David and Goliath: Underdogs, Misfits & the Art of Battling Giants

each year Microsoft Research helps hundreds of influential speakers from around the world including leading scientists renowned experts in technology book authors and leading academics and makes videos of these lectures freely available it's awesome to see so many people here I honestly kind of sort of didn't believe that many people still read books in the world given our next figure it's possible just as many people read the New Yorker but I'm pretty excited normally I like to actually read the book before I introduce the speaker unfortunately I was in Southeast Asia for the ten days the books kind of just out and I got it literally last night and I figured okay what the heck I'll start and I made it through the first two hundred pages or so so it's eminently readable is really great I think Malcolm Gladwell has been at The New Yorker since 1996 we simply know him from tipping point blink outliers what the dog saw I know a lot of us have read of stuff I'm really looking forward to him chittering some interesting conversation ideas and please give a warm like a soft welcome to Malcolm Gladwell thank you um so real pleasure to be here to see so many of you I can't remember how many times I've been here to Microsoft but it's it's now many each time I come here there's more more campus I I'm on my book tour you start your book tour in the west coast in LA and you work your way north and I don't know how many of you have had this experience but you had the very distinct sense when you start in LA and then go to San Francisco and then Portland and Seattle that the kind of average IQ rises as you but before for you before you get too big-headed I would point out that I'm a Canadian so my assumption hey so I am I thought in honor of the fact that I I'm in such a high IQ environment I would talk about one of the nerdier ideas in my book which is I have a couple of chapters a one in particular where I I deal with the idea of the inverted u-shaped curve as an explanatory model for various kinds of social phenomena and what interests me about inverted u-shaped curves is that they are there is something about them that defies intuitive comprehension I'm so sure you all know what an inverted u-shaped curve is we know what linear curves are right where there is a constant relationship between inputs and outputs we're very comfortable with diminishing marginal returns curves that are look like that where things start to flatten out but the inverted U is where things start positive go flat and then turn negative right and like I said what interests me is that we have such a hard time with that concept linear curves are the easiest of all of us for all of us to understand diminishing marginal returns is also a concept I think is not overly impossible for the general public to get but inverted u-shaped curves to feed us every time and one of the one of the things I try to do in the book is to illustrate why this is so profoundly problematic the question I don't answer in the book but which I hope to start a conversation about is why that's the case what is it about this kind of relationship that proves so difficult for human beings so let me give you two examples of u-shaped crows that I talk about in the book and you'll see the kind of consequences of our problem with them and in each case I the first is about California's experience with the three-strikes law which I'm sure all of you are familiar with this law has its origins 20 years ago when a man named Mike Reynolds who lives in Fresno California has a 18 year old daughter who is tragically murdered by some drug-crazed guy on a motorcycle in downtown fresno and he embarks on a crusade to strengthen California's criminal laws because he believes his daughter's murderer is only out on the streets because the criminal justice system in California is 2lakhs and he succeeds within a year his crusade results in a proposition being passed which dramatically strengthens the criminal laws in California so that the for any second offense a criminals mandatory sentence was doubled under three strikes and for a third offense and a third offense was defined incredibly broadly basically anything including stealing a slice of pizza for your third offense you would go to jail for a minimum of 25 years and a maximum of life and the result is that California in acts the toughest set of criminal statutes in the Western world and over the course of the life of three strikes remember it's famously it's it's it's a repealed in part last year but over the 20-year course of the of the law the prison population in California doubles and the per capita prison population in California rises to seven times that of Western Europe and Canada so you had this extraordinary increase in criminal penalties now I don't I'm gonna go and do a little bit I'm in a moment about criminology but that's not really what interests me here what interests me at the moment is the psychology of micro Reynolds's response right he sees a system where he believes a gas society where he believes crime is out of control and his assumption is that if he increases the criminal the penalties for criminal behavior by a certain amount crime will decrease by a roughly equivalent amount he has in his head in other words a notion of the relationship between penalties and criminality as linear that's what's in his head and when he makes this argument to the people of California in this campaign to get the three-strikes law enacted the the the law is enacted the proposition is supported by an overwhelming majority of Californians in 1994 why because that logic seems to them to be completely compelling it makes sense if you've got too much crime it must be the case that if you raise the penalties for crime the crime rate will will go down so what happens after three strikes is passed well what people observe in the short term in the mid-90s is that the crime rates in California start to come down by an enormous amount in the mid 1990s murder rape robbery auto theft I could go on all fall by about 40 to 50 percent in a span of a couple of years and the assumption people have is that that is a consequence of the three strikes law but very quickly it becomes clear that it's not when people look closer at the numbers they see that those crime declines started before three strikes was enacted and then when they look around the United States they realized that crime is falling by a similar amount in every state in the Union even in places where the criminal laws were untouched right so it becomes entirely unclear this is a result of three strikes and in the 20 years that three strikes was in effect there were an enormous amount of studies of this in the criminal literature dozens and dozens of studies and they could reach no consensus whatsoever on what the effect of the law was so there were huge numbers of studies which said they found zero effect there were a small number that said maybe at lower current crime by a little bit and then as a whole not a series of studies that said you know we think actually crime in California is higher after three strikes than it would have been otherwise in other words we had this model in our heads about what the effect of raising criminal penalties would be and the experience of three strikes defied the model right if you look at what actually happens you realize oh the model in our head wasn't the right wasn't we weren't carrying around the right assumptions to explain what would happen to crime let me give another example I don't know whether I imagined that I'm from New York but I imagine that in Seattle there is as much obsession among parents as there is in New York with the question of class size right people are when it comes to making to evaluating the quality of education one of the if not one of the metrics if not the leading metrics that many parents use is the average size of the classroom that their child will be a student in right and the assumption is that the smaller the class the better off your child is now that assumption has been around for an extraordinary long time if you look at the project of public education over the last hundred years in this country it has been I the first thing you see is that the increase in the amount of money spent on public education over the last hundred years has been mind-blowing we're talking about something that has exceeded the rate of inflation by several X and if you look at where that money has gone on what you discover is it overwhelmingly that money has gone towards hiring additional teachers that the project of public education of the last hundred years has been in steadily making classes on average smaller and smaller and smaller and smaller and the pace of that change has accelerated in recent years not decelerated so if you look at for example between 1996 and 2004 the average class in this country goes from I think seventeen point five to sixteen point two students now that's it is a complex methodology but let's see the other side the point is if that sounds like a trivial decline it's not that in that eight year span that difference in class size means you that we as a country hired an additional two hundred and forty thousand teachers right and if you do the math on that that's a lot of money there is almost no intervention you can do in public education that cost more than reducing class size it Dwarfs everything else in terms of cost which makes sense right you hire more teachers your if you think about a teacher and their pension and their health care and all that stuff you're talking eighty ninety thousand dollars a year and then you've got to build classes to accommodate them that's a lot of money at the end of the day and it's reflected in the taxes that you pay right and like I said why but there has been almost no argument over this project in fact there is no single intervention in the public education arena that is that has higher levels of public support than reducing class size and why are we so much in favor of class size because we have this in our heads right we think oh if my kid is happy in a class of twenty five they will be happier in a class of twenty so what does the evidence say well it turns out that the evidence on class size looks an awful lot like the evidence on three strikes it in no way does it support the notion that there's this kind of curve in fact the evidence on class size is one big unholy mess there are a mountain of study I when I say a mountain I mean quite literally a mountain I could fill this room with the studies that have been done on class size over the last hundred years and they have they reach absolutely no consensus the perhaps the best one of the best and most famous studies to give you a sense of how confusing this field is was done a couple years ago by an economist named Carolyn Hawkes P and she takes advantage of this really interesting natural experiment which is present in the state of Connecticut Connecticut is a state that has tons and tons and tons of very very small school districts so lots and lots of little schools and if you think about little schools and little districts they're going to have far more variable enrollments than large schools and large districts right and not only that a connecticut has this law which says that every class has to be capped at 24 students so what you see are these wild swings in the sizes of classes from year to year one year you've got 48 kids and you have two classes of 24 one year you've got 49 kids and you have classes of sixteen sixteen and seventeen with the same teachers in the same town in the same school with the same economy with the same parents with the same beautiful natural experiment what you do is you what she does is she accumulates 30 years of data and she looks at all 650 elementary schools in the state of Connecticut and she just says look in those years we're completely randomly there's 16 kids in a class or 17 or 18 how do the outcomes of those kids compare with the years in which there is 24 and what does she do after she accumulates is beautiful enormous data set what does she find in the difference between the performance of kids in the small classes with the performance the kids in a large class zero nothing goose egg and I saying zero I mean actually absolute zero it's not a it's not a she's not talking about something that's kind of it's there but it's not terribly statistically significant she finds eight statistically significant zero right now if you look at the take the three hundred best class sizes some of them are very good but ache the three hundred best ones and you do a meta-analysis you find basically the same thing fifteen percent of the studies on class size show a a significant although albeit very small effect from reducing classes the balance eighty five percent either show zero or a negative effect from reducing class sizes that is not a terribly overwhelming support for the notion that classes should be smaller in other words we have spent tens if not hundreds of billions of your dollars o also mine by the way I'm no longer Canadian I'm I mean I live here now so it's my dollars too we have all spent billions and billions of dollars over the last twenty five years on this accelerated course are reducing class sizes and there is absolutely no evidence to suggest that has made our kids better off we've gone around with this model in our heads and it's the wrong model right so what is the right model well I think the right model here is the right is the same as the right model in the case of crime we should be thinking about this the inverted U so let's go back and think about the inverted U for a moment an inverted u curve obviously has three components there is the left side of the curve where the relationship between inputs and outputs is positive there is the middle part of the curve where the relationship is neutral right and then there is the right side of the curve where the relationship is negative where the process that we observed to be advantageous in the beginning turns that turns into a process that is disadvantageous and what I would argue is that the U shaped curve fits many more social phenomena in our society than we care to admit in fact it is the most powerful way of intuitive way of making sense of the kinds of things that we are interested in as a society so if you fit the u-shaped curve to the question of class size it works perfectly you divide up the studies that are out there on class size according to the sizes of the class that they're dealing with and all of a sudden you see alright now this works so if we start with the very very large classes forty five kids and we reduce them to thirty are the kids better off you bet they're better off right famous study done in Israel because Israel had very very large class sizes and began to reduce them and what did they observe that kids are better off in a class of thirty than they are over the class of forty five of course they are forty five is crazy forty five all the teachers doing is try to keep order right there a traffic cop that's it they're not giving anyone any kind of intuitive individualized attention they can't even remember the names of the kids in their class for goodness sake right you go to thirty everything changes and so we do see by the way the effect sizes between 45 and 30 are not huge but they are significant right we're absolutely better off in that situation then what happens well you get into the 20s and it's pretty clear that you're in the flat middle part of the curve you can't see huge differences between 26 and 22 or 25 and 20 they're just not there and we're not carolyn Hawks be study I told you about in Connecticut that's a lot of what she's getting at most of the effect sizes she was looking at was comparing were in that low to mid 20 range and there there's just no no evidence that it makes one whit of difference whether you're in a class of 24 or 21 why is that well part of it has to do we think with the fact that one of the principal assumptions behind our belief in the efficacy of smaller classes is untested and they are to be wrong for it to work that a small class is better the teachers behavior would have to change as the class got smaller and it's not clear that it does so for example if I were to cut the workload of everyone in this room by 25% one of two things could happen you could you could either work the same number of hours but just spend an awful lot more time on the remaining 75% or you could go home early right now I'm not saying that this group would necessarily choose option 1 but our option 2 but our observation service budget but our observation of teachers is as a group they choose option 2 you cut besides their class they just work less hard right that's kind of human nature the most interesting though suggestion in the literature is that when you get below 20 you are very clearly on the right side of the curve that things are getting worse not better and the most I will give the caveat that there are not a lot of bizarrely not a lot of studies on very very small classes but those that have been done often show or have shown a negative effect and the reasoning is as follows it's principally focused not on the bright kids but on the struggling kids bright kids by the way no one should ever study bright kids in educational policy it's crazy if you're smart you're fine yeah you can if you've got a bright kid you can lock them in a closet and they'll be fine we shouldn't we should waste our time as a society worrying about the educational outcomes of kids with IQs of 130 please but the issue is with struggling kids and the argument is as follows that the single one of the single most important determinants of success in a classroom for someone who is struggling is the presence of a a real peer that is to say someone who is struggling at the same rate as you who was asking the same questions having the same problems and worrying about the same things and that if you have a troop here in the class you don't feel as isolated and as marginal you don't feel as dumb and the class slows down a lot right the more of you there are having the same problems the more the level of teaching is going to slow down and address your concerns so what happens when classes get too small is that it becomes harder and harder for the struggling students to find true peers and so the argument is we're really hurting we're so focused by the way on the relationship between the student and the teacher that we forget that for students a far more relationship more important relationship is between them and their peers and as you take away peers from struggling students you harm them that's the argument right I was so fascinated by this that I when I was doing my book I went and I I did an informal survey of several hundred teachers and the fastest the second fascinating thing with educational research is if you really want to know the answer to a thorny educational problem you just have to ask teachers no one ever asked teachers it's fascinating to me you can read I've read you know many many educational studies in my time it astonishes me how infrequently researchers simply don't just ask teachers for their perspective on a problem anyway so I I did this big survey of teachers and I said I asked the question is there such thing as such a thing as an optimal class size right is there a point at which a class could be too small not everyone but nearly everyone writes back of course absolutely as a point where it can be too small now there was some variation in what people thought the optimal size was but it was a surprisingly narrow window most people seem to think that when you got too far below 20 things got problematic and they gave a couple of additional arguments in addition to the poor students being stranded argument one argument I heard a lot was that behavioral problems in very small classes can be really well mning because as one teacher said it's like driving across country with two squabbling children in the backseat when a class gets too small one problematic dynamic between two kids dominates the whole class they can't hide from each other right you it can be so disruptive because they're just do on this other's throats the other more important argument was that you can't get discussion going particularly as kids get a little bit older when you start to get into young adolescents and kids start to students start to get very self-conscious and don't like the spotlight being shined on them and you want to have into the hallmark of a successful educational experience is a diverse amount of discussion in the class right all the great teachers that's what they do they get a discussion going because again kids learn as much if not more from their peers horizontally as they do from their teacher vertically well too few voices in the class means that the discussion suffers there's just not enough viewpoints not enough experiences and the kids are feeling very very scrutinized in a small class and you have these analytics teachers would tell me these stories about the time I had 14 kids it was a nightmare right I just could not get any kind of life going and the classic witness wasn't exciting right the um so what we see then is a clear case where the model that we should be using to evaluating this most crucial of interventions is wrong we've been doing this we should be thinking about this and that would change the way we approach all kinds of different questions in in education very very parenthetically I don't know how many of you send your kids to a private school but the chief offenders on this front are private schools who go around boasting about how small their classes are right without any regard to the notion that they this may actually be a way in which they're harming their students why do private schools do this because their principal motivation is not to provide the best learning environment for students it is to provide the most dazzling present to parents right they're interested in you not the kids which it's not all that surprising so crime if we go back to crime how well does the u-shaped curve fit our understanding of the relationship between criminal penalties and the crime rate and the answer is it fits it pretty well so if you think about our curve with our left side our middle and our right side is there a left side where we see clear benefits in terms of reduced crime for raising penalties absolutely if you have zero penalties and you raise them to X crime is going to fall hi there's a great example that one always uses is the famous police strike in Montreal in 1970 for 19 hours the police go on strike in Montreal and in those 19 hours Montreal basically turns into a medieval city there are gun battles in the streets there are the bank there's so many banks are robbed that the bank's just shut down they just like shut the doors in the whole city so it comes to a halt this is Canada right and I never know people had guns in Canada but they somehow discover them in that 99 period just go around shooting people right absolutely there's a left side to this curve is there a middle where the introduction of harsher penalties has no effect on crime absolutely a great example this is with three strikes and this is a point by the way this is one of those arguments I have tried to make this argument - let's just say less sophisticated audiences people refuse to believe this argument so I'm just going to say they wrister feel I don't know why this is really hard for people to grasp but okay here's how it goes you won't have any problem with this under the three strikes law the average age in which someone was convicted of their three third strike was 37 years of age before three strikes that person would have served an average of five years for their felony conviction so they would have gone to jail from 30 stage of 437 to the age of 42 and gotten out under three strikes they went to jail essentially for life so they went to jail at 37 till the end of their days so what three strikes had the effect of doing was locking up that person between the age of 47 and the end of their life right so the question is how many crimes do criminals typically commit after the age of 42 and the answer is none they don't you see you look at current at age crime curves they are the most fascinating thing you've ever seen the age crime curve looks like this no one commits may I say no one statistically crimes stop being committed past the age of basically by the end of your 20s your criminal career is largely over with some exceptions so there are a few people in this room who I still think are at risk for doing something nasty most of you you're out of the woods right so what does three-strikes do you think you're doing this enormous the important thing by locking these people out from 42 to the end of their days but in fact you have intervened in the career of the lifespan of a criminal at precisely the moment that he has stopped committing crimes right that's the middle part of the curve now here is the really interesting part though is there a right side to the curve is there a point at which locking up more and more people raises crime rates and this is where the argument gets really interesting so there's a whole school of criminologist recent years who have said who started to make this argument and the argument goes something like this when you lock up somebody there is a direct and an indirect effect on crime the direct effect on crime is easy to understand that individual by virtue of being behind bars can no longer commit a crime right the indirect effect is on the community that individual is a part of so they are criminals even juvenile delinquents are statistically more than likely more likely than not to be parents and if you do this the actual number is kind of fessing all I have it here one-fourth oh I'm sorry what is only one 1/4 of juveniles convicted of crimes have children and the effect of on a child of having a father put behind bars is actually quite devastating it's we have the numbers of people done this it raises your chance of being a juvenile delinquent by something like 300% it raises your chance of suffering from clinical depression by something like 200% there are these collateral consequences in other words to locking up someone most criminals are breadwinners even though they may not be getting their money legitimately that's still whole that's still money that's going to feed and clothe their family when you remove someone from a society from a community you remove a breadwinner right when you remove the biggest thing that keeps the biggest predictor of of a productive life for someone in a poor community is whether their parents are married there is nothing more devastating to a marriage than being locked up it destroys marriages and then most importantly the person that you lock up invariably returns to that community after they've served their term and when they return they are in far worse shape than when they left right they've just been behind bars for X number of years they're unemployable their peer group has now been replaced with a entirely criminal peer group and they are a burden on their family so those are all of the collateral consequences of imprisonment normally the direct effects are in most situations circumstances the direct effect of locking someone up is greater than the indirect collateral damage caused by them being locked up except if you lock up too many people then the collateral effect starts to outweigh the direct effect now what is the threshold well we think based on some very preliminary research that the that the the tipping point if you like is it's right around 2% if I lock up 2% of a community collateral damage starts to be greater than direct damage now 2% does not sound like a high number it's an enormous number nonetheless are there are communities in the United States where more than 2% of the adult population is behind bars absolutely in fact there are communities that are well above that 2% mark and studies of those kinds of communities have suggested that the more people from those communities are locked up the worse the crime problem in those universities in those are communities versus sorry I've been on the road for a long time communities becomes so the question is when was the last time you heard a politician if this is true stand up and say in answer to pleas from their constituents to do something with a crime problem to stand up and say okay I will start letting people out of jail doesn't happen even though that is an entirely logical response given these curves in fact if you look the most fastening example of this is this is exactly what happened in New York City if you know this but New York City and a crime in New York City declined in 1990s somewhat in lockstep actually a little bit ahead of criminal decline of crime declines in all other major urban areas in the United States and then in the turn of this century the crime rates in most American cities plateaued and in some cases crept up again like Chicago in New York they continued to fall in fact New York is now this such a completely bizarre inexplicable outlier in the world of crime that nobody knows what to do it's blown apart every assumption New York it just crime just keeps getting lower and lower and lower and lower on certain indices if you just if you take out it's going to sound down but if you remove murder from the equation New York is basically safer than Tokyo right now which is so weird and bizarre doesn't make any sense but the fascinating thing about this crime decline in New York the second crazy crime decline is that it has been occurring as the criminal population of New York City has been in sharp decline in other words it's happened while we've been emptying prisons right exactly what the u-shaped curve would suggest that New York was in many communities past the two percent cutoff point and what they're do what they've been doing over the last ten years is bringing that imprisonment rate back below that crucial threshold and in so doing restoring the community even as they are letting more criminals out behind bars but no one makes an argument it's completely implausible right you can't imagine you imagine a politician standing up and saying that and why not because we have such difficulty with this notion of a u-shaped curve because it seems to be beyond the ability of anyone in this country to stand up and say it's not this it's this now why can't we grasp this why what is so hard about this well that's why I don't have a good answer now we actually we only have time for questions I'd love to hear people's suggestions I'm going to throw out a couple of very very speculative not terribly serious possible answers the first one which I don't really buy some people would say well it's just this is hard right that's easy to explain this is hard to explain I don't think this is hard to explain it all actually it's what your mom said everything in moderation right what's hard about that nothing's hard about that so I don't really buy that it's hard argument you could make a kind of this is again this is a very whimsical suggestion could you make an evolutionary psych argument that says that we because we evolved in times of profound material scarcity there is nothing in our hardware that accepts the notion of too much right it would have never been an advantage during much of human evolution for us to ever conceive of something being problematic in excess because there was no excess it's only in the last hundred years that we've invented this notion of surplus that we've had enough money to be able to make classes too small or had enough resources to be able to lock up too many people maybe that's just a new phenomenon for us that we haven't grappled with that's what I do but what is interesting about all of this is that the key component that unites both the crime and the educational examples is that they are both cases where in order to see to grasp the nature of the u-shaped curve you have to look beyond the individual right if all you do is think about a student standing by him or herself a student in isolation you can't see the problem with classes being too small you have to understand that the student is a part of an ecosystem in a classroom and is dependent on their peers and so you take away too many of the peers the dynamics become problematic same thing with criminality in order to understand the problem with locking up the criminal who's in front of you you have to understand that the criminal isn't doesn't act in isolation that they have families and they come from neighborhoods and in some bizarre and paradoxical way their families and neighborhoods depend on them and if you target too many people like that you're harming everyone who's still at home in both cases in other words it requires us to have not just a different kind of mental model of mathematical model it requires us to have a different kind of social model it requires this to understand that human beings don't stand alone in the world they're part of large groups thank you I think I'd love to take questions let me just put my glasses on so I can see you um it's 1:00 in the front oh dude do I find it oh the question is have I applied this theory to DC politics no I I always give this answer and I as a Canadian I have no understanding of American politics and the longer I spend in this country the less understanding I have of what goes on um but um but certainly the I mean you know if if the lesson of all everything I just told you is what your mother told you a long time ago that everything in moderation that certainly is something that could be is worth saying in Washington right now I'm sure sure if the stock market's going up people assume it's going to keep going up if it's going down it's going to keep going they can't even recognize this simple yeah listen so yeah this is very quick eat the questions I'm not sure all have you heard it it's actually it raises a really interesting point so the questioner said I made an analogy to the stock market that people even have difficulty in a stock market and understanding that just because it's going up now doesn't mean it'll always go up right take their there ever there's some part of us that wants to extrapolate indefinitely from that kind of upward momentum what's fascinating about that is that that is absolutely true but in recent years there's been a lot of work to suggest that that is true specifically of Western cultures there are clear differences between in the kind of bedrock mental assumptions or cultural assumptions held by people in the East and people in the West and so when you so for the longest time psychologists make that observation they were doing all their studies in Western Europe and North America and a minute they started to do those studies in China or Japan or Korea they discovered that people didn't lock in to those kinds of of relatively simplistic relationships so it is that it may be something weirdly specific to the kind of analytical package that the people carry around in the West certainly by the way on the city I just just occurs to me if you go to countries like Korea the for example the classes are really large compared to American classes no one seems to assume that the kids would benefit from a dramatically smaller classroom so it's and by the way their students dramatically outperform ours in addition so sure hello so in one of your interviews I remember you saying that having too much money can be a little problematic so I'm just wondering if personal wealth follows the same yeah bankwest yeah does personal wealth follow a inverted u-shaped curve and the answer is I have a chapter on this yeah so but then the wit and the most specific way to say it is that parenting clearly is u-shaped it is very hard to be a parent if you have very little money and as I give you more money your job as a parent gets easier absolutely doesn't is there a middle part of the curve where that effect starts to level off yes we can argue about what the number is but you know most of the happiest research says that around 75 or 80 thousand dollars a year you stop seeing monotonic increases in happiness associate with wealth then is there a right side and so I hung out you you guys will be very trigged by these observations I hung out when I was writing my book with all of these children of very very wealthy people and in order to see whether the kind of thing you know we all of us have a cliche in our head about the spoiled rich kid right it doesn't can't make and if you talk to people who do counseling of wealthy kids and there's a by the way an entire ecosystem of psychologists who do nothing but do this and they will say well that's absolutely the case you would not believe the level of dysfunction among wealthy kids so I began to interview lots and lots of wealthy kids and I have to say whoa and you realize that what's going on is that exactly what I said which is that the task of being an effective parent begins to get harder again when resources past a certain threshold amount for the this it's not impossible to be a good parent if you're wealthy it's just harder I'm harder in the way that it's hard to be a good parent if you have too little money and the class example this guy James Grubman gave me this beautiful way of phrasing it he said you get into these can't won't problems that if you are a middle-income parent and your child says I'd like a pony you say we can't and the kid very quickly understands there's a whole range of requests they can make that will never ever be granted they're just impossible right you're not getting a pony I mean the you haven't said any more used to say would it's not happening but if you have a hundred million dollars and your kid says I want a pony you can't say we can't have a pony you've got a private jet of course I can have a pony right so now if you don't want your kid to have a pony you have to be capable of making not a can't argument but a won't argument and a won't argument is a lot harder a won't argument requires that you say explain what your values are and explain what your attitude to as a parent is towards your child and explain why this paradox which is why providing for your every needs now at the age of ten is not in your own best long-term interest right that's a much harder conversation to make particularly if you're guilty about working 80 hours a week right so it gets harder again and that's I think what's going on with this consistent phenomenon of of second and third generations of wealth falling apart which is their parents aren't prepared for the additional responsibilities of being rich so absolutely there's a u-shaped curve Michigan's yeah so the question is does this apply is it a natural extension of diminishing returns in other words that do you always if you keep adding resources during the period of diminishing returns do things eventually always go negative really interesting thought I don't know I mean this is kind of it's the most fun kind of a question to ask because it's ultimately unanswerable but um I kind of think so so a good thought experiment is give me a give me an example of a phenomenon involving human beings that is not u-shaped in other words can you come up with a phenomena involving human beings that is purely linear or purely just diminishing marginal returns and the answer is I I've given this question a lot slots people no one's ever given me a good one where does it you know even the even just simply ending at diminishing marginal returns can you think of that if I if I were to take the R&D budget of any company United States and increase it by a factor of 10x do we does anyone really think that wouldn't be net negative for the company right I mean think about you know I was talking to somebody you clearly would get diminishing marginal turns but I think that much no one's going to argue with right 10x I mean let's go crazy 10x you clearly get diminishing about returns but if I gave you 10 times more money I don't think you would be performing at this level you're performing now you need constrain the sum when I was talking to someone who is worked for NPR and he was doing a book on the fact that Ray Kroc the guy who was the McDonald's guy his widow took all the money left to her by Ray and basically gave almost all of it to NPR an overnight NPR got billions of dollars and she said you've no idea what a completely toxic effect this had on the culture this is not a culture that was equipped to deal with cash raining down from the skies right uh-huh it wasn't positive now maybe it ended up eventually but you know it's really hard to I thought out there just it's really hard to think of cases where where things don't get eventually start to go negative involving in things above humans sure I was wondering how you thought the u-shaped curve might apply to mastery if it takes 10,000 hours to master something is there some point past that where you're getting a negative effect oh yeah this is an interesting no no this is a very so there's a huge huge there's a significant literature on this very question so if you and it fits it does I think fit into the u-shaped model so we don't think we clearly think so what we're describing with the 10,000 hour rule is the left side of the curve so 10,000 hours is the point at which for many complex cognitively complex activities the point at which the curve starts to flatten out a bit and we're getting another way of saying that is past 10,000 hours you get diminished returns from additional practice but the really interesting question is is there a right side and the answer is yeah I think there's a wide right side so for example if there isn't sufficient variety and challenge after the point at which mastery addition ischl mastery is reached then what do you get you get boredom it becomes ro you know why do we worry about the performance of you know surgeons as they get older well because we realize that the the work isn't necessarily as engrossing as it was when they were younger and they were still they still felt they were progressing in some way so we you know the oldest surgeon works well when they have managed to contrive for themselves a situation where they continue to have to meet certain kinds of achievable challenges but it's a real problem I mean all of us have run into professionals at the end of their career and you can say I think quite quite reasonably that they have spent too much time in their profession right they've gone there on the right side of the curve about this and I'll end on this note this brings up an additional really fascinating distinction that I've always thought was incredibly important which was when you observe so if you look at any number of complex professions and you look at the productivity curves of people in them you see something that is broadly u-shaped so you see improvement over some period at beginning that you see a flattening out then you see decline at the end the question is is that curve a function of age or experience in other words is it is your performance at the end of your life going down because you're old or is it simply because you've been doing that thing for too long and the answer seems to be the latter in other words if you take people who are in mid-career and you transfer them into a new domain they will have an experience roughly similar to that same curve so it's a very hopeful thing it says that a lot of times we think what we're seeing is age-related decline we're not all we're seeing is someone who is who is just on the backside of an experience curve and then what we need to do is give them new challenges in a new environment and we can see the same improvement in performance anyway thank you all very much hope you enjoy

Channel: Microsoft Research

Views: 70,552

Rating: 4.7919073 out of 5

Keywords: microsoft research

Id: 7RGB78oREhM

Length: 51min 36sec (3096 seconds)

Published: Fri Sep 09 2016