Do you show your frustration by hammering the lift call button? Peter, Nick and Fraser discuss the urge.

For more information on Aleph Insights visit our website https://alephinsights.com or to get in touch about our podcast email [email protected]

Hello and welcome to the Cognitive Engineering Podcast produced by Tell Me Studios for Aleph Insights. In this series of podcasts we take a look at interesting topics and discuss what we think they tell us about analysis and decision making. I'm Fraser McGruer and I'm here with Nick Hare and Peter Coghill of Aleph Insights and this week we're discussing lift or elevator buttons and impatience. So just to start us off I've got a friend who gets really annoyed with me when if we're waiting for a lift because I press the button nothing happens with it doesn't doesn't come and I press it again and again and again and just keep pressing it and that really annoys him. Peter is he right

to get annoyed with me? I think I would feel a similar sort of frustration because I think it comes down to knowing more about how the lift works perhaps having a more in-depth model for what the control system is and I know that the button is a binary button. Well no no no no no

I'm going to stop you there you can't pull the engineering ace card on me. I know how a lift works as well to the extent that if I press that button that's kind of all I need to do yeah and

yet I still keep pressing it. So why is that? What makes you think that pressing the button harder or more often it's going to it was going to make you make the lift appear quicker? Oh well that's answer that for me. Well I think it's because you're not controlling your rational self. Your reptilian early brain is saying well this is synonymous with the physical world so if I want to move something faster I have to push it harder or push it more often and your higher functions aren't controlling that action and saying I don't need to push I'll push it once and it's going to come as quickly as if I so hold on let me get a question

here for you so far I mean I agree with what you said and I recognize the irrationality of what I do let's say I mean what's your expectation of when you press a lift button and in a normal situation what's your expectation of how long that it will take that lift arrives let's say within one minute for example that's even pushing it a bit but let's say you've pushed it and the button's lit up and you know that lift's coming and two minutes pass then three minutes and the button is still lit what would you do Peter? Well I might take the stairs at that point but hold on hold on hold on hold on hold on so would you take the stairs before pushing the button

again? Yes because I know this it won't make any difference pushing the button really that said though I think the control systems often aren't that simple and a very easy way to work out how many people you're inconveniencing by not getting the lift to them and therefore a way of prioritizing the where the lift moves to next is to account for the number of buttons that are pushed so if you could you could have a simple counter that just counts the number of times that being the buttons being pushed on each floor and if there's one button push that's less that has less priority than someone has pushed it lots of times yeah but then then you're just going to

get into a war of button mad button pushers on every floor that'd be very very easy to game

you'd have to have a system like that where the operate the people using it would not know that

those were the rules yeah yeah well and then and then you wonder why why have it I mean I suppose the the um what you're trying to do is assess the number of independent button pushes if everyone had their own individual button push you know but I think in real life here you've got a situation where you'd have to entertain some alternative hypotheses like the maybe the lift's broken which seems unlikely if the light was off and you pressed it and then the light came on it feels like that's probably not broken but it could be that the light is is the button is malfunctioning in some way I've encountered real life buttons where you know the light works but the but the order hasn't been processed or whatever it may be that pushing the button harder will will solve

that so out of interest Nick what would you do in my situation would you if the lift wasn't coming

would you start pushing the button yeah and I think it mainly mainly not because I believed it would be effective but because it would relieve me of a kind of tension I think I think the the issue is to do with and I think with Peter touched on it with the we're talking about our quite base instincts here a lot of the time we treat things in the world as though they're as though they're agents I think there's a lot of kind of um animism in our fundamental behaviors you know that actually if the lift was an was an animal and you poked it really hard a lot of times it would pay attention to you um and that's that I I feel like that's often what lies behind uh our frustration with with control mechanisms and uh you know interfaces is that we think of it as a kind of creature and we feel slighted by the fact that this creature is not is not paying attention to us and and certainly with computer rage that's that's a big draw I mean computers try to be your best friend don't know they try to be your mate they try to act like they're a person an agent doing things on your behalf they have things like clippy the paper clip who pops up exactly and then and I feel like well now I've got the right to get annoyed with you if you go into one of your ridiculous funks where you ignore the fact I've just double clicked on this icon especially if they humanize themselves right right and that's so that's interesting I mean you humanize computers because people relate to them but then you you the flip side of that is that people are going to get cross with a computer in a way they don't get cross with

yeah but also I remember one of the first things I ever did with Siri was to swear at it just for fun and I'm sure millions of people can say the same thing right and and sure enough Siri has got a response built in which he says something like well you didn't need to say that or something like that yeah but um which has obviously been pre-programmed so the the programmers understand people um yeah I mean where do we want to go with this well I think there are the other control

definitely imagine myself doing both of these things I wonder yeah but is it possible that actually there is some wisdom there because I know in theory obviously in theory you should select your preferred temperature once ever when you first install your thermostat and then leave it alone for the rest of your life in theory but you know I definitely feel like actually sometimes when it's very close to the uh temperature you want it to be the the noise in the system um is such that it may well just be you know it's slightly off on its detection of temperature or it's detecting the wrong bit of the house or whatever and therefore you know actually if I know I just need the heating to come on then you know I can turn the thermostat up for for an hour and uh and and then put it back down later yeah well there is definitely

engineered hysteresis in it so that it will only switch on below a lower threshold and switch off after a higher threshold so to stop a flip-flop a situation going where it's right on them right on the temperature and the boiler is just being switched on and on off on off on off um so there's there's engineers hysteresis but so yes you are kind of right but it's only half a degree it's a tiny fraction um so what you're actually probably doing is wanting a slightly warmer room in which case you are totally right to turn the thermostat up it could I mean there's also the prospect that

we our own internal perception of temperature is driven by lots of things that aren't actually that closely related at any given time to the environmental temperature so actually you know I'll be fiddling with my thermostat because 21 degrees might be fine one day but the next day because I've you know got different jumper on or I've just been outside or there's a slight breeze

you know a different different number is better I mean that's why I would say I mean certainly not for the second example you gave but the first one I mean when it's colder I actually want to be hotter than my usual 21 degrees let's say that I would like the ambient temperature I mean

these these these figures don't in my house I spend a lot of time in a back room that's got a crappy radiator so I actually have to take more manual control over my thermostat well Peter

actually lives in a super cooled vat uh and and attached to electrodes yeah I live in a server

room yeah it's uh very very chilly in there but actually yeah there's a third thing there's

another another option here which is I don't want a bloody thermostat anyway because we've got a thermostat in my house and I do not like it I don't I don't like having control over this

this other part this is a really this is a big thing yeah this is a big thing I mean so as uh you you might really hate having a smart thermostat because it's trying to it's trying to be take remove even more control from you more direct control um and be smarter about how it's how it's um changing the temperature on your part so working out when you were in the house when you're not adjusting the temperature to save your energy but then if you so the two ways to go I think you could either make it even smarter and collect more data about you so it would work out when you're likely to be feeling chilly and pre and guess ahead of when you would go and control your thermostat or go a very simplistic route and just give you a volume control like control over your temperature and give you full full manual control over what temperature it is this

this reminds me of of the uncanny valley it's almost like a kind of control or version of uncanny valley uh uncanny valley is the um it's this it's this sort of uh phenomenon where if you make something slightly anthropomorphic so it's something like a you know a cartoon character where it doesn't look very much like a human at all but we can find it cute you can find a you know a cartoon character big eyes completely if that cat thing existed in real life would be terrifying but it but it's fine because it's a just an anthropomorphized representation of a pretend thing and then you can push things to be closer and closer more resembling of humans and actually when you get when you get to a point where you're quite close to but not quite at looking exactly like a human our brains start to interpret it as a human where there's something wrong rather than a rather than a non-human that looks a bit like a human it starts to look like a human with something wrong with it and we find it really repulsive and so uncanny valley sort of affects you know if you have cgi that's very very close but not absolutely right or um or you have a robot which is quite like a human this is almost this made me think of that because i think actually you know when when you're getting smarter and smarter with your control mechanisms um in a way like a really dumb system where you have an on and off uh is is fine like you know it's a bit more work but we're kind of okay with that um but when it starts to be almost trying to do things for us in a very very smart way if it isn't getting that exactly right it's phenomenally irritating you know if it is trying to help you and failing it's much more annoying than if it's just you know actually you're the one who has to turn it on and off yes where do we go from there um yeah well i mean i suppose the question is do we try and design interfaces to take account of humans uh of our kind of intuitive feeling about the world you know reacts when you push it harder i was wondering about um i was thinking everything about harpsichords uh because you know harpsichords obviously doesn't matter how hard you hit the keys um they they will always pluck at the same same uh volume um i still wonder if people when they were playing the harpsichord weren't hitting the keys harder when they wanted it to be louder and in fact one of the things you know that's uh that that why they invented the piano was because that was such a sort of instinctive urge and something that was completely you know was wrong with the harpsichord in a way like wrong in terms of it didn't feel right because other instruments obviously you know there are there are there are you know guitar which is you pluck it harder it's louder um you know there are stringed instruments which you hit with hammers which you know i get louder the more you hit them yeah i find harpsichords really hard to play because i find them you know i can't i just

really feel like i need some more feedback okay so that being the case then i think i agree with you and so we what we need to do with lifts is what we've done with with the harpsichord right so back to you peter so we need this system right where um it it responds accordingly to either the number of times we press it or the the weight with which we press it yeah that's

the future yeah it's feasible i mean but yes you the with with a simple button press it would easily do games you just sit there and hang on the button and uh everyone that would ultimately do that and you'd be in the same place albeit you'd all be suffering because everyone's trying to hang on the button but you could you could say you could have a fingerprint reader in the button so each unique button press you've only got everyone's i got up to 10 so um you've got you've got 10 button presses then you're in the same place but somebody else comes along then there's 20 button presses on your floor so the lift could then prioritize your floor of

a floor with any one person yeah i guess where the analogy falls down is that the harpsichord there's just one person and um yeah there's a game theory element to the lift which which confuses matters a bit yeah um but what we need is we need um lifts that can tell the future right

and that can guess well i think the hitchhikers go to the galaxy had a lift that could well i

think this is a bit like that you know the issue about train timetables and stuff being you know when you when processing gets sophisticated enough when when we are able to tailor responses in as smart a way as possible then um machines will start behaving in ways that are kind of obscure and hard to read and and i think that you know with the with train timetables at the moment you know that they've designed so that they're legible so that we we can find out which train we need to be on uh you can imagine very easily in the future having dynamic train train timetables where you just turn up you use your app to tell it where you want to go and it just comes you know there might be just a train might turn up stop at a couple of random stops go back by one stop you know something quite uh chaotic looking would happen that that would be optimal um and i and i wonder actually with sophisticated control whether actually this would start to violate our need to understand how a control system works that actually whether this would i mean it would deliver more efficient outcomes but we might not like it we might not like it if the machine's doing things that we don't understand on our behalf i wonder if that's a

kind of limitation so if we had a prescient lift we would actually rebel against that well i mean

it would do things like it might do things that we would think of as weird or illogical you know why we've just been up to the 10th floor why are we going down to the 7th and up to the 9th and you know we feel like the lift ought to go all the way in one direction and then come all the way back in the other direction that's how a human would design it but actually the optimal uh might be some weird sort of fractal looking um you know series of small up and down journeys

i think that maybe that kind of distrust is rational because if you could get you could be that the lift is trying to minimize journey time for all passengers you might be that really unlucky guy that's stuck in a real small spiral of the fractal well you end up spending all day on the lift never going to his floor yeah everybody else you just happen to be the extreme in in the

distribution yeah although although you wouldn't have to that would be a slightly balmy objective for it to minimize you know i think i think you'd probably do something like mini max or you know you'd want to sort of minimize the maximum waiting time or something um but you know it's it is tricky because then then you actually look at the advantage of a simple system is it's not trying to optimize anything yeah it's just doing what it does um and when when systems start optimizing then you might be on the wrong end of it it also violates some kind of egalitarianism that we have you know um but uh but yeah i just i i think that's that's something we really are going to have to face a lot more of is is um yeah the and of course there's the other issue that that if you don't understand what machine's doing you can't necessarily tell if it's malfunctioning which is also something i think that worries us uh you know hence the arguments that i have with my girlfriend about um the sat nav you know that she she believes that the sat nav sends us sometimes in an inefficient direction and i think sometimes she's right i i whatever algorithm it's using is definitely sometimes uh actually just is not as good and um you know well it might be it's doing something really clever but but it might just be getting it wrong

yes yeah and it sometimes is i mean that's quite a common thing actually my wife is always arguing

with our sat nav as well um i think when when deep blue was first playing gary kasparov uh in the in the late 90s i think there was an occasion where kasparov you know noted one of its moves which was really surprising and um and uh sort of complimented it and said that's a really deep

sort of move and it turned out to have been a bug yeah okay right i think there's another there is another sort of solution if you can you can do away with the scarcity um then you then there's a natural solution there so the tubes aren't timetabled because there's always going to be another one along in five minutes you don't need to plan it um so you don't there's not using a brute force approach so there's no computation there it's just you just you just know there's going to be a tube along in a minute you don't need to timetable it it's only in trains where there's a bigger gap and you have to do more planning to get connections and things there's a problem so if you replaced all the trains with big conveyor belts that just accelerated you up to travel speed and you could step off at any any point then they would

that would solve it i want to this is going slightly off topic but um i want to regale you with some um stories of lifts in my life okay please do please do okay well and actually they're all annoying experiences is worst country in the world for lifts in my experience turkey okay and it's not the lifts that are the problem is that i think it's to do with this is when i was in istanbul and working and i was visiting banks um all of the turkish backs seem to be these huge high-rise buildings but they hadn't matched um the the height and and the number of people working there with the capacity for lifts so you'd have this huge building with many many floors with thousands of workers trying to get in you've got sort of one lift well i mean that's an exaggeration but a couple of lifts so you used to have lift lift jams at sort of nine o'clock in

the morning and at five o'clock in the evening so yeah that's quite well that's quite well documented in in skyscraper it's quite a difficult planning problem because you're um you have a rush hour um at the other end of the day uh and so you've got to you've got to design your capacity to meet that the maximum of the rush hour but the most then most of the day lift is sitting idle um and that's a in a high-rise building a lift is quite a big piece of space to use up that could be used for more rent rented office space or rented

accommodation did you ever play sim tower i did yeah sim tower was basically a lift simulator but with it with a kind of tower building element built on i can see that really being a pizza

type game yeah so so there are lots of solutions to this you can either have extremely lots of lots of lifts that take a brute force approach or in parallel or you could have you can have um the building sections so you you take a a fast lift up to a hub and then a smaller lift down or up from there to to your floor so they're sort of separated in in space kind of tree model a tree model yeah um so that but it's a big planning problem it's very difficult to predict how the

building is going to be used throughout its lifetime yeah um nice uh a second lift um story it was in brazil i get so fed up in brazil where i visit fairly often um because a lot people generally live in high-rise buildings um and and one of the things that annoys me there's two things annoy me one it tends to be incredibly hot because one of the areas that tends not to be have air conditioning brazil would be a lobby um and also in the lift itself so i'd spend a lot of my time well that's the first thing but the second thing is is um it's something that people are sometimes doing brazil is they hold a lift a lot they tend to hold it so if someone you know and they've got big families and they're waiting for everyone to come through and they just hold the lift and actually you can hear people on other floors start shouting um and so i just spent i just feel in brazil i spent a lot of my time waiting around in lifts for a lift or in a lift and and just being and so double with that the frustration is just how hot it is and it just

drives me mad i mean that's that's an interesting thing if you if we did design this super intelligent lift that uh that was able to predict when it was going to be needed and anything else then the again the humans are going to come along and mess it up because they people have different concepts of what manners are in terms of holding lifts i think peter's

ideal lift system is one with no humans in it at all at any time well a big conveyor belt you just

get on or or low-rise the answer or low-rise buildings with yeah low-rise or use the stairs

slides yeah what's some slides yeah yeah brilliant no i think but i think i remember seeing something recently someone said the lift was the it's the uh probably one of the most influential inventions you know of the last 500 years because it enables us to live in high-rise buildings which are the foundation of the modern city you know simply not possible without the lift and the safety mechanisms that you know the really clever safety mechanisms that are built into it and stuff

so uh before i wind up anything anyone wants to say what where did we start on this oh yeah don't push lift buttons yeah um it sounds like the problem with all the the common theme here is the problem

is not the technology the problem is it's how we yeah it's how we interpret it how we how we engage

with it how we give it agency when it shouldn't that doesn't necessarily deserve it so yeah so to round it i mean perhaps i was going to say a more aleph kind of thing but actually a more peter aleph thing would be we would just be better off without humans essentially right yes yeah okay

there we go on that well not necessarily but better off without humans but if we see we can't get rid of humans then better off with extremely intelligent systems that collect a lot of data and

processing right so i think the problem is we're in that bridge at the moment aren't we so we're kind of living through the evolution of of genuinely smart systems and it causes problems and yes the solution is to take all the humans melt them down into a kind of soup and then and then use that to build lift yeah use it use it to fuel a giant lift manufacturing

okay right let's wrap up there gentlemen thank you as always um thank you to listening to the cognitive engineering podcast i'm fraser mcgruer we've been here with peter coghill and nick hare of aleph insights thank you and until next time bye