(Talk given on Sunday 21st June, over a zoom call with 40 attendees. eukaryote is responsible for the talk, jacobjacob is responsible for the transcription.)

Ben Pace: eukaryote is the next speaker. eukaryote is a famous writer of posts such as Naked mole-rats: A case study in biological weirdness. What are you going to talk about, eukaryote?

Talk

eukaryote: I'll talk about a mental construct I'm calling 'Missing Dogs' that you might find useful. Like all good rationality techniques, this one is going to start off with a Sherlock Holmes anecdote.

eukaryote: So, Holmes and Watson are hired by some rich guy, because he has this barn where he keeps all his animals, his horse and his dogs and his sheep, and he knows that someone's been sneaking in there at night and messing with his prized race horse. So there's two entrances where the guy can get in. And Holmes and Watson think they've gotten the entrance the intruder is using, so they stay up all night. They don't hear anything, but when they check the next morning, someone has gotten in and messed with the horse.

eukaryote: So they know two things. One: the intruder must be using the other entrance. Two: Holmes says to Watson, "I think there's something else we should be considering here, which is the curious incident of the dog in the nighttime." And Watson says, "Holmes, what are you talking about? The dog did nothing in the nighttime." And Holmes says, "Yes, that was the curious incident." Because he's put together that the dogs in the stable would have barked if the stranger had come in and just messed with them. So the intruder must have been someone the dogs knew. And that is the deduction he can make from that.

eukaryote: So, this is what I'm calling the 'Missing Dog' — when you can learn something interesting from the fact that something isn't there. You might call it 'a conspicuous absence'.

Missing dog, drawn by eukaryote

eukaryote: That’s the image I spent several minutes drawing last night, so I'm going to make you look at it.

eukaryote: Here are a couple of other examples. The Fermi paradox is a really big one. Taking a few basic axioms: we think life forms on its own, we know there are billions of stars out there... where is everyone? We don't know but that's a really interesting question, the kind of question that defines a species, in my opinion.

eukaryote: And then there's some other instances where... for example, people who are blind at birth. We've never found a person who was blind at birth who also has schizophrenia, which is statistically very improbable that this would have never happened. So we learn some interesting things about how we think schizophrenia might develop in the brain.

eukaryote: Or how whales, despite having millions of times the number of individual cells that mice have, don't seem to get cancer much more often than mice. Which is weird. That tells us something interesting about cells and how cancer works.

eukaryote: This has shown up in a couple of research projects which I'll talk more about, if someone wants. But I think the point is that this is a pretty useful tool to keep in mind. So, okay, if you're trying to use this, something to note, I think this usually generates questions, not answers. So it's a way of exploring models more than directly getting to something.

eukaryote: And then, if you were trying to think about where these show up in relation to a topic, well the problem is that they're kind of hard to notice. You might try assuming that there is something you're missing like that, and just asking yourself from basic principles what that might be. I'm not really certain... I have run into cases where people have asserted that we just don't know or, the fact that these examples are missing tells us something interesting, doesn't it? And then I looked and there actually were examples of it, the people just didn't know what they were talking about.

eukaryote: So, if you think you find a missing dog, first check that the dog is actually not there – it may just be very quiet. Once you've done that, I can't really give you that much of a road map, but I think the next thing to do is to try to pin down why you think it's unexpected. Often when I've done this, I've found that my guess about what was going on, or the model of the situation was actually extremely simple. But don't stop there because I've also found that this is sort of an unusual mode of thinking and even if your really simple model shows up with a big problem, or shows up with this strange question in it, often that is just still not accounted for, even if it seems like I may not know anything about this if I don't understand what's going on here. You might not and then you will know something interesting about the situation from finding that out. Or you might just be onto a new unanswered question. That can happen too.

Q&A

Ben Pace: Thank you very much. You were successfully under the five minute limit.

Ben Pace: Jacob, you had a question? 

Jacob Lagerros: Yeah, I'm pretty curious about the slide with all the examples you didn’t cover. Something about insects and why aren’t we dead.

eukaryote: Oh, got you. These are blog posts I've written which I will just run over in the two sentence version and how they relate to this.

eukaryote: Yeah, so insect extinctions, there's a fact that is going around that recent studies have found, in really wild locations, that insect biomass has dropped by 98 percent over the last, let's say 50 years. There have been a few different studies, all of which show pretty similar results. People are like, oh no we're destroying the environment, we're going to die. And like, maybe. I was thinking, I feel like if you told me that as a teenager that in 10 years there will be 90 percent of the insects will be gone, then I would be like, oh my god, we are dead, this is the end of civilization, I should just start drinking and not go to college right now because clearly we're not getting anywhere. And yet, society seems to be going along pretty fine. Just based on what I knew about insects, that seemed completely impossible to me.

eukaryote: So I still don't know why that is, but I found out some interesting stuff about it, and I don't think anyone else has a good answer to that question also.

eukaryote: The Germy paradox is... so I spent the two years doing a masters in biodefense and learning about bioweapons and oh no, it's so easy, anyone can go to a lab and just create smallpox or it's real easy to get anthrax or whatever. And I think, okay, there have been a small number of attempted instances of bio-terrorism. There have been these huge weapons programs, and yet no one has actually used a bio weapon as a tool of war against another country since the 1940s approximately,. Whatever — the point is, it's super rare. So if these really are so cheap and deadly, and easy to make, where are they? That's the Germy paradox. And I wrote a many-page sequence about it that you can read if you want.

eukaryote: Now I feel like I understand why we have not seen those things. So that's the summary there.

Ben Pace: Interesting. Thank you very much. Anna, do you want to ask a question?

Anna T: Yeah! Hi, I was curious, eukaryote, you mentioned that sometimes you might think that you're seeing a missing dog situation but it is, as you said, a very quiet dog. Do you have any examples of that kind of thing?

eukaryote: So, I really enjoy thinking about cheery and uplifting topics in my free time, another thing I've heard passed about, biological-esque risk, is that a disease can't kill off a species. We just don't have examples of that. Species don't die from that. So, we don't need to worry about this as a species, and these diseases will never evolve naturally, blah blah blah, we're fine. And I'm like, wait, hang on... and I checked and there are actually dozens of examples of diseases killing off a species. Not on the scale that we should necessarily worry about it. It's a whole thing. Whatever. But in that case, those people just hadn't done the research. That was their quiet dog.

Ben Pace: Cool, cool. orthonormal, do you want to ask a question?

orthonormal: So, have you thought about adversarial reasoning as a way to bring these out? I've heard, and you're the biologist here, that the one useful bi-product of creationism has been coming up with a few missing dogs for biologists to look at.

eukaryote: Oh, I love it! Yeah that seems super fruitful and I’ve not thought in depth about how to do it. I gave some examples about ways of doing this with yourself, but if you can find a willing partner or just someone who disagrees with what you say and you can fight them about it, that seems pretty good. I haven't thought too much about it. But, yeah, seems fruitful.

Ben Pace: All right, thanks a lot, eukaryote.

New to LessWrong?

New Comment
3 comments, sorted by Click to highlight new comments since: Today at 5:59 AM

One example I’ve experienced is reading scientific papers. I have had the experience where I think “why haven't they presented this sub-result in this intuitive way?”. Sometimes this is just incompetence but at other times it leads me to find that the particular result in question goes against the hypothesis of the paper and that result is included only in the footnotes/supplemental material.

Nitpick: I think there's a minor transcription error, in that "biological-esque risk" should read "biological X-risk".

whales, despite having millions of times the number of individual cells that mice have, don’t seem to get cancer much more often than mice.

Is this all mice, or just lab mice? I ask because of Bret Weinstein's thing about how lab mice have abnormally long telomeres, which causes them to get cancer a lot more frequently than normal mice (though in googling for the source I also found this counterargument). So is it that whales get cancer less often than we'd expect, or just that mice (or rather, the mice that we observe) get it a lot more frequently?