real concrete things. Repeatedly re-loaned microcredit keeps counting because it stays under the control of a large respectable institution, as the abstraction of money.

This is bonkers. It has little to do with doing the most good, and a lot to do with the worship of smooth, respectable official-seeming vagueness. Where else am I still making this mistake?