Future of Coding - Man-Computer Symbiosis by J.C.R. Licklider

Episode Date: April 12, 2022

Jimmy Miller joins the show as co-host. Together, we embark on a new series of episodes covering the most influential and interesting papers in the history of our field. Some of these papers led directly to where we are today, and their influence cannot be overstated. Others were overlooked or unloved in their day, and we revive them out of curiosity and wonder. A few even hint at an inspiring future we haven't yet achieved, placing them squarely in line with our community's goals. We give these papers all the respect and deep reflection they deserve and, perhaps, the occasional kick in the shins. Today's paper is titled Man-Computer Symbiosis, authored by J.C.R. "Licky" Licklider in 1960. The title sure is outdated — but how have the ideas aged in the eternity since it was published? Listen on to hear your two hosts figure that out, and delight at just how wildly right and wrong some of its predictions turned out to be. Thank you to the following sponsors, all of whom are doing important work in our field, and all of whom want to hire you to do even more of it: Theatre.js — Enabling designers to code, programmers to design — empowering everyone to create. Glide — Anyone can make their own apps, with a GUI builder backed by a spreadsheet. Replit — An online REPL-driven dev environment with all the batteries you could ask for. (I keep hearing about more and more people in our sphere landing jobs at these co's — high-fives all around!) Show notes for this episode can be found at futureofcoding.org/episodes/55Support us on Patreon: https://www.patreon.com/futureofcodingSee omnystudio.com/listener for privacy information.

Welcome to the future of coding. Good to be back. I've got a brand new thing that we're going to start doing on the podcast. And I say we because I am pleased to introduce you to a new co-host. Jimmy Miller is going to be joining me on a series of episodes of indefinite length. This might just be what the show is now, in addition to interviews when they come, where we're going to look at a bunch of classic papers in computer science and adjacent fields, do a deep reading of them, and then reflect on them here on the

show and share how they relate to our visions for what the future of programming could look like. We're going to start the series by looking at a foundational paper by J.C.R. Licklider called Man-Computer Symbiosis. This episode has been sponsored by Replit, Glide, and a new one, TheaterJS. But just before we get into the paper, I'm going to let Jimmy introduce himself. I've been a big fan of future of coding for quite a while. I reached out to Ivan to kind of do this little, you know, talk about some papers simply because I really like where Future of Coding is going.

I like these interviews that are kind of more speculative, more in-depth, and less about exactly just showcasing work. And so I don't really have any work to showcase. I don't ever finish my projects. I just play with things. Never really have anything to show for it. So I think coming in and talking about it sounded a lot of fun.

Yeah, I think that puts you in the same company as everybody else in the slack and i don't i don't know that anybody in the slack other than mariano has actually finished a project and and put it out in the world that's unfair but i mean that's definitely the the sort of the norm of our community is everybody's working on their far future ambitious goals. And I think it makes for a nice change of pace from the regular working world where you have to, you know, grind out software that doesn't feel very meaningful or significant. And, you know, it pays the bills. And then we can all come here and talk about like, oh, let's radically transform the way we all live and work. And that's a kind of a utopian fantasy kind of thing that I am very much on board with. So I mean, exactly like

finishing software is the the least fun part of software, right? Like, yeah, at work having to do that, like, well, yeah, I guess I need to make this CSS all nice here, or I need to go add in metrics and ensure that if something like those are just never the fun things right it's always the that initial discovery phase that thinking through the problem it's it's learning right like that that to me is the the fun part of all of this it's just learning new things and imagining some possible future rather than uh necessarily realizing it there's that saying art is never finished, it's abandoned. I can't remember who that quote belongs to,

but that I think parallels very nicely the adage from engineering, the first 90% of the problem takes 90% of the time, and the second 90% of the problem takes 90% of the time, and yeah. Yeah, absolutely, right? And I think that this is, I think that we should, you know, foster this, right? Like, I don't think that this is, some people think like the fact that no future of coding project has like taken over and become mainstream is, you know, some indictment, but I don't really, I think the fact that we're all learning and discovering new ways and exploring things

is really the end goal. I love that it has filtered into the way that I do my own work in my regular job. I'm now routinely looking at the kind of the ho-hum SaaS software that I play in with a different lens and thinking about it in a different way. And it's much more like there's all this talk about like user-focused design and and sort of obsessing about the experience for the end user as a very important part of the concept of what software should be and i i like having this as a lens because it takes me i feel even deeper into that realm where i'm thinking about things not just in terms of like how will somebody use the software I'm making to

do their job at their, you know, at their business or whatever it is, or, or in my case, like, how will they use my software to learn about the things they need to learn about? Because I do learning software. I, it gives me this, this awareness that like the computer is supposed to be so much more of an empowering tool than what it currently is. Like what it is now is so much closer to a closed down, dumbed down appliance that is dynamic only in the most limited ways. And that there's this rich lineage of thinkers who said, you know, computers are supposed to be the bicycle for the mind they're supposed to be you know this transformative tool for thought and they're supposed to allow you to work with other people and the world around you in ways that heretofore were unimaginable and in some ways it feels like that came true it may be in a limited extent or in certain local situations but it's definitely not as broadly distributed and as broadly realized as I think many of the people 40, 50, 60 years ago thought it would be.

And so that's something I think I'm going to really enjoy about reading these papers that we're going to go back and read together. It's going to give me even more of an acute awareness of what could have been and maybe someday what could be. Yeah, absolutely. I think that was a great segue into this paper here. Man-computer symbiosis by J.C.R. Licklider. I do want to get a disclaimer out from the beginning that we probably today would not call this man-computer symbiosis. We'd call it human-computer symbiosis.

Oh, yeah. See, that's weird because I took this as like male computers as opposed to female computers because i figure you know at the time they were really big on clear gender roles and so we've got to make sure computers can participate in the in patriarchy the way scientists and researchers were at the time but uh no your your explanation makes a lot more sense yeah i mean if you look at like the turing test it's actually like trying one person has to pretend to be the woman which is really weird uh but yeah so no uh this is i think we should definitely call this human computer symbiosis yeah uh you know to lick lick is how uh jcr lick lighter preferred to be called by everybody so i call him licky but uh

you know you teach their own uh to lick's credit he mentions human beings quite a lot so i don't think he was trying to like yeah so heavily gender this but you know it's 1960 and uh yeah that's what they did and i think we can both agree that like we're we don't want to do that and maybe if we read some quotes we'll try to change it to be gender neutral but we might you know slip up in general, we mean, because we're reading, right? But like, we mean human computer symbiosis. Yeah. Awesome. Yeah, I just thought that was important to get out there. It's one of those like, great, I think great paper, but kind of title doesn't really stand the test of time here. So yeah, I, you know, this paper, I chose this paper, because I think

it's a really good introduction into exactly the things you were talking about. This kind of grander vision for what computers ought to mean to people, to society. I personally think it's a really... I don't know if I would say it is the exact vision I want, but it's a really interesting take. We have the bicycle for the mind as

like kind of the slogan that a lot of people have used but i feel like symbiosis has kind of fallen by the wayside as a metaphor to use and i think it's a really provocative interesting one yeah it reminds me a little bit of like sort of early transhumanism and and eventually i think it became more and more science fiction than than something that people were pursuing in a serious way. But this idea of like augmenting the human being, like we've got Engelbert's famous work in that area. And there's just something about focusing on the relationship between the person and the computer so early on in the history of the development of the computer um that i think makes this paper have this energy to it and excitement that i felt reading it just because it uh it does such a great job of of imagining where things will go and i struggled to find examples in this paper where it actually made a prediction that didn't turn out to be true. Like it does an amazing job of looking at very primitive technology at the time where

computers were still, you know, room scale batch process things where you have to prepare a program offline and then feed it to the computer and it will do its crunching and give you back a result. And then you take your result and do more work offline while other people were using the computer. There wasn't much in the way of time sharing yet. And it imagines, you know, with the march of technology and the advancement of what it is that enables computation. And this is before the transistor era too, if I'm not mistaken. This is before microprocessors were a thing for sure.

It imagines something that i think became a reality in the personal computer revolution in the 80s where people and computers had this much more intimate relationship and it to me it just nails it like it's it's uncanny um so i think well i have no segue jimmy it turn. I would say, I think that's really interesting because, uh, I do think it gets a lot right. Right. Like, don't get me wrong. I do think it predicts a lot. There's also a lot of things though, that it like really gets wrong and like being too

pessimistic about where, what's going to be possible. Yeah. Timelines and that sort of thing. I don't mean timelines. I mean, just in terms of like capability and what the relationship between the person and the computer will look like when it happens like it's it it doesn't arrive in the order that licky imagines it will uh but um it's i like there were very few if any things that i noticed that stuck out as like oh yeah that's nothing like what we have today see i guess I would actually take the opposite.

I think that he predicts some good things, but I don't think we've made it yet. Yeah, I mean, and that's, I think that's going to be in the eye of the beholder. Like there's definitely a pessimistic sort of hate read of this paper where you're reading it and you're seeing all

the ways in which the computer industry has failed to deliver on the promise and i mean that's definitely on brand for for us but um i don't know i i kept catching myself reading it in that light and thinking you know what that that might be me projecting into this my imagination of what i think licky would have wanted based on you know his wild imagination or like like an alan k type figure where they're just sort of perpetually underwhelmed by the the industry but there are i think ways of interpreting what lick is saying that that are fulfilled like the criteria are fulfilled by what we have today but i think this this is probably something we can hash out as we go through it in more detail yeah i i totally agree i think that there are ways and but i i i do think

i think it's okay let's let's just dive in i think to how he so we can try to figure out you know where are we on this you know i i tried to find like an interview with him on him like definitively saying whether he felt it but i mean not that that would really matter but i i couldn't find anything where he's like yep we reached it and my goal happened or not right uh but okay so he starts off by defining symbiosis, the general concept, right? He talks about the fig tree and this larva that helps the fig tree grow. And he talks about this as a living together in intimate association or even close union of two dissimilar organisms. That's his idea of what symbiosis would be so if we make this a human computer symbiosis it would be that we are living together in intimate

association and we are two distinct organisms and i think that one's really interesting because he goes on you kind of talked about this like augmenting idea, but he goes on to say that what he really wants to contrast symbiosis with is what J.D. North called mechanically extended man. And I couldn't find actually, I tried to track down like the original, who is J.D. North and what did he say? It was very hard to find i could only find this paper when googling that yeah on that note i i wanted to point out something um the the reference is jd north's the rational behavior of mechanically extended man which uh is also credited to bolton pa Aircraft Limited in England. And I just like reading that made me long for the era

where like an aircraft company, like some random aircraft company in Europe was like paying some researcher to think about like, oh yeah, this hypothetical technological future where machines and people are unified in their existence. And it's like that's a a glorious age of great ambition in technology and business together that in some ways still exists but it definitely is a different vibe than it did looking back on it back then through rose-colored glasses

so that just made me very happy i think all of that's been outsourced like i know there's a whole consultancy that like reads sci-fi novels and then helps companies like read the sci-fi novel to predict the future yeah that's i mean that's cool but it's also totally boring compared to like you know like an aircraft company hiring somebody to go do this like tinkering workshop sort of experimental like you know we're building the the future of of humans with you know mechanical arms or whatever it is that is this like man machine unification because yeah that's something lick uh distinguishes here when he's talking about person computer symbiosis there's this separation between machine and computer they're differentiated and that's i think interesting and probably pretty important at the time considering how

early computers were in their evolution that i think has been lost to time because these days i will often call a computer the machine like to refer to refer to it in a different way when I'm writing. And I think of it as a very machine-like thing, even though it's not very mechanical in the conventional sense of the word. It's still something that differentiation, I think, has diminished over time as computers have become more sort of ubiquitous maybe and and this is an era where the computer is a very sort of special new kind of thing yeah i i definitely agree that there's this strong distinction here between machine and computer but i almost feel like today we couldn't draw that strong of a distinction like everything that is a machine is computer controlled it's you know

like i feel like the only reason that this distinction is being made is just because of the lack of ubiquity you know that the newness that computers have and i feel like today we could even say like uh you know mechanically so he talks about mechanically extended man uh being that there are certain mechanical parts of the systems that were mere extensions, first of the human arm, then of the human eye. So, you know, we're thinking of like hammers or spears or, and then like human eye glasses, telescopes, microscopes.

I assume these are the sorts of things he has in his mind. Yeah, that would make sense. Not like prosthesis and, you know. Yes. This isn't the futuristic mechanical extension. This is like, yeah, we're doing that, right? Here's these things that we couldn't do without our tools.

Yeah. And so I think if we're, you know, extending that to computers, there's lots of things that we just couldn't do as well without computers, but they're just mere extensions of things we already could do. They just make them like more convenient or more powerful.

Right. Not like not like different in kind, just different in magnitude. Yeah. Like he talks about the ways in which these mechanical extensions have been sort of adopted by industry as taking functions that humans were already doing and sort of automating them and it being about maybe reducing the role of the human factory worker or something like that um and that sort of relegating human operators to the to the to the jobs that only human beings can do and anything that can be automated is automated by machine and that it's sort of there's a i think a sadness i felt in reading lick talk about this because it it puts the machine in a very kind of narrow unfulfilling category where it's just doing this kind of rote work that's of a mechanical nature and it's just displacing human workers. And it's not something that I think

he really craves in this piece, which is that symbiosis between person and computer should enrich both of them. And there's a passage in the previous paragraph where it says, the resulting partnership will think as no human brain has ever thought, so that's a benefit to the human and process data in a way not approached by the information handling machines we know today. And that's the computer part. And I see this, like this desire for both sides of the relationship to be enriched. And I think there's some, some, like we have the term these days, mechanical empathy, but this is, you know, once once again we're disentangling those things this is like computer empathy in a way that i think is um it's very

interesting considering the era yeah and i love his little phrase to like encapsulate this idea of like we just put humans in the places we couldn't automate as uh humanly extended machines yeah yeah um i i think that that is a very little clever turn of phrase and i mean i think that's kind of the the natural thing we do today i mean you think like factories of course but also like call centers right like uh the the owners of all these call centers or the consultants or you know the whoever is running these call centers right and making all the money off of the labor of these people would love to just get rid of all of those people if they could, right?

They would love for this to be a fully automated system, but they can't. And so they kind of slot people into them, but then don't even give them, like, agency or freedom. Yeah, it's very much the half-step that Mechanical Turk sort of popularized among computer programmer type people who suddenly became aware that that is a whole thing where there are many jobs that people do where they're serving not as a as a free thinking individual person but as something that is easier to program to do a certain kind of complex task than a computer is and so for as long

as it's cheaper to employ human beings in this sort of automated mechanical role they'll continue to do that right up until the point it becomes truly you know mechanizable automatable and then and then the humans are pushed out of the picture yeah exactly and so i think this is where his criteria is interesting, because even augmenting human intelligence doesn't seem to be enough to lick for us to have this symbiosis. It really has to be that the computer is kind of living a life of its own, and we are also living our lives, and then we're somehow meeting up in the middle and helping each other yeah um and if the computer's kind of sitting there idle unless we need it for a you know a particular use it seems like to me at least that wouldn't be quite this symbiotic vision yeah yeah that makes sense yeah so he moves on from that though to say like um and i i think this is a fun little glimpse into the era, right? That, you know, a lot of people think we don't need symbiosis because artificial intelligence is right around the corner.

In fact, here it says that the Air Force has done a study and found out that it's going to happen in 1980. And so why even make symbiosis? Because computers are just going to take over. I love his argument about this. That it's like, okay, so it's going to happen in 1980. So there might be like, you know, 10 or 15 years where we can actually have human-computer symbiosis before full AI kicks in. And it's like, well, those are going to be 10 or 15, as he says, intellectually the most creative and exciting in the history of mankind so it's like

this vision that he has is something where it's like well even if it is doomed from the outset we might as well make hey well the sun shines and enjoy this this uh period of unification which is of course what came to pass and now we live in the post-ai world and uh it's not uh it's definitely not okay fine so some of the some of the things I said earlier about accurate prediction are called into question here. But well, actually, I think in this case, he hedges his bets in a nice way. And almost like, makes me think, was he a little bit more skeptical than he wanted to let on? Because he says that it, it might be 15 or 10 or 500 years right that we before we get ai and uh 500 might as well mean like yeah we're not getting that right um yes but at the same time he like he double hedges where earlier

he says in short it seems worthwhile to avoid argument with other enthusiasts for artificial intelligence by conceding dominance in the distant future of cerebration to machines alone and so he puts like he doesn't just say it seems worthwhile to avoid argument with enthusiasts for artificial intelligence he has to say other enthusiasts implying he is also enthusiastic for artificial intelligence because of course ai enthusiasts are so argumentative that if he doesn't you you know, pull that punch, they're going to jump down his throat about this. So, yeah, I love it. I think it's very clever little uses of language.

It kind of reminds me of like philosophers who were very clearly atheists who couldn't admit as much. And so wrote text where they're like clearly, I don't know where they're like clearly praising God's existence or whatever, but at the same time also questioning it. Yeah. Kind of like they're on both sides of the fence so they can have, you know, plausible deniability. Yeah.

That, that is a great analogy. Yeah. Cause in this case, like AI is definitely God. So, uh, we're, we're just, you know, in the glory days before God has been invented and then we're going to all have to live in, in holy servitude for the rest of our pitiful existence but thankfully we're not we haven't lived out those 15 10 or 500 years yet yeah i know that i can't remember who but there's one person saying some thinker saying well you know once ai happens they'll look back and see who who was who were the believers and who weren't and yeah so uh praise praise our machine overlords whenever they arrive.

Absolutely. Yeah. No question. So yeah, this is kind of his first, like, maybe you don't buy my vision. And, you know, I think he does a good job of like, okay, we can have both uh but then his second maybe you don't buy my vision is uh really that we don't need symbiosis because uh he says that programming right forces us to think clearly and that just that disciplined thought process that we need to like bring our problem to the computer might just be good enough and i i hear this sort of argument actually quite a lot today.

I think this is still a very commonly held view. When I talk about kind of like REPL-driven development, I'm a Clojure person in my day job, and I really like Clojure. And so I work at the REPL interactively doing lots of things, and I'll kind of hear like, oh, you don't need that. Like, think through your problem very precisely and you're all good. Yeah. It's like some people could uncharitably say that it's, you know, it's being lazy because

you're just kind of bashing away at it until you get the right result. Kind of like the same criticisms you throw at types or at tests or at any any number of tools that are supposed to help um give you more i would say dialogue with the tool that you're using to do problem solving and i i think it's like it's a fair criticism i just think it's misapplied like there's great value in realizing that um and that's like the whole point of of oblique strategies and other techniques for creative thinking and spurring that kind of creative problem solving mode of thought is that there are benefits to having a tool that gives you a way of thinking about something and that forces you into a way of thinking about something and that way of thinking about something being different from the way you are naturally

inclined to think. And so I guess this is like thinking fast and slow, that kind of thing, where sure, you know, a good programming tool might be so much easier to use that it keeps you in that fast thinking mode where you're not deeply thinking about the problem you're not engaging with it but i think that that's it's one of those lenses where i think there's no clear right or wrong about it and what you really want is you want the ability to have both and so i think that's where this argument fails here at least is it's saying oh we don't need symbiosis because you know programming helps you do thinking anyways and it's like well sure but not always and it shouldn't necessarily be always and just because you can have something to have a desirable outcome from the status quo doesn't mean we

shouldn't stop looking for alternatives to that status quo yeah absolutely yeah and i but i think that i do think is a very good argument against it but i don't know if it's licks like lick wants to say that uh really what what we need to do with symbiosis is not only like have it help us solve problems but help us come up with the problems to begin with and maybe those are kind of like the oblique strategies and some of those things but i think this is actually you know to have our contrast here of uh you know the the positive reading that we've really gotten there and then i guess my pessimistic reading that we haven't uh i i you know i think that he says we need to bring computing machines effectively into the formulative parts of technical problems and i don't know that we really have that

yeah and that like that was even i had that highlighted in the uh the abstract or the summary at the very start at this this distinction between um formulated problems and formulative thinking the sort of the problem finding part and that's something where i once again i like i feel like there is an optimistic read of that where something like a spreadsheet or something like muse or something like you know any of these tools that that one could call a tool for thought in that sort of contemporary sense they they do give you a little bit of that you know relationship with the dynamic medium that you don't get from something like a static pencil and paper or you know other kind of outside the computer tools where you're you like it's not

gonna help you with thinking in the way that lick is describing where it's like you know that the computer can summon data and perform calculations and do all that kind of stuff automatically and and sort of imagine the right kind of data to go and fetch just in the nick of time sort of thing but it is letting you use the computer as a little bit of an out like an external mind in a way where there is more that you're able to think and and more that you're able to imagine through using the computer and you're thinking through the machine rather than it just being like a you know like a slightly fancier analog of paper where you have to drive the entire relationship but it's you know it's it's keeping a good record of your thoughts as you go or something like that like i think there is

i think this is one of the cases where that symbiosis is something that i actually feel came to pass where computers are doing more lifting in that relationship than we might think but this is a this is a tricky and subtle thing, and I'm not sure exactly how to point to it, other than to say it's something I feel. Yeah, I mean, I definitely see where you're coming from. And I do think this trend towards tools for thought is definitely a positive trend, and definitely is leaning more into this realm, right? I guess where I find that I'm always, I guess I always feel like I am doing all the driving, right, in some ways. I mean, don't get me wrong, I love Muse, I love,

I don't love spreadsheets, but that's just because I'm weird, not because they're bad. I just can't ever figure them out. But I know they're really powerful. But anyways, I do really love these tools. But when I want to think about something dynamic, when I want to think about a process over time, I feel like I am being the one that has to be driving every little thing

and that my computer can't even help me think about just the unfolding of events over time. If I want to think about something static, the computer's really fantastic at it. If I want to think about data, if I want to think about calculation, but as soon as I'm like, I want to think about this computer system, and how do these messages go across them,

and what would happen if this queue backed up, or whatever, these are the things I'm thinking of because I'm a software engineer. messages go across them and what would happen if this queue backed up or you know whatever right like these are the things i'm thinking of because you know i'm a software engineer but like yeah i do feel like there's some uh some lifting that the computer could be doing for me that uh i guess i i wish it were yeah like if you look at nikki case's work work on things like the transmission of viruses or their early work, Parable of the Polygons, these dynamic simulations that are for learning, like people call them explorable explanations. Those things are made by writing code. And the design of the code is something the computer does not currently help with. You have to come up with that design for the

dynamic system entirely yourself and then once the system exists other people can use that system as a thing to think with and to bounce ideas off of and to experiment with and explore and learn through and it there's i have this acute feeling i'm sure many many people do that the creation of those kind of explorable explanations is something that computers should do a much better job of helping with and that if they could we'd be a lot closer to i think something that uh lick is craving through this paper something that he's imagining might come uh but yeah that's that's definitely not the case yet yeah completely agree i think that that's a great way to put it uh and and if we continue on with this paper i think this

is where uh for me at least the the organization's a little strange here we go from like all right here's some symbiosis here's what it. Here's why you might be skeptical to, hi, I'm Lick, and I decided to keep records about myself to see what my day is like. Yeah. Can I actually read a passage from this paragraph? Yeah, yeah. Let's, for sure. Despite the fact that there is a voluminous literature on thinking and problem solving, including intensive case history studies of the process of invention, I could find nothing comparable to a time and motion study analysis of the mental work of a person engaged in a scientific or technical enterprise. In the spring and summer of 1957,

therefore, I tried to keep track of what one moderately technical person actually did during the hours he regarded as devoted to work and i read that sentence and i'm like oh shit you're just you measured yourself didn't you lick you used yourself as the subject of this experiment like it's it's one of those those sentences that i think he's intending for it to foreshadow where this is going but it has that sort of feeling where i expected the you know like the narrator from arrested development or some of like, meta textual layer to come in and say, like, this is going exactly where you think it's going. And he continues, although I was aware of the inadequacy of the sampling, I served as my own subject. That just, this made me so happy reading this. Like, it's such a,

such a quintessentially dorky, but in the way that just warms my heart. So yeah, the rest of the section is just a description of like recording his own work in a, in a time and motion study. And I had to look up what those were. And they're basically like just keeping records of like, okay,

at this time, the person is doing this at this time, the person's doing that. That's the time study. And I believe the motion study is like, first, at this time, this person is here. And at that time, that person is there. And it's like, just a way of recording data about the activity. Is that right? Jimmy, does that line up with your understanding? Yeah, that lines up with my understanding, at least I do have to say I love this. That the right after the paragraph you just read. It soon became apparent that the main thing I did was to keep records, and that the project would have become an infinite regress

if keeping records had been carried through in the detailed, envisaged, and the initial plan. He's like, okay, if I keep records about keeping records, I'm just going to be keeping records. Which is just, like, of course. But yes, I do think this is just a a fun little passage and would totally not fly yes in scientific world but even like the corporate world like if i was like yeah i did a case study of me yeah and i found that i don't like this and we should change it like or even like hey i'm gonna publish this paper in nature or

whatever and part of it includes a passage of how i wanted to record myself but recording myself recording myself would have been recursive i think the the reviewers would have said no you have to take that out that's too cute by half very very funny no that's not acceptable but yeah that's like so so dorky so heartwarming yeah so what he gets out of this study though i think is probably pretty uh pretty accurate at least i think almost certainly of of his time but i think for me as well uh that he's trying to do this intellectual activity he's trying to make decisions he's trying to think through problems but that he finds that most of his quote unquote thinking time was really clerical or mechanical. It was searching, calculating, plotting, transforming, determining the logical or dynamic consequences

of a set of assumptions or hypotheses, preparing the way for a decision or an insight. So it's, you know, it's not actually making the decisions. It's not actually doing the end product and pondering ideas. It's doing all the things you need to do to be able to think about the thing. This is also where my pessimistic reading here. I feel like that's what I do.

The computer is definitely a convenient tool for doing that, but even just plotting data, I have so much data I want to plot that I can't even just use... I try 12 different plotting libraries and all of them are inadequate in some various way and I end up spending 12 hours of time

just trying to get this data plotted rather than actually getting to think about my data and this is something brett victor touched on really well in drawing dynamic visualizations and some of his other work where he shows you know there are libraries code libraries you can use for plotting and diagramming and that sort of thing but to use them you have to use a very narrow um very technical very, symbolic programming interface that does not feel related to the actual product that you're trying to make the actual graph or chart or whatever it is. And it doesn't feel related to the data you're trying to bring into that chart

or that graph. That's on the one end of the spectrum. The other end of the spectrum is something like illustrator, where you can just do your plotting by hand and you have some put down every data point pretty much entirely manually and there are some things that are kind of in the middle ground but they fall short for various reasons and those are explored well in brett's work but there's no great tool afforded by current software for doing exactly this kind of work and i think that it's it's worth generalizing from that to imagine all the other different kinds of work that when you're not using a computer requires or when you're using a very primitive computer it requires still doing a lot of manual orchestration and a lot of you know hard data measuring and crunching and preparation or information gathering

or whatever it is that's feeding into your work. And imagining how the computer software that we have today, in some cases can help with some of that, in some cases doesn't help at all. But I do feel like there are some tools out there that do a very very good job of this already and i'm thinking of things like non-linear editing for video where the the old process of editing a film was very manual it required actually manually cutting strips of film and working out timings per frame and and it was a very destructive and slow and methodical process. And nowadays, like video editing and audio editing and other sort of artistic tools are so much more flexible because they are non-destructive and because you can work with tremendous amounts

of data and processing very quickly. Like another example is in music production. The Beatles famously had a four-track recorder for most of their early records and later had an eight track and and they did a lot with it but they certainly never broke away from the pattern of having people in a room playing a song sort of through from start to finish or at least pieces of a song from start to finish and and capturing those and then maybe adding ornamentation after the fact in other layers but they never approached something like what you'd get with electronic music nowadays where the the music can be made entirely from little microscopic snippets of sound being assembled almost like you know pieces of

scrap paper on a on a table spread out in a collage or something like that like there are different ways of working that are afforded by the dynamism of the computer and i think this is one of those situations where sure like when it comes to data processing or when it comes to number crunching like there's still there are still some domains that are underserved but i think at least in the artistic domains we are living in the future that is promised by this this vision lick is putting forward yeah that's actually a really great point i guess uh maybe maybe a lot of my uh my pessimism comes from like being in the one domain where the you know you're doing something quote unquote unique right like? Like it's, it's not, you're doing general data processing rather than domain specific, you know, uh, data processing and, and I'm building the tools to make

people be able to do the domain specific things. So, yeah, I think that's actually a wonderful point. I hadn't really thought about, you know, music, film, et cetera, and how much less yeah how much you really do get to think a little bit more about those i i guess though from the outside take it almost feels like having never done that old those old processes it still feels very manual to me all right and maybe that's just me like always uh being unhappy with computers right whatever right but like i don't know i'm not uh i'm not i can i can play i can play the bass you know like i played the bass the the stand-up bass in in high school i have a bass guitar i can you know make a few chords but i'm definitely not like a musician by by any stretch of the imagination

but it does feel like it could be way easier for me to make music. Sure. I mean, there's, there's garage band and other things like that. Like there's, there are computer tools out there that mean you don't have to, like you can, you here's, here's another way of looking at it. Music and acoustics and the theory thereof can be split into a lot of different categories and a lot of different aspects and you can pretty much pick any one of those aspects and say i don't want to think about that specific thing and there will be a piece of software that says okay i will think about that thing for you whether it's like uh chord progressions there are software programs

out there that will just generate chord progressions like if you look at microsoft song smith or whatever where you can feed it some musical ideas and it will generate a whole progression and set the the singing that you do to uh to an accompaniment for you so you don't have to worry about what the chord progression is or there are things for handling rhythm like quantization there are things for handling i don't want to have to generate tones. Okay, there are all these sample libraries and tone bank libraries. Just another example of a different domain that I think might be easier to relate to is the difference between using a typewriter and a word processor, where a word processor does like a staggering amount of work that is so well handled that we just take it for granted and it's it's basically invisible to all but the people who have actually tried to implement one of them from some degree of scratch compared to what came before where you know a typewriter also a tremendous advancement over having to write out notes by hand where you know you have to develop a style of writing and a

style of thinking that works well for putting down an entire page of text all at once without making mistakes like the the way things were written before the typewriter versus after the typewriter is um just just amazing to me and the difference in in how people phrased their thoughts and now you know with word processors it's it's so much more flexible than that and in the ability to rewrite and rewrite and rewrite and rewrite at a at a microscopic scale um you get so much more like delicately crafted writing so much more easily that i just i i can't imagine what it would have to be like if, you know, the, the computer word processor went away and I had to do all my writing in the

analog way from now on, that would just be debilitating for me. Yeah. You know what? I'll, I think I'll give you this one. I think that you're, uh, looking back, I had a, uh, in high school I had a AP Kim and, uh, my, the first AP Chem. And the first semester, we had like 14 chem labs and we had to write up these big reports. And they were like 20 or 30 pages each. And in the second semester, my teacher decided that we had to do them all, all these lab reports by hand on carbon copy paper.

Oh, yeah. And we were not allowed at all to be typing them out. And every single mistake we made on this carbon copy paper would be half a point off. Oh, man. Any bad handwriting, et cetera. And so I failed the second semester of AP Chem because I refused to do the,

I went to all the labs, I did all of the actual labs, but I refused to do a lab report handwritten like that because it would have taken me, I already spent 50, 60 hours of work on each of these chem lab reports on the computer.

And we had plots, we had to make graphs and charts, and I wrote programs to do calculations for me, and unit conversion, and all this stuff. So you know what? I think you're right. Not only does the computer make that one so much better for me, I just actually refused to do it without the computer's help. Yeah, and I had similar assignments certainly

they weren't in chem they were more commonly in english or at least other places where you could make a defensible argument that that restriction was actually relevant as opposed to some sort of you know anarch not anarchistic anachronistic it's okay it's anarchistic also but some anachronistic you know throwback to here's how real chemists did it in the 1970s or whatever but yeah when i had those assignments where it's like you have to write this out by hand i'd write it on the computer first and once i had come up with the thing i wanted to say then i would actually write it down by hand like i think that's the the indictment of the manual method is that given the choice you know this goes back to our previous

unreleased episode of this podcast about you know given the choice between the augmented thing and the non-augmented thing i choose the augmented thing every time yeah yeah no that makes perfect sense so you know we're in the i think we're in the weird section here right where we now we're talking about like lick going and studying himself. And then we're talking about just like in general, how do computers and humans compare? And like, it's like, these are all fine topics. It's just the ordering of all of this is very strange to me.

Yeah, I'll trust you on that. To me, it seemed to make make sense at least in that i could kind of feel how each idea built on the preceding ones but certainly like you could chart a different course through this space and it'd probably be more sensible i just it didn't stick out to me but that's yeah yeah so so what he i mean ultimately what he says is really humans are flexible and we can go and do many different things, and computers are fast and precise. Which I know probably needed to be said in 1960. That they each have their own strengths, and that we want to have those two strengths complement one another. I'm very pleased to introduce you to our first sponsor this week,

TheaterJS, which is a project that comes from a member of our community, Aria Minai. And if I am mispronouncing your name, I deeply apologize. TheaterJS, it's this project that I have a long relationship with, and I'm going to talk about it. I'm going off script. Aria sent me this lovely script for what to talk about. I will read that in a minute. But this is a project that is near and dear to my heart. And so I just I can't help myself but share some of my own enthusiasm for it first before we get into the, you know, here's here's what it is on the package. And, and here's the call to action kind of stuff. Theater.js, it's a, I'll describe it in my own words, it's a JavaScript library that you include in a web project that you're working on. And it gives you an interface for animating properties of that project. So you can pick, you know, a DOM node, or something in 3.js, or just some variable. And if you want to have that property varied over time, you know, typically how you do that is you'd have some linear interpolation or something like that. And you'd write the code for like, okay, start at this value and end at that value and take, you know, one second to do it. And you just math that out. And, and there you go where you say like oh the property has this current value and this change each frame and so every frame just increment the value until it hits some goal and then stop theater js does a wonderful thing in that it takes

a kind of an interface that i know of as called an f curve. And it lets you use an F-curve editor to control how these values in your scene, in your project, change over time. And this F-curve editor, it's something that comes from animation tools used by real legit artists, real legit animation tools, like all the 3D animation packages that I came to know and love as a teenager because that's that's my background my background's in 3d animation and in those tools what you have is a timeline view so you have you know a horizontal axis in the cartesian plane that is your passage through time and then a vertical axis that can be sort of numeric in nature for whatever value you're manipulating and you can put down key frames along that time so at this moment in time the value should be x and at that moment in time the value should be y and then not

just set okay i want to interpolate between them in this linear way but actually draw using bezier curves or other methods of of drawing curves, how the value should transform as it goes from one to the other. And this is a lovely interface for animating and for creating motion and for creating something that moves with expression or that changes with expression. And there was this project from Airbnb called Body Movin',

which would let you do this sort of F-curve animation in Adobe After Effects, and then export it from After Effects and then import it into Body Movin and use it in the browser to animate your properties. But that workflow, that should remind you of something, right? You prepare something in one program in one environment that is disconnected from the thing you're actually working on and then you pass it through some compiler-like process and then you finally get to see the result later on separately removed from the place where you were working on it and i hate that you know that about me i hate that disconnection between the environment where

you're doing your work and the environment where you actually get to experience the result of that work. And Theater.js closes that gap. It lets you do your animation work right in the same context as your project. It lets you open up a window within your browser and plug in the animation curves that you want to have right up against your own running code. So whatever it is that you're building, whether it's an explorable explanation, or it's some kind of cool bit of motion graphics on the web, or it's some kind of 3d game or whatever you're doing, you can have your system and your animation interface right in the same space. And I think that that is incredibly cool. theaters being built by a small team of designers and developers, and their mission is to enable designers to code, programmers to

design, and someday empower everyone to create. For that goal, they're building a design environment that enables direct manipulation, composability, and remix. If that sounds like something you want to build, head on over to join.theatrejs.com. And once again, thank you to TheatreJS for helping bring us the future of coding. All right, so we're in this section, the separable functions of men and computers, an anticipated symbiotic association. Yes, and he kind of starts off with some interesting things like a progression. He thinks that over time, computers are going to play a stronger role.

But initially, humans will set the goals and supply the motivations, at least in the early years. They will formulate hypotheses. They will ask questions. They will ask questions. They will think of mechanisms, procedures, and models. And some of this, though, he kind of gets a little bit wrong. Maybe it's in the early course of things, but at least maybe we succeeded in doing this. Where, like, he says we have to remember that

such and such person did some possibly relevant work on a topic back in 1947 or at any rate shortly after world war ii basically he's saying like the computer might not be able to search all of this for us and we might have to remember these things yeah and actually this is probably the thing that computers are the best at today right like i don't have to memorize relevant work at all i can just search for it and find fountains of of relevant work which is funny because i think it's not necessarily that lick wouldn't have thought that the computer could help with that but maybe that that would have been a hard thing to convince people of that you know

computers would have that kind of search because it's certainly in keeping with other things he says in this in this paper that you know the computer is going to be able to make these sort of um dynamic connections and surface suggestions and that sort of thing which we'll talk about in a minute because i've got some thoughts about that well it seems and we'll get to this as well but it seems that he really believed much stronger in the computational aspects of machines more than the storage aspects yeah that it was really about processing that they were going to be really good at and not information retrieval and storage yeah which i guess makes sense given at the time storage was really impoverished compared to computation at least in terms of you know what it physically took to make

storage happen like you know giant drum memory and uh if i remember correctly this was before tape memory was a common thing even um but i might have my history wrong about that i'm sure yeah i think we're still doing like mercury delay lines and things at this point yeah yeah which would just be wild yeah um yeah i can't and and we'll get to that he mentioned he talks about like kind of the current state of computer memory and like the next section on this on this bit that you read about people setting the goals and supplying the motivations formulating the hypotheses asking the questions thinking of the mechanisms procedures and models as opposed to what the computer will do i found that interesting because and this is going to be

a perennial theme on this series of uh episodes where we look at papers and talk about them i think this is almost the inverse of what a video game is where a video game now granted i'm gonna have to you know put uh sort of a caveat on this. Yes, video games are designed by people. And a lot of the things I'm going to describe and asking the questions thinking of the mechanisms procedures and models and then it is you the player who has to work within those constraints to fulfill some systemic objective and i found that really interesting to think about because it's possibly what makes a game a fun thing is the sort of the artificiality of it, the sort of, you know, hey, let's switch roles, you and me for a little bit. And I'm going to role play as the one who has to do the busy work to fulfill some objective. And you, the computer are going to, you know,

continually surface new objectives for me to fulfill. And that's like, I think the inverse of what Lick is suggesting the role should be, at least when it comes to this sort of like creative problem solving kind of work. Yeah, no, I think that's really interesting. I do think that at least the games that I'm always really interested in are these games where there's, you know, some system established and you're trying to figure out the bounds. You're trying to explore the conclusions of the system being in place. So yeah, I hadn't thought about this in relation to that, but I do think you're right. This is very much inverted with video games. It does kind of make me wonder when we talk about programming should be more like video

games if that's not some of the elements we're looking for. Yeah, exactly. And maybe this could serve as the foundation for like a counter argument or some sort of rebuttal against the suggestion that programming should be more like games, which of course they should, but that's neither here nor there.

Yeah, there's this aspect, like as soon as you, and I guess that's inherent to the idea of symbiosis, where it's like two or more different organisms whose whose different natures complement one another in this very fundamental way it sort of suggests that there are these defined roles that computers and people are supposed to slot into which feels very gender-y to me just to take that one up again there are the two genders human computer and um it might be very fun in a in a once again another

fun exercise to to do more inversion of those two roles and to imagine what it would look like if if humans played computers and computers played people and i guess that's the turing test in a certain uh light as well so yeah yeah there's just there's so much to this that i i really enjoyed i know i said that at the top but like especially now that we're getting into this this part of the paper like it's there's some there's some very fun things to think about in here that i i loved yeah no i i completely agree with you uh i think that what lick i think especially like to me it's the the beginning and like this middle bit what lick is offering us is such an interesting way of thinking

about this and i do agree with you that in some ways we've gotten there and in some ways we even found that like like you said with video games like actually it's the opposite that's the more interesting thing yeah but then there's some things like here where he talks about that once we give these goals and motivations and hypotheses, the computer will convert the hypotheses into testable models and then test the models against data, which the human operator may designate roughly and identify as relevant when the computer presents them for their approval. And it goes on to say, you know, the computer will simulate the mechanisms and the models, carry out the procedures and display the results to the operator. It will transform data, plot graphs, and then in a parenthetical, cutting the cake in whatever way the human operator specifies. Or in several alternative ways if the human operator specifies or in several alternative ways if the human operator is not sure what they want that uh that raised some flags for me because yes like that is definitely a hundred percent one of the things we have today is this like

several alternative ways if the human operator is not sure what they want this idea of the computer doing a bunch of data processing and saying hey here are some things you can choose from because i wasn't sure exactly how you wanted this cake to be cut so pick your slice um those definitely exist and i tried to think of examples where they weren't terrible and all the examples i could think of where it's like the computer's like i'm not sure which one of these you mean which one is it they're all they're all bad like i don't know why and i i would love do you know any jimmy like any examples where the computer is gonna say like here are several different ways of interpreting the thing you've asked me for um that are good like is that something that we have yeah no i've never found any of them to be

good right like that's that's one of the things that i think you know to go on my pessimistic reading that we've fallen short of this is i can't think of anything where i can really give like some sketchy idea and be like hey computer like help me think about this and show me different possibilities and then it will you know help me right if i want that i have to make it myself yeah right like i have to code up what those possibilities would be and like the only reason like i have to consider i think beforehand what those possibilities would be and then really what i'm using the computer is is to draw those out right like to actually make a graphic of them because then once i look at the graphic i can be like yeah that one actually works no that one doesn't etc right like it's never like oh if i group by these four criterias then i'll make sense

of this data like try grouping by all of them and just tell me which one works right like yeah it's it's almost never uh quite automatic in that way i usually have to pre-think of the ways i want to slice and dice yeah so just for the for the sake of exploration here are some examples that came to mind for me uh and and i think also relevant see also um hey check it out learn more click the link in the description um is our previous episode on this podcast with ella heppner uh where towards the end of the podcast she talks about aesthetic selection and that as a as a useful tool in doing generative artwork this is related and i think that there's some some stuff there, but I'm going to put that in a box called It's Creative and say that the criteria that you're using any video game where there's a character creator, where you get to move a bunch of sliders around to decide what your character looks like, where that

game also gives you the thing you can click that says like similar faces and it pops up like 20 or however many different variations of the character. And you can pick one of those variations and it'll move all the sliders a little bit. Those similar faces are all always worse than whatever i came up with and they're only useful for making some just monstrous repulsive you know horrifying monstrosity i use two different uh variations of the word monster in there um another example is the the disappointment that is search engines telling you you know did you mean blank um because when they say did you mean blank it's usually taking an uncommon word that you used intentionally and correctly and saying oh it snaps to grid to this

much more common word that is not what you mean and And it's, you know, that's, that's a frustration that is felt whenever it does that automatically. And it says, including results for this thing, that's not at all what you wanted. And that's why you didn't say it. And what I never get from that interface that I wish it would do, and I'm sure what the people who initially designed it thought they were reaching for was something where it said hey here's this other way of formulating your query that will get you to a more precise result so if you have like let's say you have some like diagnostic task that you're doing like let's say you know your your furnace is making some sort of weird sound and you want to figure it out and you go type in like you know weird

sound furnace popping gas smell whatever it is and the thing could come back and say oh here's a different way of phrasing that query that just works better that will get you to you know queries about that instead of queries about you know something completely unrelated like there's a kind of a a need for magic in there a need for like do what i mean not what i say that computers are just bad at and so i think it's it's it's like this if i had to pick one thing out of this paper where lick completely failed to predict how things were going to emerge it would be this it would be this aspect that's but is it that computers are bad at it? Or to kind of bring us back to earlier,

is it that really most of those functions, we are the humanly extended machines, right? Where what they're trying to do, I think like the YouTube algorithm, right? I want to watch some interesting programming content and I never know exactly what I'm looking to watch, but I hop on YouTube and yeah, all those suggestions are really terrible,

but not because they couldn't be good, but because what they're optimizing for is to get me to watch longer and get me to engage more with content that is controversial or whatever, get me to engage more with content that is controversial or whatever. Get me hooked. And so I do wonder, and apparently they are very good at that.

They do hook people in and they make lots of money doing so. And is it really that all the techniques we have that would surface those things, they just require a lot of investment and we invest in them not serving us but us serving them yeah i don't know this feels a little bit like a computer apologist i think you're a sympathizer and uh should be regarded with skepticism um yeah i don't know like it it's the kind of thing where i'd be much more sympathetic to the suggestion that the failing here is that we are wielding the computer to destructive and you know just to the ends of self-enrichment and you know corporate enrichment rather than fulfilling some grander goal it would be it would

be easier for me to agree to that hypothesis or that suggestion if if there were any counter examples that were ready at hand and that just might be like a lack of imagination on my part i'm totally willing to admit that i haven't done a rigorous you know analysis of the field of computers making suggestions um but it's just you know as a lifetime computer user um well that feels dark to say um i i don't feel like this is something that we've figured out a way to do that is beneficial regardless of who's at fault for that failure i mean small at a very small scale spell check is is pretty good at at those things at least there are really good spell checks right if i have no idea how to spell

a word and i google it oh yeah that's almost always right about what how i actually spell that word now i did have i did actually think of that one earlier before i went all off on my bullshit um and i did think of that one and i did think of a really crushing rebuttal but i've forgotten it so i guess it just has to stand unrefuted you're right um yeah no spellcheck is good. Our second sponsor this week, returning once again, is Glide. Glide's mission is to create a billion software developers by 2030 by making software dramatically easier to build. We all marvel at how successful spreadsheets have been

at letting non-programmers build complex software, but spreadsheets are a terrible way to distribute software. They're an IDE, and the software built in it rolled into one, and you can builds a basic mobile app from the data in the spreadsheet. You can then go and reconfigure it in many different ways, including adding computations and building some pretty complex interactions. Then you click a button and you get a link or a QR code to distribute the app. The data in the app and in the spreadsheet will automatically keep in sync. For the Glide team, that's just the beginning. Glide needs to become much more powerful. Its declarative computation system has to support many more use cases without becoming yet another formula language. Its imperative actions don't even have a concept of loops yet,

or of transactions. Glide needs to integrate with tons of data sources and scale up to handle much more data. To do all that, Glide needs your help. If you're excited about making end-user software development a reality, go to glideapps.com slash jobs and apply to join the team. As usual, I'll add my own little commentary in here, as I am wont to do. Gl glide have been very actively involved with our community um i keep talking to people who say oh yeah i was working on this project with glide and we just published this blog post and i'm excited about this work that i did and it's it's very cool to see how they're engaging with people in our orbit who are exploring this space and how they're pushing it from within as well and so i just wanted to to say that I have very much enjoyed having Glide as a sponsor because

that's made me especially aware of all the things that they're doing and to see how they are really pursuing this vision that we all share. And so I just wanted to call that out because if you are looking for a place to work on this kind of stuff, do definitely consider them because it seems like they are very committed to this and they're doing interesting things to push on it. So once again, you can go to glideapps.com slash jobs and apply to join their team. Thank you to Glide for helping bring us the future of coding i think that the the most interesting section for me uh where he gets both like wonderfully correct and wonderfully wrong at the exact same time is this speed mismatch between humans and computers on the top of page seven here i just i have to read this sentence because it just it just warms my heart uh i don't know just how wrong he is but how right he ends up being which is any present day

large-scale computer is too fast and too costly for real-time cooperative thinking with one person and i just love this idea too costly totally give. That a large-scale computer is too fast for real-time cooperative thinking with one person. So I looked up what the fastest computer in 1960 was. Yep. And it is... So I don't even know if he had this in mind, of course, because he might not have known about this very fast computer.

It is a a cray 1604 came out just in 1960 it had a 48 bit 208 kilohertz processor with 192 kilobytes of of memory it was uh mips is measured as 0.1 it had no floating point operations and no storage so it didn't have mips it had kips yes yes that's a good point wow uh yeah so this is this is the too fast for one person right is 208 kilohertz now to be fair that's still pretty fast like like kilohertz is when you're thinking in terms of you know raw data crunching and that sort of thing like i i can actually relate to this i'm not just digging my heels in because i said this paper was without error and all of its predictions were great um that is actually like that's more than i would

have guessed i and this would have actually been fun to do it's too bad we can't edit podcasts and splice this in earlier but i would have guessed um not even in the kilohertz range so that's that's fascinating to me that's really interesting to think about i mean this is a cray right like this is a supercomputer this is you know not your your ordinary computer i wanted to just see you know if we push it as far as possible yeah right and i doubt actually that this is what he had in mind was you know the world's fastest computer in 1960 like like top three or whatever it was back then whatever the ranking is yeah yeah so but. So, but like, uh, I, I get it.

Kilohertz is, you know, it's legit. Like it's not, you know, nothing. Yeah. Um,

but it's, it definitely, I have right now in front of me, you know, my Mac book, I have my iPad and my phone. And then I guarantee this microphone has a processor in it faster than kilohertz.

Right. Yeah. Yeah. Um, and then i guarantee this microphone has a processor in it faster than kilohertz yes right yeah um like it definitely is not too fast for uh for real-time cooperative thinking uh it's really the cost i i understand that he was really talking about cost here but i just i love the suggestion that it's too fast well here's here's let's i mean we're we're an hour and a half in some of this is going to get cut out so we're an hour in uh let's dig into this for a minute so i wonder if what is happening here that would have been likely true at the time and is likely no longer true is at my my best guess and i have no i'm basing this on nothing is that at the time computers were still simple enough that certain people who worked with them still understood everything that they did from top to bottom and it was doing you know kilohertz worth of work but that work was

fully understood and could probably be understood by a single individual whereas today that is not even close to being the reality where there is no single individual who understands what a computer does top to bottom in in full detail there are people who understand in broad strokes and there are people who understand very narrow slices of it and um there was a time not too long ago where it was conceivable that a person could have a pretty good sense of things and i'm thinking like maybe you know around the time of the cathedral and the bazaar that sort of um debate about how operating systems should be built but we are we are far past that era at this point and so it's it's it's the kind

of thing where i think if you're thinking of it in that way if you're thinking that the computer is still something that can be fully comprehended um if it's doing you know hundreds of thousands of operations per second and you're trying to like mentally contain that and think about it at a at an engineering level and not do this sort of big-minded like oh imagine the future where it's you know so much more ambitious and it's it's thinking it becomes a brain unto itself and it has this you know magical quality to it but it's still you're thinking of it like a machine it's a machine that is starting to get scary in its performance and it's in the the rate at which it's improving and

the capability that it's already demonstrating to do amazing feats of data processing it's it's the sort of thing where it might have already started to feel like you know like a monster was being turned loose or something like that like there's just this nuclear explosive sort of power to it that is coming out and it's it's something where i guess i i'm i'm reaching for what i might imagine is informing this thought of it being too fast because it's certainly like lick at the time had that thought like thought you know computers are already too fast to be thinking as a peer to a person and so it's not like he's imagining oh in the future computers will be too fast. He's already experiencing that.

Like that's something present in his day and age. And so. I think there's something to that. That maybe we've lost in the time since. Something that he was aware of. That we're not aware of anymore. Just because the computer is so much more. Abstract and nebulous

and distant from us now than it was for him then. Yeah, I think that makes a lot of sense. And I think that maybe also kind of, you know, playing into that would also be the kinds of things computers could do at the time because they were so limited on memory. And that's what, like, the next section is about memory. I don't want to go there quite yet. But I do wonder if you limit a computer to pure computation, not on lots of data, not repeating that same computation on millions of elements,

but just like, yeah, I'm computing something. Computers are very, very fast. It's almost always that memory bottleneck that makes it so that, and it's the size of the problems we want to solve on gigabytes of data or terabytes of data that really make them feel slow. So yeah, I do think that being able to fully understand them,

but also they're only focusing on the CPU, the computational aspects, and all the the memory requirements could have led him to to feel this way yeah and then we see i think right after this like it's too fast you know we see a vision that is now uh you know kind of being brought into maybe reality through dynamic land. It says, It seems reasonable to envision for a time 10 or 15 years hence a thinking center that will incorporate the functions of present-day libraries

together with anticipated advances in information storage and retrieval and the symbiotic functions suggested earlier in this paper. So here we have this idea of, like we have power lines pumping in power, we end up getting computation lines pumping in computations, and there's the community center that you would go

and be part of this computational thing which which just to me at least very much reminds me of brett victor's vision here for for dynamic land yeah though i i i'll i'll push back on that only slightly in that i think brett victor's um reason for wanting dynamic land to be a public thinking center is a sort of a like a public commons like a public good kind of framing where it's not about um the computer being a centralized thing for the sake of efficiency or anything like that which i think might be a little bit of what's informing licks thinking here just based on the earlier parts of this paragraph but for brett it's it's more about access and not having the very profit-hungry computer industry be the thing that mediates is thinking of here but in either case i think it's like an interesting thing that yeah maybe that will come back someday maybe we will find a way

to get back to the public library being the place that you go to do your computing which certainly like when i was a kid in the 90s we had a computer at home but we didn't have the internet because i lived out in the woods i'm always living out in the woods um and so we would go into town and i would sit at the library and that would be where my internet was and i'd spend hours and hours and hours there just you know surfing final fantasy fan pages or whatever it was but yeah i mean i do agree with you that it looks like lick here is making an economic uh argument you know for this but i think you could also kind of take it slightly different and say that uh you know if you kind of put this brett victor dynamic land spend on spin on it right it would be that and by having

these thinking centers it would be more symbiotic because humans would get to be using our full capabilities right if you imagine you know this is this is how like brett victor could could talk about it right like it's more humane it's it's more we're actually able to be up and about and not sitting at little at our desk looking at little squares but with other people and that's how humans are supposed to be is we're social animals and you know all of those sorts of things so yes i do think that lick maybe didn't quite have that in mind but you can kind of extend his thinking here of symbiosis to to apply to that as well yeah yeah so then we get to memory we started at computation we're now down to like memory hardware requirements and uh again i have to laugh here uh when we start

to think of storing any appreciable fraction of a technical literature in computer memory we run into billions of bits whoa and enough things change markably billions of dollars um which is good to know that i'm a billionaire yeah i have billions of bits here that i can sell and uh on on the black market and make billions of dollars i and like is it so it like there's is i and so there's definitely not a linear relationship between the cost of memory and the size of that memory even at that time like there's no way it would be like a dollar per bit like he's got to be doing some amount of like oh what would it cost to you know manage to store billions of bits like there's that's um i don't know i just like the thought of like one dollar per bit in terms of storage is just like staggering to me like i'm

sure at some time in history that was a very you know a realistic thing but uh by the time we have computers that are operating at the kilohertz scale and considering these are 1960 dollars also these aren't 2022 dollars so yeah this like that that passage is is amazing and um yeah yeah it hasn't aged the best yes and and you know okay so my bias here is there's a lot of talk here about memory kind of on the assumption that memory is going to be really like difficult yeah and so we have to be really clever yeah um you know and he talks about the the try or technically it's tree but t-r-i-e for retrieval yeah i like to call it try anyways i know it's tree but broken like a closureist i mean tree means multiple things try is clear like a closureist yeah i mean you know this data structure does

become important right like it's used for lots of things but he kind of acts like this is going to be the way like all of or at least a possible way like all of our memory would be stored and then there's this like division between like well we might have some memory that's immutable, and that's like the big bulk memory. And anyways, I think that maybe, you know, if you think there's something really interesting in here, I do think that this is like one of the places where I think it was important for his time, right? Because he has to like justify the possibility here. But it just didn't, you know, live. It didn't you know live it didn't it didn't uh age well really you don't

think immutable memory and having some things that are marked as indelible and some things that are marked as as published um you don't think that as a closurist you don't think that that's uh that has relevance to this day interesting interesting interesting yeah so uh i think that that we can kind of just skip over that yeah uh oh yeah and here's where he talks about the different technologies core thin film or even tape memory so yes tape memory did exist at the time though he listed it last in the list and said or even tape memory who can you know perish the thought um yeah i this is uh this yeah this section is just it's here to appease the people who are gonna object on the basis of of contemporary technological limits and that's

clearly not relevant to the to the main arc of this work and then we get into uh his next section is the language problem thank you for uh calling that out, because otherwise, yeah, because he does two sections in a row on memory. So, yeah, the language problem. This is more, this one's more interesting. Yes, let's talk about this one. Yeah, so the big thing that he wants to call out is that there's this dissimilarity between human languages and computer languages, and that this is the most serious obstacle for true symbiosis yeah true to this day yes yeah yes i think he's exactly right and i guess this is where my reading my pessimistic reading to keep calling it that uh also uh comes in because we we haven't

solved this like at all yeah like siri is not a serious solution to solving the language problem uh nor is sequel um nor is any of the the attempts we've had of like yeah yeah you can kind of do natural language-ish stuff yeah and just to just to like describe what this section talks about because i don't think we did. We just said the language problem like it's, it's going to be very familiar. My only note for this whole section is that it slaps. It's basically, you know, if computers and people are supposed to have the symbiotic relationship, they need a way to communicate with one another. How might we do that? it looks at some emerging ideas called programming languages like fortran and algol and um ipl yeah ipl um and and how they are uh sort of a promising direction but not um not a uh full realization um and and and subsequent sections are going to get into this more and more interesting ways but yeah that there's just this this sort of limitation in the bandwidth of communication between the human and the computer and some ways of of navigating that um it also references

this idea of telling the computer what to do through the specification of goals and the way that computer algorithms might be able to work with those goals doing something like hill climbing or self-organization and it looks at those two things and sort of says oh hey yeah they're both kind of promising hill climbing has been researched a fair bit and show some promise and has some implications and then self-organizing programs which feels a little bit ooe to me maybe um has not been uh explored enough but could could mean some some really powerful things but it's it's a pretty short section that just basically holds true to this day saying hey you know this is important computers and people need to be able to communicate

with one another and um and we still haven't cracked that nut and then he's got a little dig making fun of los angeles for being covered in smog uh yes yeah which also holds true today yeah so this this section aged very well and that i i i think he's pointing out problems that you know we still have yeah um and even kind of admitting that there's not great solutions yeah right um and i think that that's very accurate and then the next section i think also i think aged very well yeah um this input output equipment so um i'll just i'll go to where i think like kind of the meat of this whole thing is to me is is nowhere to my knowledge however is there anything approaching the flexibility and convenience of pencil and doodle pad or chalk and blackboard used by people in technical discussion so he says there's all these input and output

devices we're coming up with lots of things but uh they're they're not as good as writing stuff down that stabs me right in the heart with a number two pencil um because i like i i'm sure by the time this episode comes out the the work i've been doing with ink and switch will be published um that's something i've been thinking a lot a lot a lot about is the relationship between pencil and doodle pad or chalk and blackboard there you go ken perlin um and and what we have as the equivalents or the the simulations or the the the computer things that borrow from that tradition of pencil and paper um and just like i've i've spent quite a few hours writing down exhaustive explanations of how pencil and paper are so much richer than stylus and touchscreen or stylus and tablet. And it's something that it's just a real bugbear of mine.

It's a real hobby horse. we could go so much deeper into this but things like at the time and even to this day like working with a a tablet and a stylus interface like you have that tablet and stylus kind of designed together and prescribed for one another in a very narrow way and sure like with resistive touch screens you could use any kind of firm thing that could press in and make contact but now with capacitive um and with the the digitizers used for styluses it's so much more narrow and and traditionally where the stylus was on a wire and was sort of like the light gun on the nest or something like that with a you know some way of figuring out where it was using a camera or some other mechanism like it's even more limited where with actual paper and pencil like you have

just in this sort of way that brett victor references a lot like you have the benefit of physics you have the benefit of i can get any one of a hundred different pencils and any one of a hundred different kinds of paper and spread them all out around me in space and and have you know multiples of them and mix and match them however i want and there's this like just all these beautiful emergent properties from them that nothing in the computer has that yet there's nothing in in the computer that is so compositional as pencil and paper and we have so much theory and we have so much you know development like we we praise things like open doc and and other attempts at interoperability and open standards and like the you know the glorious

era of the late 2000s where it was like apis are going to be the the revolutionary thing and every web service should talk to every other web service and if ta-ta-ta and all of these you know wonderful uh promises being fulfilled about the computer being so general and so open to computation and yet it's like we have never even cracked um or even like put a chink in the shell of the nut of pencil and paper when it comes to the the richness that you can get from an interface with a tool for thinking like it's just it's it's not even close and so that like that little mention of pencil and doodle pad like it's it's a little almost throwaway line in the paper but it just it's it's profoundly resonant to me and so i am yeah i appreciate that how many years has been 62 years later

we're still we're still asking ourselves these same questions. Yeah, and I think he points to, in my opinion, I think you captured that beautifully. There is so much that we can't do, and I think he almost doesn't want to achieve that. He at least doesn't think that's achievable. And so he kind of gives this, I think, almost like a trade-off. Like, yes, we can't quite reach that flexibility in the same way,

but what could we trade off for that? And so I think this is where he talks about being able to sketch out the format of a table roughly and let the computer shape it up with precision. In this section, desk surface display and control, the bottom here. They could correct the computer's data, instruct the machine via flow diagrams, and in general, interact with it very much as they would another engineer, except the other engineer would be a precise draftsman, a lightning calculator, a mnemonic wizard, and many other valuable partners all in one. Yeah, that rings true. Like that feels like, you know, Sketchpad was happening around the same time. And this idea of, you know, flow-based programming is,

I know there are people on Twitter who talk a lot about visual programming. I don't know who those people are, but those people are out there. Like this is an area where, yeah, once you narrow that interface and you set up some reasonable expectations for like okay you know computers can't emulate the richness of physical interfaces that we're used to for other thinking tools like pen and paper but what interfaces could they offer that might be nice to think through and in what way would that allow the relationship to be and it's it's interesting that and this might just be a framing device because of the the fact that licky is um i think he's doing this this publishing this uh this paper here for the air force i think so it might be that that is the focus but it's interesting that the that the designs for the interface are very much engineering focused and it's very much

about that domain and and doing the sort of work that would be useful there and that that dictates the interface and it makes me wonder you know if if he was being funded by you know the national arts center or something like that uh would there have been a very different interpretation of the interface that would be desired and what that might look like, what that articulation might be. And, you know, another one of the accidents of history that most of technology has been invented by people who are of the engineering persuasion rather than of other persuasions.

And there's an alternate reality there that might be very interesting to explore if we ever get that technology. Yeah, I totally agree that it would, it is interesting just to see how much of this is colored by the military context he's in yeah i would just say i think that we have we have nowhere near gotten to computers being able to do this really yeah like i'm not trying i'm trying to be as optimistic as i can okay i really am uh-huh i just i could interact with them very much in the way I would another engineer. My means of communication, even trying to be as charitable as I can be with another engineer,

are so vastly different from how I have to communicate a flow diagram with a computer. Well, but there's the second half of that sentence, which is except. So we're now narrowing the space. So we start with another engineer, except that the other engineer in scare quotes would be a precise drafts person, a lightning calculator, a mnemnemonic wizard and many other valuable partners so it's sort of it's it's narrowing the expectation a little bit and saying that this other engineer has these specializations and they just so happen to be the specializations that involve calculation memory um and and precision and i think that that's fair like autocad is that thing like when you're doing some sort of architectural work or design work in autocad like it's taking care of making sure you always know exactly what measurements you have like it's doing the drafts person work

as like in my regular job i work a lot with schematic diagrams and i'm very familiar with the let's just say the variety of schematics that one encounters um and some of them definitely feel like they were made with software that is doing a very very good job being consistent and and offering aid and other ones definitely feel like they were made by somebody using uh like a you know just a general vector graphic tool instead of something that is more domain specific and you can really feel the difference like i i can imagine somebody in the 1960s looking at the variety of blueprints and plans and diagrams that would be created by people at drafting tables well-meaning people i'm sure but there'd be variability and imprecision and error introduced

into that and it could be very appealing to imagine having a tool like autodesk or autocad to do that that work in a more technical precise way because that's that's something that's chiefly valued in this context so i don't see it as like lick necessarily imagining that the computer is going to be like some smart assistant like siri or something like that or like what siri purports to be um but i i totally see like gooey software something that would have been very foreign at the time very rare very unusual if you know something that lick had much interaction with at all i'm sure a little bit by this point um oh 1960 i keep thinking 1970 yeah so there there we go yeah 1960 i don't know if he would

have seen much gooey software yet no no definitely yeah yeah i've been thinking 1970 this whole time yeah so but like like the idea of software like autocad i think would have he would have looked at that and said yes that's exactly what i have in mind yeah see i guess i read the accept a little differently than you do i don't see it as narrowing i see i think you could write this sentence as in general interact with it very much the way you would another engineer, except they're better in every regard. Yeah. Right?

And many other valuable partners all in one. It's like, yeah, you can interact with the computer, and it's a whole team of engineers for you. Yeah. I mean, it is a whole team of draftspeople, right? It is that kind of an accelerant like there's there's definitely a tremendous amount of that sort of dull preparatory work that it was talking about earlier where it was saying like you know i spent most of my time preparing figures and crunching data and plotting things like that's that's gone that is handled

here that is taken care of and not only that like auto i keep calling it auto auto desk auto cad and other and like solid works and these other sort of um of um like architectural and engineering design tools like do you actually have pretty good capability to suggest oh hey you are doing this in this way and that's going to cause these problems like like they there there are the tools that can do material analysis and say like hey you're making this out of this material but because of the you know the sheer resistance of this material or it's you know it's ductility or whatever this shape that you've made or this tool path that you've drawn is going to lead to a break at this

spot because we can physically model how that material behaves in the shape that you've given and there's like that work is what a brilliant engineer used to be able to do and can now be in a large part simulated or or automated or handled through software okay Okay, so maybe, I honestly have not done much CAD, and maybe this is just true of CAD, although I would maybe push back even there and say, is the CAD being a precise drafts person, or is it a precise tool for drafting? It's not like I can tell the computer,

hey, please go draft this idea for me, and I come back eight hours later, and it's done. I have to be actively directing it. Yes. Which, if I'm talking with another engineer, I don't have to be. I can interact briefly, walk away, come back, work has been done. Maybe not exactly what I was looking for. I can correct, I can, you know, et you know etc right that to me is the difference um but to even like make this not like you know maybe again i just have this bias because i'm doing you know programming stuff but i'll try to make it

like not that like you know my wife is a horse trainer and if you talk to another horse person and even if we like ignore like doing the effects in the physical world and just we're like oh there's a horse that's colicking you know what should we do i don't know a computer program that would be able to give me the answer yeah at least not a computer program where the the intelligence of the computer or the like the magical spell that we've cast with software to make it appear intelligent is what is giving you the answer.

There might be information retrieval. You might have a medical database you can consult, but that is only a slight accelerant on having a library of books and you go and you find the right book and you look for C and then you find the calling. Yes, exactly. The intelligence there

is me yes like i'm the one going and doing it it's just the information retrieval or the intelligences are you doing the look up the person who wrote the original material placing it into this information repository and then thirdly the person who constructed the information repository the programmer and like the the computer in there is just a kind of a facilitator between those three parties. In the case of something like CAD, I think it's just a little bit nicer than that in that it's not like there's a person at Autodesk. There, I got it right that time. saying oh if the you know if you're using aluminum and you make it less than you know five millimeters uh in diameter and it's like a tube sort of shape then it's going to have this problem and if you make it less than four millimeters like it's not like they are hard coding all of the different

circumstances for all of the different shapes there's some there's some generality there there's some generativity it is um it's where the role of simulation comes in and that's something that is not talked about at all in this paper unless i missed it or i'm forgetting but simulation is a really big aspect of what gives computers this sort of human-like ability to respond to novel circumstances in a in an intelligent seeming way and that's i think what enables cad to be you know effective in this example where you could easily find other domains other examples where it totally fails in this regard and in the way that lick is describing so yeah it's like i think this is very much an eye of the beholder kind of

thing for this one yeah to make your art examples, right? If you tried to get the computer to do the same thing that you can do with CAD, but for art, I do feel like it's lacking. It can't be like, oh, that imagery will evoke certain emotions that you might not want in this culture. Well, you say that, but there is a lot of and it's like the example with music where it's like pick a part of music that you don't want to care about and there will be software out there to automate that thing for you it's not good but it exists um the same thing with video like there's um a lot of tools out there apple's really big on this where you just drop a bunch of footage on it and it automatically edits together some kind of you know little film that has careful timing along with music so that beats in the video match up with beats in the music and not just like cuts

but actual moments in the content of the video like they are doing some yeah um computer vision kind of stuff to figure out like oh this is the moment where the person jumping off of the video like they are doing some yeah um computer vision kind of stuff to figure out like oh this is the moment where the person jumping off of the diving board landed in the water and there's a big splash so that would line up with a beat in the music like it can figure that kind of stuff out but it's not good like it's it's it's yeah it's not like you hired an editor no and that makes me wonder like is, is there an inherent thing here? Or is this just a matter of time and eventually it'll pass beyond the threshold of, of, you know, perceivability where, you know, all but the most art damaged among us won't be able to tell the difference between the work of some, you know, very competent director or editor or what have you, and the work of some you know very competent director or editor or what have you and the work of some ai powered video editing tool we'll see that's a whole other debate yeah i mean i think it's i think in some ways it's like uh you know the history of photography right like

yeah just be it's not as if photography somehow ruined, you know, uh, art, right. I know that some people predicted it would cause like, Oh, we're just trying to, uh, represent the world. And you know, it gives rise to other things.

So I do think that you'll, we'll see it, it continue to improve, but that won't, uh, that won't be the end, right. That won't be the end in itself.

And I would say it's not that photography ruined art. It's that like everything ruined art. Like art is so, so destroyed at this point. Like pinning the blame just on photography is a little too cute. Like there's just, there's no hope for art. Art is a lost cause. This is going to be an ad read, but I'm going to start off on a different foot than normal. So one of the things that I found interesting about the Swift programming language is that when it was first introduced, it was described as something

that should scale as small as what you'd expect from a scripting language. So it's something, you know, if you just want to write a really, really simple script that you can run on the command line, it should work really, really nicely for that. That was a design goal. And so it had a really plain, easy to work with syntax and that supports being used in a scripting context where you don't want a lot of boilerplate. But it was also intended to scale all the way up to be a systems language in the large, like something you'd use to write an operating system. And so it's supposed to be very memory safe, and it's supposed to be, you know, very predictable and performant and have all these other attributes that are desirable for that kind of use case. And that's an interesting way to approach the design of something is to say, I want it to be very quick and easy and

pleasing in the small, but I want it to be very dependable and performant in the large. And somewhere else where I see that, that sort of dual set of design goals is with today's final sponsor, Replit. They're an online repl, that's what it says on the tin. You can load them up, pick your programming language of choice, including Swift, and start coding whatever you want to code without having to install dependencies, without having to spin up a dev environment. They just let you get started right away. And so in that sense, they're nicely optimizing for that use in the small. If you want to just kick the tires on a language, if you're, you know, looking at a bunch of different languages and considering which one

to use, and you want to try them and compare, they're great for that because you can just spin up each little language's preferred environment, do some coding, see how it feels, and then switch to the next one. And you don't have to worry about cleaning up anything that you might have had to install on a much bigger project, if you know what language you want to use, and you're just looking for a nice tool set to work with that language, that's something that Replit have been working very hard to make into a good experience. And this doesn't just include nice features that you want for more serious projects like GitHub integration or their multiplayer editor, which is great for getting a couple of people together in the same REPL, and you can have that collaborative editing experience that we all love about online tools. It also means that they're working on features that benefit different domains where you might want to use their tools, where you might want to do some coding in a specific domain and build something there. So for example, they have a game engine that they've made called Kaboom,

which is neat. And you should go check that out if you haven't seen it, because it does a great job of introducing itself. It's one of those things that I like to look for is, hey, when does a project have a fun way to let you know what it is and what it does and what it's like? What's its character? What's its personality? And I think Kaboom nails that. And that's something that Replit seems really intent on doing, is looking at what are all of the different pieces that you'd expect to have if you're going to work on a project, whether it's a little tiny project or a big project, and making sure that those pieces are in place. And so if that's something that you're interested in is building one of these systems that can work very well in the small and very well in the large and solving all of the interesting design consequences of spanning that whole spectrum, you should go work at Replit. You can go to replit.com slash jobs to see what positions they have open.

And once again, I'd like to thank Replit for helping bring us the future of coding. So I'd say, so we've been talking about this for a while. I think that we've got this computer posted wall display, which I don't think today is anything. Yeah, no, it's more like memory talk kind of thing. Yeah. And then we got automatic speech recognition, which I think is pretty straightforward that we want to do this, right?

Yeah. But my favorite part in here, and this is just one of those, again, like old-fashioned-y things that I have to point out. Like, you know, why would we want automatic speech recognition is what Lick asks. And he says, in large part, the interest stems from the realization that one can hardly take a military commander or a corporate president away from his work to teach him to type. And I intentionally gender here because it was very gendered at the time right um and if computing machines are ever to be used directly by top level decision makers it may be worthwhile to provide communication via the most natural means

even at considerable cost yeah this is amazing this is so good uh and i'll give you um i'll give you first dibs it at uh ripping this apart what about this uh stuck out for you i mean so first off of course you know at this time we have these very gendered notions that like to type is you know feminine yes right like this is what secretaries did but if you're're not going to be a woman typing, then you should be a computer specialist is what he later says down here. And I think this is just so not in step with everything else in this essay that it almost is dystopian to me like okay like he says here it it seems reasonable therefore for computer specialists to be the ones who interact directly with computers in business offices yeah so just imagine that he's right and that human computer symbiosis will usher in the most creative and intellectually intellectual period of human history but only for the chosen

few yeah no is this the priestly class yeah right of computer specialists and then they're beholden to the kings of corporate ceos it just it just strikes me as such so out of step with everything else that's been said that it's just laughable well and it came true though like that's computer lib dream machines in a nut is this this this order of computer priests and their uh you know their their exclusivity and how they they keep other people away from the computer and they're this protected class and that the computer deserves to be set free of their controlling influence. And I just, I, it's, it's painful to think about and it's awkward and it's out of sync with the rest of the essay, but it's once again, it's something where reading it, it, it rang true to me in, in, at least in my interpretation.

Yeah. And I think that, you know, obviously today think that obviously today corporate CEOs know how to type and so do military commanders. But there was this real backlash and I think that reading these old papers and we've mentioned gender quite a lot in this, it is I think an important reminder that the social situation that we're coming out of when computers are being created

really structured what we have today. It's not as if all of these biases they had in 1960 just evaporated and became completely neutral in the technology they created. Yeah, exactly. They would have been a present reality in the early days of deciding what it is that the computer should be and how it should be formed. Exactly. And that still impacts today. And so this is one of the things where like, I think it's, I think it's important to look back at these old papers and kind of, they offer like a very grand

vision that I think is really attractive, but also to see like some of those biases and and things that you know also shaped them negatively yep and this is where this paper if this paper i think you know ends in in this section right about like uh doing automatic speech recognition and in fact says that probably five years from now we can have a moderate for like 2 000 words um speech recognizer yeah uh but of course it says like without an accent yeah um and you know like all of these like caveats right yeah um clear speaking no accent and only 1 000 common english words and 1 000 technical terms yes yeah and i just love without unusual accent as if you know there is uh such a thing as a usual accent yes exactly as if there's the correct uh accent to have yeah for more on this see uh see

the recent episode on uh about accents on the podcast robot or not yeah uh but you know it just kind of ends here and if it were i feel like if this weren't a a memo for the military and this were like a manifesto yeah this would be a very strange place to end it there's no recapitulation there's no big conclusion it's just like oh yep and now here's my budget proposal so he doesn't specifically say like please give me money but he mentions all these private corporations and how they would totally take on this work and be able to achieve it. What an anticlimactic ending. Like for a piece that is otherwise so good at forecasting where we have come in the 60 years since, it's just baffling to me that it ends with such a whimper, like with such a yeah on on such a weird branch of the tree it doesn't return to the trunk it doesn't have any summary statement like i guess there is a summary

at the beginning that probably actually would have made for a better ending and it just it falls out there so that to me that was the thing i i was meant i was uh alluding to earlier when i said like yeah i have some feelings about the structure of this piece like to me nothing more than than the fact that it ends literally the last sentence is just saying achieve practically significant speech recognition down to perhaps five years the five years just mentioned it's like that's it that's such a strange thing yeah that's not important but yeah and no it is it is just really interesting that that's how it ends and i think it you know it does show its context right like yeah what does he want to leave people on is and here's the research program we can start it makes me wonder if it like that that this piece of writing is so good at in my opinion as an optimist

nailing the next 60 years of the advancement of technology if not in the right order and on the right timeline is like a consequence of the fact that it is a very you know let's roll up our sleeves and get down to brass tacks kind of like serious proposal meant for stakeholders, as opposed to some grand visionary manifesto that's supposed to, you know, inspire and, and delight other researchers and paint this, this beautiful picture of what the computer what sort of, you know, utopian future the computer might usher in, like the fact that this is kind of structurally forced to be very focused and very um you know attainable might be what makes it have lasting resonance whereas if it was grander it wouldn't it would be too science fictiony too

imaginative too given to speculation yeah i i do think that this is a model for despite like you know it some of it not lasting the test of time in terms of like predictions allegedly yeah well you know memory costing billions of dollars right we can agree on that one but uh but like it is you're right i do think that uh even if i don't think we've gotten there yeah the course he he lays out is very much the course that was taken yeah and i do think that this is a model for a really good paper about the future of coding right about the future of computation where it sets out the grand vision but then really addresses the current issues of the day. I have to wonder if inventing on principle

was followed up with a more, and here's how you achieve in your languages, in your runtimes, whatever, this way of doing direct feedback, even if it's just sketched out in the manner that this is yeah not you know in full detail if that wouldn't have been if we wouldn't have seen more things than we even saw i think it was i think drawing dynamic visualizations which came after inventing on principle and also um uh stop drawing dead fish also came after inventing on principle i think both of them do that i just i think that they for whatever reason didn't catch on among the programming crowd perhaps because inventing

on principles vagueness and the plurality of different examples that it offered allowed people to kind of project something onto it that maybe brett didn't intend because i know like not to not to get all kremlinology about this but brett the purpose of inventing on principle wasn't to say that immediate feedback and direct connection with the tools are the principles that we should adopt it was meant to say those are brett's principles and that everyone should find their own set of principle principles to motivate their work and that here is what it looks like to have a rich set of principles and to follow them in the direction that they lead but that like that sort

of fuzziness in his message meant that i don't know a lot of people i think got the wrong impression of it and took the wrong things away from his work so that when he followed up with you know like drawing dynamic visualizations is a very realistic tool like somebody could conceivably build that and yes it does have shortcomings and and um pitfalls and unexplored areas but it's it's definitely something that we could have today and some people have tried to build that like there's apparatus by toby shockman previous guest on the podcast there's charticulator which is for power bi for generating dynamic charts in a very sort of brett victory kind of way there are some other examples out there of tools that have done this to

a good degree of maturity but aside from those handful of examples which are like literally implementations of the same thing that drawing dynamic visualizations was about like they're just there for whatever reason i think nobody has done the hard work of actually picking a domain and saying let's let's you know figure out the principles that the the victorian principles and actually pursuing them to fruition so i i want to say i completely agree with you on the the point of brett victor's uh talk and it's something that i really did take to heart that like finding a principle is what is more important than adopting brett's yeah right like i've wanted to write an essay on this for a while because I do think that that's the thing that is often missed.

But drawing dead fish, which I do enjoy, and then whatever the other one was. Stop drawing dynamic visualizations. Stop drawing dead fish and then stop drawing dynamic visualizations. They don't reinforce the point that you should find your own principle. They're living out of Brett's principle. I do get that. But you're right, there are tools that did go off and do those,

and I do think that's in part because he followed up with the dynamic visualizations, etc. He went a little bit deeper. But I guess what I kind of meant was, yes, you're right that his point was not, you should adopt my principle. But the programming tools or the anything, yeah, the programming

tools is one of the things I was kind of maybe focusing on, which yeah, admittedly, I have a bias towards those. But he wasn't presenting, here is a way to realize this yeah right from my understanding you know it was it was very much demo wear right it's not like he went and built a system that could generally do this yeah or he wasn't he didn't write a thing or or present a

work that is like here is my framing of the problem that needs to be solved and here is the relationship between that problem and the state of the art and the upcoming you know reasonably expected future of technology the way lick did in this paper that's not something brett did yeah whereas magic ink i feel like is much more in the vein of what Lick did here. Magic Ink is very much a, here is the idea behind what I think interaction design should be, meaning it shouldn't be all about interaction, and here are some case studies and examples of how you can do that. Yeah.

And I do think that while it's maybe not as wide-ranging as Inventicon Principle, it definitely has been very influential. Yeah. But it's's much narrower than i think lick's thing is here like there's definitely a kind of a goldie locks going on here where this paper from lick like it does this beautiful balancing act between a truly grand vision a truly staggering um idea of what will come in you know 20 or 30 years after this is published to some extent with the pc revolution and the gooey going mainstream um and you know squaring that off against you know what is the first step that we can take towards that vision? And which step is the step in the right direction? And that's, that's a really interesting way to present a vision is to say, here is the distant landmark that we're going to walk towards and here is you know here's a because it's certainly not an outline of the full field of technology it's just an outline of the things that are relevant to achieving that vision so i would i would want to uh you know just because we both kind of had different takes i completely agree that the things that that lick has said here a lot of them has come true he was

very before his time he put this vision out. And partially it happened because he was also in charge of funding people who made it happen. The best way to predict the future is to invent it, yes. There's no question that that's true as well. But what I'm curious about is, do you

personally feel that your relationship with the computer is one of symbiosis. Yes, I do. I have no hesitation in saying that. Okay, interesting. And I don't think that, because there's like the different kinds of symbiosis as well, right? Like there's mutualism, there's parasitism, and there's the other one that I can't remember.

And I don't think that it is a purely positive symbiosis, but I do definitely think that it is a purely positive symbiosis but i do definitely think that in so much as i have the ability to use pencil and paper to extend my mind into the physical world and that i have like if you came into my office and looked around and i'm sounding different on the microphone because i'm kind of looking around my office as i say this like if you looked around the space is a reflection of the kind of person i am and the thinking i do to the extent that if you took me out of my office and put me somewhere else my existence as a human being is different like that's one of the reasons why it's nice to travel that's one of the reasons why it's nice to um have an office is because it it's like being a like a transformer or some kind of robot or

something like that and being able to swap like different tool arms on or different you know augmentations of your person it's like the the japanese school of of science fiction character design where uh samus in the metroid games and like uh whatever the guy's name is in megaman like they have an arm cannon because the cannon being their arm represents this idea that the power that you have is something that is within you it is your own power that enables you to overcome adversity whereas in american science fiction it's so often the thing that gives you power is a gun and that is a thing that can be taken away it's a thing that you have not a thing that you are and to me like being with a computer that's that extension of my person that's that you know that um fulfillment of of my existence

in a particular direction and i think that that that is a symbiotic relationship and and maybe the part of the symbiosis that is arguable is in what way is the computer benefited by being with me um i'll certainly tell you my computers are better taken care of than computers by other family members of mine that tend to be over ridden with garbage all over the desktop and and covered with dirt and whatever else not to throw anybody under the bus um but yeah that's not a meaningful benefit so they're in that sense i don't know but yeah like from my perspective i do think it's symbiotic what about you how do you feel about your relationship with the computer is it symbiotic is it something that um that you feel similarly or differently about?

Yeah, I mean, this is one of the things that I think really confronted me with this paper here was I relate to the computer a lot, right? Like I am on my computer a good amount, you know, I'm a programmer. And I kind of like, got into programming at a very form, very formative years in my life, like I kind of discovered programming at an a very formative years in my life. I kind of discovered programming at an early age. And so in a lot of ways, I wanted to think that this is a symbiotic relationship. I do get a lot of, honestly, solace in computers. I feel comfortable when I'm on computers in a way that I never did before I kind of discovered them. But as I look at it, I feel like I almost am always in one of two modes, meaning the mechanically extended mode,

where like the computer is a tool for me to accomplish some goal. Even if that's like a very personal goal, it's still a tool towards that. Or I'm being, you know, there's the humanly extended machine that I am being, the computer is using me, right? And not, of course, the computer itself, but the systems that I'm in. So my corporate job, right? Where through the computer,

I'm fulfilling some role for the larger system. And I guess where I find the lack of symbiosis, like why I think I'm in one of these two modes, is that I am always, either the system's the impetus or I'm the impetus for all of the action. And I never like come back to the computer surprised. I never come back to the computer and it's done something for me in a way that I didn't expect before. It's never feels like there's another organism there. It's always either I'm imposing on it or a system is imposing on me through the computer if that makes sense it perfectly does but i don't know that the larva that lives in the fig tree would feel that the

fig tree is another organism well yeah but that's only because i don't know if larva are conscious well as a larva as somebody who is pupating like my dog even feels more like a more of a symbiotic relationship than than my computer yeah right yeah definitely like i get a lot out of my relationship with you know my tiny little dog even though you know she's not providing for me or you know. Like there's just something there that feels a lot more two-sided. I mean, I just think that I would love for it to feel like a symbiotic relationship and maybe some of it is not even the computer's fault. It's how I choose to use it. Right. Like I could even admit that. that uh but i i guess i just don't i feel like if there are uh computer human symbiosis symbiosis today i think they're pretty rare i think most people would not feel that sort of symbiotic relationship yeah and i guess that's going to depend on how each person

defines symbiosis and and what it means to them what does symbiosis mean to you the uh another angle that i wanted to take on this that you may or may not find permissible is that the ways in which a computer does surprise me or the things that it does that i wasn't expecting or you know when i come back to it and it's presenting me with something delightful that I didn't know it was going to present me with. Those are things that originate in some sense, in some interpretation, from something another person has done, but they're mediated by the computer, and that the computer enables me to have a relationship with that other person in a way that I wouldn't have had otherwise. And that's not suggesting that i have a symbiotic relationship with the computer but rather that the computer is able to

have some of those properties that we're talking about you know emerging from symbiosis um but they they originate from other people so once again you, games where a game is designed by a person. I've had experiences with games like I played Inscription last Christmas, and it is just loaded with surprise after surprise. And sure, those surprises are all designed in a way where everybody who plays the game is going to have a similar experience with it but the the person who created it i can't remember their name the computer was essential for them to be able to arrive at a system that is capable of creating those feelings within me and the other people who play this game and it it just affords a really different kind of relationship one that is a little bit more dynamic than what you get out of you know being able to write a book or direct a film or act in a play or do any of these other kinds of expression or you know draw a diagram for an architectural project or something like that or you know calculate the ballistics needed to launch a shell into another ship across the atlantic ocean or whatever it is so it's it's the kind of thing where maybe

it's not symbiotic in a in a in a way that is in keeping with the biological use of that term but i think there's certainly something here that is interesting and novel and different and that has changed a lot over the last 60 years and especially the last 30 or 20 or wherever you want to draw the line either the pc or the internet or social media or what have you like there's there's definitely something here that and maybe it is the you know the the humanly humanly extended machine to some extent but there's there's something there that is new and different about the computer and maybe symbiosis is just the wrong framing device but if you took that exact framing away from this paper and said you know here's here's some framing that we don't

have a word for that's going to describe the relationship between the person and the computer and here's how that that framing is going to be enabled by this technology like i think that holds i think that that's that's something that um exists now that didn't exist before yeah i i definitely agree that this communication aspect is something i don't actually think is talked about like at all here right that's this like social aspect and connecting through to other people through the computer the computer more as a medium than an organism so i definitely would agree that that is a very important aspect of our of our you know real existence but i guess i just don't see lick you know quite going there yeah but i mean lick doesn't have the benefit of having you know yeah grown

up with diary land or whatever it is and in fact that's a good example like like like uh online journaling you know the the early social networking like services like deviant art was a was one for me because i was an artist as a teenager and that that is for all intents and purposes a social network it's just centered around uploading and sharing artwork that you make. Those things are more like books or magazines or correspondence through letters or phone calls or other things where they are like technology facilitating a direct communication from one person to another person see also our previous unreleased episode about uh long distance communication um and that's not even what i'm thinking like i'm thinking i'm thinking something much deeper and more profound man um i'm thinking like like the ability for the computer to be programmed and to have some dynamic behavior and to have some not even like unpredictability or uncertainty but to have like to be more like a partner or to be more like like even a even a you know a petty simple dim-witted simulation of something vaguely resembling a person like to be a puppet

even like there's there's something i think profoundly different about that than from any other previous mechanism for quote-unquote communication no i i get where you're coming from i mean video games aren't exactly communication in the way that you know i was maybe cashing things out yeah i guess what what i think what i would need right like you know you you already feel that you have that symbiotic relationship. I think if I were going to, I do think video games are beautiful and wonderful, but you're right.

Like what they give us in their surprises are, you know, also what everyone else is discovering, right? Like anyone who plays that game for, you know, yes, I know that there's randomness, there's et cetera, but for the most part, anyone who plays that game will have a similar experience anyone who uh you know consumes computer specific media will have the similar experience the applications we have are all kind of designed around the common case and i feel like if if we're going to have a symbiosis, I feel like there has to be something more personal, more customizable, more moldable, whatever word you want.

It needs to be... If I go onto someone else's computer, to me, it should feel different in a way that I don't think that it often does. Really? You don't feel sitting down at somebody else's computer, like picking up somebody else's phone? Not even the physical sensation of touching somebody else's phone which is gross but like like opening somebody's phone and looking at the home screen or whatever app they had open like that that gives me this this intense feeling of of panic and discomfort that I'm sure is exactly the same as the feeling of a larva that lands in

the wrong fig tree. And to continue circling back to the larva and the fig tree, like, isn't the experience of every larva and every fig tree going to be basically the same? Isn't the experience of everybody's romantic relationship going to be so similar? Like, you know, you and your true love meet each other in high school where you go to the same high school um and you know have a have a an infatuation and then an admission of feelings and then a you know an early flirtatious relationship and and you know a deepening appreciation for one another and then a moment of of betrayal or something like that and then heart-wrenching

and the undoing of a relationship and mourning the loss and and finding someone else and rebounding like all of those are universal experiences as part of being a human and even though they are all deeply personal they are relatable they are imminently relatable. There's, you know, like the love song is a thing. The breakup song is a thing. So I don't know that that individuality of experience is the right criterion for determining whether something is a profound relationship in the way that the word symbiosis suggests some kind of profundity which which again makes me maybe think that that symbiosis is a is a word that we are now stretching far past its its technical definition yeah but we're not going off the technical definition we're going off licks definition right yeah like yeah right and he paints it as a utopia like let's just you know well it's this. No, he paints it as computers can help you with formulative thinking instead of formulated problem solving. So it's like. He says that we should be intellectually the most creative and exciting in the history of mankind.

Yeah, that is pretty big. To me, this is the criteria. I guess that's where I would go back to is I don't. I think that, of course, we're very creative and intellectually exciting, but I don't think it's a difference in kind. I think it's just a difference in, we use computers to do all sorts of things,

and yes, maybe we can do it on certain scales, etc. I'm not saying there were things that we could have achieved without computers that we had you know whatever right like we probably couldn't have gone to the moon without computers i'm not i'm not suggesting otherwise um but like i i do think that i wouldn't say that there's a a clear difference and the this being the most creative and exciting period of of history so just on the record yeah jimmy does not think that the current era is the most intellectually stimulating and creative and exciting period of all of human history yes on the record and i do on the record i think that

we are living in the most intellectually stimulating and exciting and invigorating period of history it may be because i'm alive now and don't have a basis for comparison if i was alive in the 1920s or whatever and it's the kind of thing where it's like sure you know you could everybody who imagines being alive in the in the 1700s or whatever imagines themselves in some high court or you know as the you know the patronized composer to some some lord or something like that, a nice cushy job, or maybe apprenticing under Da Vinci or something like that. But of course, most people in that time were not having that experience.

Yeah, to be clear, that's not the sort of thing I have in mind here. I would say if we are in the most creative and exciting time in the history of humankind, it's just because humankind has continued to progress, and it's more about the social systems and the social technologies that we've created more than computer-human symbiosis. I think that we've continued to advance by standing on the shoulders of giants, where giants here are not the great men of history,

but all the people who came before us, right? The masses who came before us, who forged ahead and made changes in our social landscape where we don't have as much injustice before etc right i i don't think it's because i'm really intimately connected with my computer yeah that would be my my stance on that yeah i i have to agree social progress is definitely the the biggest giant shoulders to stand on because it brings us to the present era where if we do actually truly want human computer symbiosis um we'll be able

to achieve that by finally giving computers genders so we can have you know male computer female computer man computer symbiosis achieved thank you social progress um we've made it that okay i'm i hope that that's the note we end on yeah i that was what i was going for i don't think it landed but it's close enough yeah we needed a way out that would have gone forever yeah yeah you're all good and that's all for this time thank you to jimmy miller for joining me on this adventure thanks to theater js glide and Replit for sponsoring the show. And I will see you all in the future. Bye.