Gideon Yaffe, Brain Injury Survivor

 

92-131. Gideon Yaffe

SPEAKERS

Gideon Yaffe, Will Bachman,

 

Will Bachman  01:03

Hello and welcome to the 92 report conversations with members of the Harvard and Radcliffe class of 1992 I’m your host. Will Bachman, and I’m delighted to be here today with Gideon. Yaffe Gideon, welcome to the show.

 

Gideon Yaffe  01:17

Thanks. Thanks. Happy to be here. So Gideon,

 

Will Bachman  01:21

tell me about your journey since graduating from Harvard.

 

Gideon Yaffe  01:25

Ah, well, let me see so well. So the first thing I did after graduation was I gotten this crappy Volkswagen Jetta and drove across the country with my then girlfriend, now wife who was also in our class, Sue Chan, who’d be a much better person for you to interview. Will but

 

Will Bachman  01:54

ready, I’m ready to tell him. Well, yeah,

 

Gideon Yaffe  01:58

yeah. So we drove across the country and we went to San Francisco. Neither of us had jobs or anything. I’d applied to graduate school in philosophy, which had been my undergraduate major. I’d applied in my senior year and at college, and I didn’t get in anywhere. So I was like, feeling very despondent about that. But so we got to San Francisco, and I asked her to marry me, and she said, Yes, so that was nice. So we were already on our way to getting married at like, however old we were 12,

 

Will Bachman  02:35

so you were a no job prospect, no grad rejected from grad school philosophy major. Yeah, that like, what a powerful prospect she was signing up for Exactly,

 

Gideon Yaffe  02:49

okay, yeah, that’s right. So that was the parents were thrilled. All right, yes, yeah, her parents were also thrilled, because I was white, and that was, like, the last thing that they wanted her for her. But, yeah, but yeah. So then, yeah. So then, well, and then, to even make it worse, in terms of my prospects, my first job out of college was at a pet store, and it was, it’s actually a really cool pet store that I recommend. It’s called the East Bay vivarium. It’s in Oakland, and it’s, it’s a reptile pet store, and that had been like a big hobby of mine, keeping reptiles, snakes and lizards and stuff, since I was a little kid. So I was really into this store, but my job there really was not good. So I mean, I was making think I was making $3.75 an hour, which doesn’t go far even in 1992 and and

 

Will Bachman  03:57

then I have to do the math to see how that compares with my navy instant salary given, you know, like it depends on how many hours per day you want to count it if you’re at sea working. But, okay, but, but either way, we were not, we were not breaking the bank there.

 

Gideon Yaffe  04:13

Okay, yeah, right, I actually think, yeah, your job. So my job mostly, mostly consisted of murdering rats and bunnies to feed to huge snakes that they had there, and well as as well as, like, counting bugs into jars to sell to little kids who had lizards at home. So it was not a great job. But, yeah, so I didn’t work there for very long. But yeah, so I worked. I was working there for a little while, and I got a then I got a job at like a computer magazine, and then I was and I was mostly just like sitting in coffee houses and feeling sorry for myself, I think is the main. Thing I was doing, and then you

 

Will Bachman  05:02

had applied. You had applied to grad school in philosophy.

 

Gideon Yaffe  05:06

Yes, in my senior year college, I had Okay, and then I but so one of the other thing I was doing, and I don’t really know why I applied in my senior year, honestly, I was like, I think I was just wasn’t sure what I wanted to be doing with my life, and I liked my philosophy classes, so I was like, Maybe I should just do more of that, like that was out as well as articulate as my position was at that time. But so then I, I was, you know, just hanging out a lot in San Francisco. And I was, I was reading, a really hard, long book, Proust’s remembrance of things past, which is like eight volumes and many pages and incredibly difficult so, but I was like, I was finding that incredibly rewarding and amazing to read. I really loved it. So I thought, Well, what do I really want to be doing with myself? I want to be like, thinking really hard, like, that was what I wanted to be doing. So I thought, okay, I should just try again apply, apply to grad school again. So I did it. I applied again to grad school, this time again in philosophy, and this time I got in a lot of places. So I was sort of like, yeah. So I was then thinking about where to go to college, where to go to grad school, and I decided to go to and I really loved the bay area where we were living already, and so I I signed on to go to Stanford. So I went to Stanford, thinking that I was gonna study like philosophy of art and esthetics and stuff like that. But I found very quickly, I found a professor there who I adored, Michael Bratman, who worked in philosophy of action, like basically connected to the Free Will problem, but related to other kinds of questions about the nature of action and agency and attention. And so I worked with Michael, and that was really, really super rewarding. And I also got interested at that time in history of philosophy. And yeah, so I wrote, I wrote. I wrote half my dissertation about John Locke, and then half about contemporary problems, about the Free Will problem. And then I went on the job market in academia, which was like, I mean, it was a brutal job market. I mean, it was insanely rare for people to get jobs. So the first time I went on the market, I got some interviews, but then I didn’t get a job. Then I went on the market again the following year, and I that year, I didn’t get a tenure track job, but I got a one year job, which was felt like an extremely lucky thing that was at UCSD, so, so I’m and by this time, Sue was working somewhat in the film industry. And she so she was so it kind of made sense for her to make the move to LA and be in the and where, she could be in the in the union and stuff for the kind of work in the film industry she was doing. So we moved to LA and I was commuting down to San Diego. I had a place there to crash, and I was on the job market again. And I was teaching a ton and I so I went on the job market again, and I was sure that if I didn’t get a tenure track job this time, I was just, like, gonna pack it in and, like, just do something else. I wasn’t sure what, but I knew it wasn’t going to be, you know, I wouldn’t go back to the East Bay vivarium, but I wasn’t sure where I was gonna what I was gonna do.

 

Will Bachman  08:56

But philosophies like, on the side to sort of, yeah, right, exactly.

 

Gideon Yaffe  09:00

Yeah. Like, I mean, so this was, like, a classic joke at the time, was that this, this guy, apparently, I don’t actually know if it was true, but he used to tell this story about how his daughter would go to like, kindergarten, and she’d say to her friends like her, she had some friend who asked her, so what does your dad do? And she said, he’s a philosopher. And her friend said, Oh yeah, my dad’s unemployed too. Yeah. So, you know, I don’t know. I mean, I right, but so I went on the job market again. That was my third time on the market, and lo and behold, I got a tenure track job at University of Southern California, USC, which was, like an amazing stroke of luck. You know, I was already living in LA, and it was where, where Sue’s work was likely to be. And, you know, so I was, I was jonesing about that job. I really liked the job. And then I, I. I started there in 1999 1999 2000 school year, and let me see. So one of the things that I did early on there was that I taught a philosophy of law class because it needed to be taught. And I also wrote a paper, a long paper about addiction, which was super interesting to me for a variety of reasons, some of them, you know, because I knew a lot of addicts, but also because I was interested about freedom and addiction, and the way in which addiction would undermine freedom and free will, was one of my main interests. So I wrote about, I wrote this paper about addiction, and then I can’t remember the exact order of events, but I ended up being asked to be part of Oh no, no. Actually, the first thing that happened was that, yeah, so a friend of mine was a victim of a serious crime. He was a victim of a carjacking, and largely out of my, you know, desire to be helpful or supportive or something to my friend. I read a lot about carjacking, and which why she didn’t make it particularly helpful to my friend, but it was nonetheless, I don’t know, made me feel better or something. And then there was a, yeah, there was a legal problem at the center of a lot of carjacking cases that involved intention. And I got really interested in this legal problem, so I wrote a paper about that. So these were these two things. I’d written this paper about intention, about

 

Will Bachman  11:37

what’s the confusion of the intention The carjacker wants to take your car like that seems pretty intentional. What?

 

Gideon Yaffe  11:44

Yes, well, here’s confusion. Do you want it? You really want this kind of detail? Will

 

Will Bachman  11:49

I do? This show is all about detail. Like it seems clear to an average lay person, intentional. They want to take your car. Yes, you prefer not to give it to them, but they say, have a gun. Interestingly,

 

Gideon Yaffe  12:03

the so car theft is, is a car is, is what you know they want to take your car, and so they take your car, right? But carjacking is defined in the federal law. One of the elements of carjacking, one of the things that needs to be proven against somebody to make them guilty, carjacking is that they had the intent to kill or seriously injure the driver. Oh, okay, so if they, if what they intended was to take the car, yeah, that’s then they wouldn’t be guilty of carjacking. They’d be guilty car theft. Yeah, what they they have to also intend to kill the driver. Oh, I see. So now a lot of carjackers had argued. Well, no, I mean, I was just, I did intend to kill him the driver, but I only intended to kill him if he didn’t give me his keys. Yeah, but that’s like a conditional intention. That’s not the intention to kill or injure the driver. That’s the conditional intention to kill or injure the driver if he doesn’t give me his keys. So the question was whether, under what conditions, is a conditional intention enough for the crime, when the when the pet, when the statute in question calls for intention that’s unconditional. So this is this weird, little funny mens rea problem. So I wrote this paper about this, which, which I called conditional intention and mens rea. So there you go. You should look that one up. And yeah. So that paper got me feeling like, wow, philosophy of law is really what I should be working on in a more serious way. This is, like, this felt like a real world problem somehow, like it was going to make a difference. Like, I mean the difference. So carjacking, you know, would be typically sentenced to, like, 25 years in prison, where car theft would be much, much lower, lower sentence than that. So there’s, like, some 20 year difference depending on whether the person who has the conditional intention is actually just as bad as the person who has the unconditional one. So we want to know, like, when is that the case? So anyway, I’d written about that it was a philosophy, a philosophical problem and philosophy of action, but made me feel like, you know, legal stuff was important, and so I wanted to learn a lot more about the law. And so I coerced USC into giving me a year in the law school there. So I was, I had a year as a student in the law school there on USC dime, which was great. Okay, so you’re a student, yeah, so I was Yeah. So I was actually sitting in a contract class when the news came down that Bush had won the Bush v Gore case, like. So that gives you an idea of how long ago this was. But, yeah, so, yeah. So then I but. During that year, taking all these law classes. So I basically did the first year curriculum of a law student that year, okay? And I, and I got to know, you know, all the professors in the law school, and I was all, and I was a very unusual law student, because, you know, obviously I wasn’t interested in being a lawyer, and so, and so. Then at the end of that year, the law school offered me the job at USC, so I would have been partly in the philosophy department and partly in the law school. So I was like, Sure. So I took that and then, and then. So then I started being I started teaching. Like part of my, part of my, part of my teaching load was in the philosophy department, then part of it was also in the law school. So I started teaching criminal law there in the law school. And then I teach, continue to teach philosophy classes in the philosophy department. And then let me see. So then the next thing that happened, I guess, is that the MacArthur Foundation launched this thing called the law and neuroscience project, which was a big project. I was they dedicated, I think, 10 million originally, to it, to bring together a lot of different people in from philosophy, from law and from neuroscience, to talk about the ways in which neuroscience might be, might be relevant to legal proceedings of various kinds, especially in criminal law proceedings. And because of my work on addiction, and because of my law and philosophy kind of connection, they invited me to be part of this think tank thing. So I started doing that. The first phase of that lasted, like, I guess, three or four years, and then there was a second phase, when the size of the thing shrunk a lot. And I think it shrunk from like 60 to like a dozen. And I was part of that second group, too. And the So, what that meant was that, for there was a I was kind of like thrown into collaborative engagement with neuroscientists on various problems, on various kinds of experiments that you might run in neuroscience, that might you know be of use to the legal system in various ways. So I started to do that, and what kind of sunk into me as part of that was that I really needed to learn more neuroscience, and I already had my year as a law student. So hey, why not so? So I I got a grant to to support that. So then I had another year as a law student, as, excuse me, as a neuroscience student at Cal Tech. So that was another year without teaching, where I was like, yeah. I mean, it’s like, really, yeah. I mean, these are academic sinecures, I suppose. But it was awesome. So I had this great year at Cal Tech where I learned a lot of neuroscience, a huge amount. And I also was collaborating with various people through the MacArthur project. And then I these launched some collaborations through with Caltech people as well. So I started to be kind of like the middle author on some neuroscience papers where, you know, the primary authors were the neuroscientists, but we would design the experiment together with some kind of legal question in mind, and then we they’d run the experiment to collect the data, and I kind of roll back in to write papers and things like that at the end. So, yeah, so I started to do that. And then also around this time, I can’t remember exactly the order, I wrote this book about attempted crimes, about the crime of trying to commit a crime, but failing which, you know, is a law. I mean, there’s like lots and like lots of laws governing attempted crimes, attempted murder, you know, tempted attempted assassination of the of a presidential candidate, for instance. You know many examples of this. So I wrote, I wrote a book about that. And then. So this combination of things like, it’s

 

Will Bachman  19:01

actually like a double failure. If you go to jail for failing, it’s one thing to like crime,

 

Gideon Yaffe  19:10

Yep, yeah. No, I know. I know. I’m well, in fact, actually. So some, some of these, some of these failures, are even worse than your than you think, because some of them, a lot of them involve, don’t involve, you know, missing the presidential candidate and striking only his ear. I mean, a lot of them involve things like sticking pins in a voodoo doll, trying really hard to kill the judge, or whatever. That’s one real case, actually. That was a Florida case,

 

Will Bachman  19:43

wow. Or just, like, disorganized, you know, conspiracies or something, yeah,

 

Gideon Yaffe  19:49

1. I mean, the the number of cases that you read about attempts where they just really just don’t seem like competent people at all, is, um.

 

Will Bachman  20:00

Striking, yeah, yeah. Like, one thing to go to jail for, for being a successful bank robber and getting caught, but to, like, fail at being a criminal, that’s, that’s, yeah, that’s really, you’re just a real failure. That’s, that’s very sad, like,

 

Gideon Yaffe  20:14

you know some, I mean, here’s a, here’s a statistic for you, in a very, very large percentage. I don’t know the exact number, but it’s a very large percentage of people who are serving sentences in the United States currently in prison. What among their things of which they were found guilty, are attempts? Oh, really,

 

Will Bachman  20:33

it’s interesting, very large percentage. Yeah. Wait, so the voodoo doll thing, so that would count as, like, attempted murder if you were using a pin on a voodoo doll like it

 

Gideon Yaffe  20:43

doesn’t, but then the reason why it doesn’t is because it’s what’s called inherently impossible attempt. Oh, interesting. Then, you know what? Which attempts are inherently impossible? You know? I mean, that’s so that’s one of the things that I wrote about in the book, about what makes what makes an attempt, possible versus impossible, and what kinds of considerations go into the calculation. And that’s just one example. But yeah, I mean, so there’s just, like, lots of these funny little problems about attempted crimes that I that I wrote this book about. So then the combination of, like, wait

 

Will Bachman  21:18

a minute, wait a minute. Gotta promote it here. What’s the title of the book,

 

Gideon Yaffe  21:22

the type the book is called attempts, right?

 

Will Bachman  21:26

Gideon Yaffe, right? Yes,

 

Gideon Yaffe  21:27

exactly. That’s the one. Yeah. I mean, yeah. The problem with most of my writing is that it’s unreadable. So, you know, I don’t know it’s, it’s like, I write these things for like, seven or eight people. And you know, you really

 

Will Bachman  21:40

know how to promote this. Gideon, a master promoter here. Oh, yeah, you really don’t want to read my books. They’re unreadable.

 

Gideon Yaffe  21:50

Yeah, yeah. I mean, you can talk to me about them. They’re They’re fun to talk about, all right, yeah. So anyway, so because of the MacArthur thing, and then the attempt book, and because of my legal thing, and I don’t know so Yale suddenly comes calling. So they kind of like, called me up and were like, Hey, you want a job here. There was a little bit more to it than that, but more or less that was what happened. And so I, I, I guess I was, like, flattered into it. I mean, like, I really liked my job, and I really loved LA, but i don’t know i It was very prestigious, fancy job, and it would have been entirely like my primary appointment would have been in law school. But I could have had any kind of connections I wanted to other places in the university, and so I, you know, I don’t know I did it. It didn’t hurt that. At the same time the film industry, which is what Sue was doing, was leaving LA. It was really moving away from LA. So almost all of her work was not in LA. So, she, I mean, so, yeah, so she, she was working in Bulgaria at that, at that time, at one on one project that was the 300 movie and, you know, and so, and then, since then, I mean, she, I’m not sure she’s ever worked in LA again since that period. So she works all over the world, but not in LA. So it was kind of like I didn’t really need to be there. We had a kid, and the school situation in LA was terrible, and so we thought, you know, we could make it work wherever. So why not go to this cool job for me? So yeah, so that’s what I did, but I kept my house in LA. So that was one of the very first things that I said to the dean at Yale when I when he made me the offer was, you know, if I have to give up my house in LA, I’m not going to take this job. So they made it possible for me to keep my house in LA, so that was nice. And so what that means is that now I’m we’ve moved into a situation we’ve been now doing it for a dozen years, where we spend the academic year in New Haven and we go back to LA whenever school is not in session. Um, oh, wow. Okay, so that’s our BI coastal arrangement. And yeah, and then my daughter, our daughter, would be in school in New Haven, and then would spend the summers with us in LA, yeah. So yeah, and yeah. So that’s what’s been going on professionally, I guess. I mean, I

 

Will Bachman  24:38

Yeah, it sounds like you are touching on, you know, I mean, philosophy, neuroscience, law. I’m curious, to what degree are you involved in all of the generative AI stuff? It sounds like you’re a perfect advisor to one of these companies.

 

Gideon Yaffe  24:54

Well, yes, I’m not advising any companies, but I am involved in that stuff because, let’s tell us. About that? Well, one project that I’m now just, we’re just getting started, and we need funding, appropriate funding, but so a neuroscientist who I’ve collaborated with some named URI Maus, he is starting this big project on AI intentions, basically, AI agents when you haven’t so you know one, one example of the kind of problem that this would, that we hope to solve are things like, if, if a machine runs down a person, or kills a person. There seems all the world of difference between the self driving car that does that and the drone that does that. So what’s the difference? Well, what’s very tempting is to say, well, the drone intends to kill the person and the self driving car doesn’t intend to kill the person. It’s just an accident. But, you know, these are just machines. So what’s the difference between the between this, between these two things, and between the drone and the self driving car? And what is it for a machine to have an intention? And increasingly, this seems like something that we should know the answer to. So that’s one example of the kind of AI stuff I’m kind of involved in. And then the another is that there’s a so Sergey Brin, the the billionaire. He has a daughter who suffers from autism, some very, very severe form of autism, and so he’s funded this giant group at Yale that I’m part of that will that is trying to build AI models of the brain of a particular person, and then use that and then see how that model will respond to various forms of surgery, to imagine a change to the model that would be then duplicated in a surgical change to the human brain, and then run it forward, and you get a much better picture of what might happen in that brain over time thanks to the surgery, then you without ever having to cut open anybody’s head to Try it out. So I’m also involved in that one to some degree, yeah, so I don’t know. I mean, we’ll see whether, where all this goes, I’m not sure exactly. So, I mean, you know, a lot of this there’s, what’s interesting is that there’s, like, these, these huge overlaps between neuroscience and AI study. And exactly, what role I’ll be playing in any of that. I don’t know, but yeah, I might need to spend another year learning about AI. Maybe I can do that another

 

Will Bachman  27:51

another year slacking off and going to school. Yeah? What? Okay, so you have thought about this more than any other person I’ve probably ever met. I mean, you did a PhD on it of free will. So I’m curious to hear your thoughts about, I mean, we’re kind of potentially getting close. Maybe it’s an asymptotic close. But how would we know if these llms of chatgpt, you know, 7.2 or 8.3 you know, are they? How will we know if they’re conscious or have free will or have intention? So, yeah,

 

Gideon Yaffe  28:27

well, so, so consciousness is interesting, because I think that they’re probably just never going to be conscious. But it depends a lot on what you think consciousness is. But the what’s interesting actually,

 

Will Bachman  28:39

why do you say that? Why do you say that?

 

Gideon Yaffe  28:42

Oh, well, it’s just that. Well, it really depends on what you think consciousness is. So so some so for some way of understanding, one way of understanding consciousness is that it involves self representation of certain kinds. That is, you need not just having the thought, but thinking about the fact that you’re having the thought, so you have the second order thought about the thought. And certainly, if that’s what’s all that’s required for consciousness, then without a doubt, these llms can be conscious. But there’s another way of understanding consciousness, where it involves what are, what are sometimes called qualia, ways of feeling so way. So what it’s like from the inside? What the What is it like property? So, what is it like to to feel a co to get an ice cream headache, or what is it like to be tickled under your under your chin? Or these kinds of, what is it like? Questions, what is it like in the in the classic example, what is it like to be a bat? There must be something that it’s like to be a bat. So you might think that, but is, what is it like to be an LLM, it might be that there is no more to what is it like to be an LLM, there is to. What is it like to be a toast? Or there’s just sort of nothing that it’s like. So that at least seems possible to me. But how would you know? I mean, it’s very, very hard to know, right? So this is, how do you know even about each other, right? I know that you have that there’s something that it’s like to be will Bachman, I

 

Will Bachman  30:20

was gonna ask you the same question, except I was gonna say Gideon Yaffe, okay,

 

Gideon Yaffe  30:24

right. So that’s called the problem of other minds in in philosophy. And it’s not an easily solved problem. It really is not. So the same problem of, you know, the weather and what is it like to be an LLM, and is, is we encounter that exact same problem of other minds with respect to the llms, for sure? So you know that that in some ways. So my hunch is that there’s nothing that it’s like to be a toaster, and that there’s nothing that it’s like to be a more complicated toaster, even if what you’re doing is generating various kinds of sequences of words, but that that I don’t know that for sure, but it also seems to me to be I don’t know either thing for sure. Maybe there’s something that it’s like to be a toaster. It just seems to me kind of unlikely, given the small amount that I know about the way in which human consciousness arises out of, out of stuff in our heads. We don’t have anything quite that complicated going on inside the heads of of, you know, any other, any AI being, but, you know, maybe, but, but this, but the question about whether or not aI have intention is actually different, I think, from the question of whether or not they’re conscious. Okay, that is intentions. So intentions serve to organize behavior in a whole variety of interesting ways. So I intend to go to the store to I intend to have, excuse me, I intend to have ice cream for ice cream for dessert. So I go to the store, and because I have the intention to have ice cream for dessert, I get ice cream. I don’t think through over again what to have for dessert, because I’ve already decided I’m having ice cream for dessert. So this means that one of the that one task that I might have, that one task that I tip them that I sometimes have to engage in, I don’t have to engage in this time. I don’t have to think through the pros and cons of various possible desserts, right? Because I already decided that earlier. So I have the intention, and it serves to to commit me to getting to getting this later, without having to think through what the what the pros and cons of that are now, that now. So now, do do AIS have that kind of thing? Do they organize their behavior over time in that kind of way? Do they make decisions now so they don’t have to think about them later? Or do they not? That seems to me one relevant question to whether or not AIS have intentions or similarly like, for coordinating like. So we’re on a call right now. I intended to be on this call. You intended to be on this call. I knew that you intended to be on this call, and that was why that was part of the reason that I didn’t that I show up and I showed up at the call because, hey, right, Will’s going to be there. It coordinates our behavior with each other. So are AI’s going to do that kind of thing with each other? Are they going to be in a position where they think, oh, that other AI is going to be at the corner at this time? How do I know that? Because I know what it intends now? Or do they not reason that way? So the these so interpersonal, not just excuse me, in, in, yeah, interpersonal, as opposed to intrapersonal organization, is another thing that intentions provide us with. What kind are those kinds of organization available to AI? I don’t know, but that’s that that feels like a more tractable problem than the question of whether or not AIs are conscious. Okay, so I mean, then, similarly, with freedom, it might be as hard as the consciousness problem, depending on again, depending on what you think about, what you think freedom is,

 

Will Bachman  34:10

okay, so it sounds like, what? What are some other examples where, like the law and intention and AI intersect, you talked about like a vehicle, a self driving vehicle. Yeah, might hold that vehicle liable differently if it was just super rainy and it was skidding along and just skidded into the person, versus like, turn to the right and went off onto the sidewalk to hit the person, or, you know, just missed them or something. But what are some other examples that are cutting edge here of intention and AI and yeah,

 

Gideon Yaffe  34:52

I mean, well, so, so, so there’s other other various places in law, where. Intentions matter, and then there’s going to be a question about whether or not AI can have intentions for that purpose. So for instance, one of the things that’s relevant for making a contract with somebody else is that they intend to be bound by the contract that they sign. The contract intending to be bound by it, as opposed to sign aid inadvertently. So now we’re likely, in fact, we may already be making contracts with AI and AI. Now, when AI contracts with us, are those contracts binding or not, it will. It would depend in part on whether or not the AIS have intentions, and what does what the content of those intentions are? I don’t know. I mean, I don’t know how to think about that, but that seems to me be another another area where AI intention might matter. Another place about intention is with the interpretation of texts, of legal texts. Yeah, we have aI writing text, and then writing law. Do we need to think about what the intentions are, with which they were with, with those words were written.

 

Will Bachman  36:20

So do you expect? So is there a need then, to kind of rewrite some of these laws or update them to, you know, address this because say, hey, intention doesn’t matter. It’s just it’s if you sign it or not, you’re bound by the textual language. It doesn’t matter the intention of the drafter, or your intention when you signed it, you sign it, tough luck.

 

Gideon Yaffe  36:43

Yeah. I mean, there might be all kinds of, all kinds of changes that we might need to have. So it depends a lot on what intentions aI have, or if they even have intentions. And that’s the first question, which, you know. I mean, that’s where we’re hoping this project will go to try to figure out, well,

 

Will Bachman  37:05

I imagine it’s tough to really understand what a human’s intentions are, although maybe you have evidence. You can say, well, we caught a voice recording of the person talking to their you know, conspirator, you know, maybe in that case, but with AI, at least, my understanding is it’s such a black box of all this. LLM kind of just weightings and stuff. There’s no kind of you go in there and it’s like, oh, my intention equals, you know, and then a quote, so, how are, how would you, having studied it and our neurosciences and thinking about identifying what is the intention of the AI, given that when you look inside, it’s just kind of a black box,

 

Gideon Yaffe  37:50

I think largely the hope is going to be to try to try to do it by looking at the role that their representation plays in guiding their behavior. So, so, so, for instance, so, so you say it can be that you need, you don’t actually need confessions very frequently in order person’s intentions are. So you get in an elevator, and the guy in the elevator presses the number three, from which you infer that he intends to go to floor three, right? Wrong? Like he might have pressed the wrong button, or he might be just intending to delay, how the delay the elevator, or a whole variety of things, right? But infer that, and it seems like basically correct. And then the So, so then the question is going to be kind of like, you know, Are there similar sorts of inferences that you can make about the about about AI or not. So can you include anything about what their what their representations are, what what organizing role those representations play in from their behavior. And so it’s not a matter of looking inside, right? You don’t look inside the head of the guy in the elevator to know that he intends button, right? You just watch him do it. And so similar you can you the question will be kind of like, can we make those kind of inferences about the about the AI or not, even though

 

Will Bachman  39:18

you So you gave us an uh, you talked earlier about you mentioned about addiction, and then you gave us the cool story. Tell us any other, either breakthroughs, or, you know, kind of advances that you’ve made in any one of these areas. I’ve been kind of like, you know,

 

Gideon Yaffe  39:37

have I made any advances? I mean,

 

Will Bachman  39:39

if you’re, if you know, if Yale come, you know, hired you, that you probably done something that they, you know, published something that they thought was pretty good, tell us, what was, you know, what are you proudest of?

 

Gideon Yaffe  39:49

What am I proud of? Stuff? Well, I mean, so, like, well, good question. So well. So one thing I did, I mean, this is actually since I’ve been at Yale, but one thing that I’m kind of proud of is a joint proud. Project that I did, actually, it was originally began with the MacArthur project. It was a joint project in neuroscience, where what we did was we gave people a task where what they had basically, what they had to decide was whether or not to carry drugs across the border when there was some probability that that was in the pursuit case that they were being asked to carry was drugs, some probability that it wasn’t drugs, some probability that they were going to be caught at the border, some probability that they wouldn’t be caught. And we gave them the idea was we wanted to induce them to be in one of one or the other of two mental states, either the state that the law defines as knowledge, which is state where you know for sure you’re carrying drugs, or the state which is described as recklessness, where you’re aware of a risk that you’re carrying drugs, but you don’t know that you’re carrying drugs. And so we had some subjects who knew that they were carrying drugs, some subjects who merely you know, were reckless about whether or not they had drugs. And then we were scanning their brains while we did this, and then we went back tried to distinguish the knowing and the reckless based solely on their brain activity. And we found we could do it pretty successfully, probably more accurately, actually, than people watching them,

 

Will Bachman  41:25

yeah, their

 

Gideon Yaffe  41:27

brain was a better indicator of what their mental state was, then people could infer from their behavior. So I was this

 

Will Bachman  41:36

one, yeah, so you could look at the mental scans and sort of be blinded, so you didn’t know if the subjects were the A or the B, and you could say, Okay, we think that these groups are the knowledge people, these groups of recklessness people, and you were actually better at predicting than someone just looking at them, yep.

 

Gideon Yaffe  41:53

Yeah. This was, this was kind of what was called Mind reading, by brain reading. Yeah. I also wrote another paper about that, about what conditions mind reading and brain reading, what, when? When should you be worried about our capacities to read minds from brain activity? I’m kind of proud of that one. I don’t know. I wrote a book about the grounds for leniency to child criminals. Kind of proud of that one too.

 

Will Bachman  42:24

What was the message there? Oh,

 

Gideon Yaffe  42:28

so that was kind of the opposite side on the neuroscience stuff, because I was really struck by how much people say, Oh, well, kids deserve a break because their own they haven’t fully they don’t have fully developed brains. And then I argue in this book that that’s not the reason why we should give them a break. We should give them a break instead because we deny them complete political representation, we don’t give them the vote. So it’s disenfranchisement, rather than neural immaturity, that warrants giving a break to kids. The one, one sort of intuitive way to see this is to think about differences in development between based on, you know, x y versus x X chromosomes. So, like X X chromosomes, we know they mature a lot faster than X, Y chromosome people. But what we don’t think, Oh, well, so then the age of maturity for X, Y, Z, for, you know, for girls, for women, should be much lower. They should actually, we should be throwing them in prison at 16, where we should wait until the boys are, you know, 22 like, that’s not the line we take. Why it must be that it’s actually not maturity that we’re reacting to and feeling that there’s something monstrous about giving a full helping of punishment to a child. So what is that? What is monstrous about that? I think it’s that it’s not fully self inflicted their lack of political participation. So that’s, that’s, you know, that’s, I don’t know. That’s another thing I’m proud of. I guess that’s also a book I wrote after I got to Yale. So I don’t know. Maybe my best work is here, I don’t know. Or maybe my best work is all before that. Or maybe my, maybe none of my work is that good. That’s also possible. But I wrote the, I wrote the book about kids in part, and I dedicated it to my daughter in part, because she was an adolescent at the time. She was, like, super precocious adolescent,

 

Will Bachman  44:28

I hope not a criminal, though I hope not. Okay, so on career

 

Gideon Yaffe  44:38

stuff, I haven’t talked about the personal stuff at all, which is, actually, I don’t know. I’ll

 

Will Bachman  44:44

get into that. Yeah, go ahead. Tell us about outside of being a philosopher, neuroscientist, you know law philosopher

 

Gideon Yaffe  44:51

well. So here’s something that happened. So in the 22 or the. Be the 2122 academic year. I was scheduled in the summer beforehand to teach a class on law and neuroscience I was excited to do and the pre and then I wasn’t able to teach it. The reason I wasn’t able to teach it was because the previous summer, June of 2021 I was in a massive, terrible car accident, Oh, no. And basically, my whole life, as I measure, my life is sort of before that accident and after that accident. Oh, wow. Okay, so what I ended up with as a result of that accident was a very, very severe brain injury, what’s called diffuse axonal injury, which is a very large percentage of people who have that injury die of it. And then many, many, many who have such injuries are never able to speak again, or never able to walk again, or so. I mean, in my case, I’m massively, massively lucky. Yeah. So I, yeah. So, I mean, my, I, I think a lot about the like, I think about this seriously every day, and yeah, and I’ve taken, I’ve made a lot of sort of steps in my life to try to express my gratitude about this, but

 

Will Bachman  46:24

tell us about that. Tell us about that more

 

Gideon Yaffe  46:27

too personal, too

 

Will Bachman  46:30

Okay, well, well, don’t do the too personal stuff. But what are the what about the stuff that you can share publicly that you’ve changed your book online?

 

Gideon Yaffe  46:42

Well, I mean, I’ve thought a lot about meaningfulness, about the way in which we make miserable events in our past meaningful by taking them as reasons to do good things, reasons that, I mean, it’s instances where they don’t actually give you any reason, really. So I think a lot, a lot a lot of things that we do are like this. So, you know, so if you, when you do something to honor somebody who died, what are you doing, actually? I mean, the fact that they died doesn’t actually give you a reason to do whatever it is that you do, right? I mean, whatever the good thing is that you do. So, like I was thinking about this in connection. So there was a, there was a New York Times editorial, come some time ago, now, maybe a year and a half ago, about a woman whose son died unexpectedly, and she became a vegetarian. The reason that she became a vegetarian was because he was very dedicated to climate activism. But, like, his dedication to climate activism was actually not really a reason to be a vegetarian. Like, I mean, there’s, there’s various kinds of reasons that we have to be a vegetarian. It’s like, you know, it’s it’s more healthy, and it’s better for the world, and a lot of things like that. But the fact that somebody you really care about died is not a reason to be a vegetarian. So why? But, but she makes it a reason to be a vegetarian, and she thereby gives a certain kind of meaning to his experience. So I don’t know. So the there’s various ways in which I’ve been trying to do good things, taking my my miserable injury as a reason, to try to make something out of that, to make it meaningful or important, or something like that.

 

Will Bachman  48:38

Wow, what? I’m glad that you were lucky on that, my goodness, tell us about Yeah, it’s I am. Yeah, I was in a car accident just a year ago, and the train was fine, but the car got totaled and but unlike you, I walk, yeah, I walked away with I walked away completely fine, a little shaken up, you know, but no injuries at all, other than just being a little shaken up, the car was totaled, and does put a new perspective on life so well, the car I was in, we’re just coming down the Farm Lane to the railroad tracks, and a little bit of brush in the way, so the car got a little bit too close to the railroad tracks, you know, to stop look, you know, looking for the train to come. And then the train was, like, 50 feet away. And, and, you know, a few more feet forward, and it would have been a goner, but, and, but the car, the train, just kind of clipped the front of the car and spun it 180 degrees. So the car was kind of totally destroyed, but airbags deployed, and they. Yes, they did very good job. So

 

Gideon Yaffe  50:04

I’m really you walked away. Me too.

 

Will Bachman  50:09

So Well, let’s start to this segment of the show. Tell us about any courses or professors that you had at Harvard that continue to resonate with you. Oh,

 

Gideon Yaffe  50:18

you know, there’s a lot of them, actually, what? So one, what was probably the best class that I took at Harvard, like, I don’t know whether you took this. One was Simon Shamas core class on the Baroque period. I missed that. That was amazing. So Shama Is this, like, famous dude now, but he was a course where all he did, pretty much in the course, was he went through really, really famous European paintings from the Baroque period and talked about them. Man, I mean that course, like, hey, basically every time I go to a museum, I think about that course, like, it just, you know, so, I mean, you see these Rubens paintings, or Las Meninas or whatever, I mean, these great paintings he would show in the class, and they that, so that I think about that class a lot, um, those were some, I don’t know. I think there was other. Were others. There’s, like, lots of others, actually. Some of them are, you know, some are like, I mean, there were philosophy classes that I took, like I took a seminar with Hilary Putnam that was really influential to me. Took the class on rational action with Robert Nozick, which was really influential. I took math 25 which was hell, and convinced.

 

Will Bachman  51:38

I convinced you not to be a math major. I guess

 

Gideon Yaffe  51:40

just it kicked my ass so hard that I could not be a math major. Yeah. But, like, there were, really, there was a lot of math professor, like, now, math professors who were in that class.

 

Will Bachman  51:56

So Gideon, you have a couple books out. You told us. You told us about one of them, where can listeners find you online, and perhaps a faculty profile or, or, I don’t know if you’re on social media, share any links here that where people can find out more about your work. I

 

Gideon Yaffe  52:13

don’t know. So I have a there’s, there’s web pages with my name was on them at the at Yale Law School where I work. I think there’s also one in the philosophy department, in the psychology department, I don’t know, but those are all kind of the same. But I like, never check those, never update them. But there’s like, CVS and stuff up there. I mean, I don’t know. And I mean, you can find an email address for me up there too, which I’m happy to have to get emails from people or whatever, but, all right, I don’t know. I don’t have any I’m not, like, I don’t have an Instagram My daughter has like, 15,000 followers or something on

 

Will Bachman  52:49

Instagram. Oh, 15,000 followers. Okay, and, and what? My goodness, what is she? What is she posting about? Is she Instagram influencer? She’s

 

Gideon Yaffe  53:01

not an Instagram influencer. She’s an actor. Who’s she’s taking a leave of absence from her freshman year at Harvard to shoot a TV show at the moment. So,

 

Will Bachman  53:11

oh, wow. Okay, so Gideon, that is amazing. It’s been so fun speaking with you, and thank you so much for joining today.

 

Gideon Yaffe  53:26

Yeah, thank you. Appreciate it. You.