Video: New Data Structures (DocDrop)

00:00:00

foreign [Music] ERS linkers and mappers we are really happy to host a series of conversations around the topic of tools for thinking our longer term goal is to spark a

00:00:19

diverse connected shared memory that will help us make important decisions together our near-term goal with these podcasts is to blow more oxygen on the growing tools for thinking sector addressing key issues and talking with the people who are doing the work

00:00:33

this podcast is created by Beta Works a New York city-based startup Studio I'm Jerry mikulski your interlocutor and obsessive mind mapper our topic today is data structures that

00:00:45

improve tools for thinking our desks are Paul Roney founder of cosmic Alex obenauer the founder of mindsets and John undercover a gesture UI Pioneer you may recognize from our third episode on

00:00:57

Space pixels and cognition there was just a moment where it made a lot of sense to invite John into this conversation and he has sort of free license to jump in and so forth but with

00:01:08

with that I would love just uh Paul if you could just tell us a little bit of the story of uh what you're doing and how you got there yeah so nice to meet you all Saint Paul I'm the founder of a cosmic uh Cosmic is

00:01:23

a special web browser or a special tool for thoughts uh based on an infinite canvas where you can drag and drop any kind of data from text to web pages to PDFs and the nice thing about Cosmic is that it's based on database called Foundation that we've also developed

00:01:35

based on ipfs to a decentralized file system and that decentralized system allows us to implement features in the canvas that makes sense like backlinking linking transclusions so the idea that an object can lives in two spaces at the

00:01:48

same time and be edited by several people while maintaining this thing of data and we do think that it was Forest essential to build that database to make the UI shine we see more and more visual

00:02:02

canvas but if we do not change the underlying structure it will be difficult to take them kind of in a more um I would say hard Direction and Beyond the simple whiteboarding so that's kind

00:02:14

of what we do we do the database and the new spatial UI for IP bad for the magnet for the web sounds great thank you Paul um and just ipfs the interplanetary file system by itself is a juicy and

00:02:27

interesting topic and right away there's this sort of tension between sophisticated data structures but distributed data management and so ipfs wins you to the distributed data part but then you still have to solve the

00:02:39

other the other pieces of it um Alex you've been wrestling with these issues as well uh could you could you give us a little Trail for how you got here yeah for sure

00:02:51

um I yeah so my my primary work is uh independent research in the future of personal Computing it's primarily shaped by exploring you know the OS of the future what does that look like and how should that work um I document this work on in my lab

00:03:05

notes on my personal website um it it uh it's primarily looking at trying to figure out um what is the mental model that we have when we're interacting with our digital

00:03:17

things and I it's you know it's not apps apps an app is a is a as a metaphor um but but that doesn't necessarily fit into our mental models very well so what are those mental models with which we

00:03:30

think with which we do personal Computing um and can we put together the the The Primitives uh to interact with um on our on our you know our personal Computing devices

00:03:42

um more fluidly so that we can more closely represent our thinking within our personal Computing domain that kind of thing thanks Alex and and at various points in time I've often thought a little bit

00:03:55

about and I am no information architect or coder but I've often thought about sort of the difference between databases operating systems applications and where does data lie and could we somehow just

00:04:07

treat files sprinkled across the file system as a distributed data set and infer database operations and other kinds of things on them we'll sort of I think we're going to walk through some

00:04:20

of that as we as we open the conversation but I love like the thing that you're focused on um and and John could you give us a little bit of background uh on yourself well that's going to start sounding

00:04:32

redundant because we did that back in episode three but I'm I guess a lifelong bemoaner of the lack of uh of uh advancement in UI UI being that most fundamental of all uh computational

00:04:45

experiences the one piece of the computer we can actually touch see smell you know like that that is computation for us and for a better or much much worse I think the the sort of emergence of the internet but really the web

00:04:58

around 94 distracted on mass the entire CS world to the extent that you know I watched uh grad school get kind of hollowed out as

00:05:09

everyone went off to start a a web-based startup and so forth but more more importantly and more deleteriously uh that that mass diversion meant that people were no longer developing new uis

00:05:22

and it seemed to me that 1984 from the Macintosh UI to 94 a decade was about what we should expect uh a healthy a healthy kind of General UI to last and

00:05:36

then you know the soul is all full of holes and the shoe leather is falling apart and it's time to get a new shoe but instead we got the web and we got stuck with this desktop UI so I think

00:05:47

we're we're sort of 35 years in an ubliet in a in a pit um and I'm hoping to stack up some little Pebbles to help climb our way out I've um I've put examples of some of the ways we might do

00:06:01

that in films that I've been very very very fortunate indeed to be able to participate in in the design of like Minority Report and Iron Man and Eon flux and so forth but most of the time I I spend my life building building

00:06:14

examples and I'll just say that I'm really excited to uh speak with with all of you today uh the podcast and the the kind of growing tide of interest in these topics

00:06:26

um out of which the podcast comes makes me feel a little less lonely I love that thank you John and and you're reminding me that in in episode three where you were our guest um I brought in this notion that the

00:06:40

desktop metaphor has has sort of all sorts of applications in a tractor beam because every application needs users and users are really familiar with the office suite and the desktop metaphor so every interesting application gets

00:06:52

tugged in that way I'm including that by reference here if anybody would like to go deeper into that go live go back and listen to episode three go ahead Paul yeah so no I was just just to rebound on the desktop metaphor one thing that to

00:07:05

me the to me the fundamental problem with that metaphor is not the fact that we try to create a virtual desktop but the problem is that as OS has evolved like as we got from Mac os1 system one

00:07:17

to Mac os9 to micro S10 to where we are right now we diluted the power of that metaphor I'm going to give a few examples that to me are very very uh that users do not see them but they actually make the mantle model way way

00:07:30

worse so for example in system one when you double click on a folder on Mac OS it will open a new window that window has had metadata that would store its size its position on the screen of the computer and if that folder contained a

00:07:44

subfolder and you double clicked on that it would not refresh the content of the window as Mac OS 10 is doing today it would open a new window with its own coordinates its own size and you could create arrangement of Windows specially

00:07:56

so the finder was actually a spatial metaphor it was not an infinite canvas in any way but has that idea of a trail that you could follow and where you could find your files very easily and we lost that with Mac OS and as we go

00:08:09

especially with the new version that came out yesterday it's worse and worse because windows will not even remember to press that like would you like an icon Arrangement a grid Arrangement a list Arrangement or a columnar Arrangements like we completely lost

00:08:22

that and the other thing that goes even before the Macintosh was that the desktop as a metaphor was active so if you look at the work that was done at the park and if you look at the computer like the Dorado the alto more even

00:08:35

clouded on us and commercially available the Xerox star on the desktop you had icons like you had a printer you had an an outbox and an inbox and there is something that is to me very very

00:08:47

significant in the operating Paradigm that we have today today if I want to send you an email I have to open an email client or go to a non-provider type an email and that email will be stuck in that program or I can copy and paste it let's set that aside I'm going

00:09:01

to send send it to you the whereas on the trucks computers you could just open a text editor type your mail it would have an icon on the desktop anywhere just drag and dropping that into the in or out basket and it would get

00:09:13

distributed or you are drag and dropping that on the printer icon and it will get printed out so the desktop was actually functioning not only as a file organizer but it really has a space to create and

00:09:25

to manage work and so we lost that and to me this is the fundamental problem today like we lean too much on the file metaphor on the desktop metaphor as just a file browser like we made the desktop

00:09:37

dumb and um the work we do Revels mainly around that and on how should how could we do away with files like um maybe you want to get me there

00:09:50

because it's already very long but another thing that I'm really really kind of um not please pattern my English sometimes because I'm French so I mean let's use the right to a lot of times but something that I'm kind of pissed off about is how the iPad designing

00:10:02

files like if you look at the iPhone and the iPad they are not app oriented the Mac is app oriented but you can still open a file in different apps actually that's one of the Magic Magical thing about the mac and the Mac will just pick

00:10:15

for you which app is the best translator that's how they were called before Mac OS 10. on the iPad I would call the iPad a container container oriented computer which is that you have apps that contains their data and because it

00:10:27

became so tedious to organize because you had to handle files in each app you had to manage that in in the app itself like the iPad is really app oriented and I would say container oriented because

00:10:38

the files themselves are trapped into the app so Apple came out with the files app in I think it was iPad OS 8 or 11 and the files app is just a number and harder to use finder that has an even

00:10:51

worse metaphor because you cannot even organize it visually and so now the magic of the iPad which should have been to say we hit the file system we have a data pool and apps are just going to talk to that data pool was completely lost

00:11:03

because of that love that so so from what you started saying I should not upgrade to Ventura depends on what you want your computer to look and behave exactly but but you're also describing this Historic

00:11:18

Trail that I hadn't noticed that much and I'm aware of Xerox star but I've never sat down at one and so and so this progression uh and and as you're saying sort of weakening of the metaphors is

00:11:31

super super interesting if I if I may actually I think uh with your permission Paul and maybe this is yeah go ahead you know maybe we can carry on the hockey puck off my face to Alex because I bet he has a bunch to

00:11:45

say about this also but it feels to me like the original sin is even farther back than the dilution of the Macintosh metaphor this the original sin lies in the difference between

00:11:57

uh the small talk environments the small talk GUI environments of Xerox Park and the Macintosh version where in the Macintosh version there's no uh there's no direct efficacy no direct Agency on

00:12:11

the part of the user all you can do is open apps and use the apps in the small talk environment you've got a GUI you can build apps but there's always a code

00:12:23

editor right there there's always a transcript window where you can uh where you can directly manipulate the substance of the machine and that's trapped behind or it's occluded

00:12:35

deliberately behind a layer of abstraction that that the Mac UI gave us and that we've been making we've been making that blast Shield thicker and thicker and thicker which with each passing year until as you say with

00:12:48

Android and iOS they've worked really hard to pretend that files don't even exist and of course they do and you can't get away from it and the underlying os's habit which is why you have to eventually

00:12:59

introduce the files so the the files metaphor the files mental model for what's happening down there is at least for me way way better than the well there aren't really files and it's sort

00:13:12

of magic but the but we don't we're not going to help you understand what the magic is idea that mobile devices try to foist on us Alex the Alex the puck is in your rectangle

00:13:25

yeah I um the original sin I definitely look at that that move that exact spot is the original sin so a lot of I'm constantly designing new things into you

00:13:37

know my fictitious or semi-real os's of the future one of the big ones is non-volatile workspaces you gather together a bunch of different things that you need to do some line of work or you follow some train of thought and it

00:13:50

causes a number of different things to be you know on your screen and you have them arranged in some capacity that's some amount of work you restart your computer that work is completely lost you have to redo it so a non-volatile workspace you know is you know kind of

00:14:04

intuitive at this point right the idea that I should be able to return to a workspace have all my items that I had up arranged the way I had them and it turns out that was a part of the if I understand correctly original uh GUI

00:14:16

desktop idea that you would have kind of multiple I I don't know if they were called multiple desktops at that point but uh that you would you know have different workspaces and you could flip from your kind of office workspace to

00:14:29

your home workspace and it would it would keep all your you know all your windows all your files all your things would be open uh as you left them and and so we lost it every time I feel like I and so whenever I describe my work on my website I say new and renewed ideas

00:14:42

because I constantly find 90 of of what I put of the new thinking that I put in uh to my work it was there and often at Xerox Park so many good ideas were dropped in the

00:14:55

world way back when including Doug engelbart's famous mother of all demos which was very multi-user right and then and then when these things get commercialized they get dumbed down they get broken they get you know both gates

00:15:09

and jobs had to be blackmailed and coerced into group work networking any kind of connectivity among people they both were thinking about personal productivity how did how did all that happen

00:15:22

so I I think there is something about jobs uh um and Gates here so when jobs was kicked out of apple and started next he gave an interview when he said I saw three things at the Xerox Park I saw the

00:15:33

GUI I saw object-oriented programming small talk and then I saw ethernet networking and I was so blinded by the first one that I forgot about the other two so I focused on that and next and so one of the things that it usually like

00:15:45

you can find it on YouTube but one of the demos of Steve Jobs that is not very famous because it's not very glamorous but one of the best one is actually Steve Jobs at next just after the stop commercializing Hardware called interpersonal Computing which describes

00:15:59

exactly that like how a computer is a window towards uh something and that window can be shared with someone else but we've also lost other metaphors like for example one of my main culprits was today multiplayer software like figma

00:16:12

Google which are in themselves very good products is that they completely lean into the what you see is what I see metaphor like you have to be on the same canvas we have to share the same documents that's not how creative work

00:16:26

our work or productivity happens like usually the creative people will kind of work in a security workspace with all their stuff all their resources all the material and then share a different artifact and so one of the metaphors

00:16:38

developed at Park that we also lost what's called what you see is almost what I see which means that we should be able to have several pointers to different uh um places in a file and to

00:16:50

me that problem also comes from the fact that users because of those metaphors of the of those metaphors are completely mixing window files and apps and especially now as we

00:17:03

go towards more and more integrated software like for example figma is completely self-contained into one window you have to pay in the file explorer everything is is inside one window whereas a file should be kind of

00:17:15

like email you should be able to choose an app that would display a file in a different Manner and you as the user should be able to set the parameters that's completely impossible today with the five structures we have we have to

00:17:27

do away with the funds and as we move ahead with more and more cloud storage it's going to to become a larger larger problem I had a wishlist uh software item years ago that I called linkido sort of a

00:17:41

neologistism a blend of link keita of Link management and Aikido and the idea was I it dawned on me one day that I was using eight different text editors and back then it would have it might have been Outlook word PowerPoint blogger uh

00:17:55

something something I don't remember what it was but I counted eight and they each had different kinds of commands and different ways of embedding things and this was before command K was a general way to embed a URL and stuff like that you know they had cut meaning perfectly

00:18:06

different I assume uh exactly and cut copy paste was standard because that had sort of won the day there which is a keyboard Miracle like like the one useful thing that came out of the QWERTY keyboard for me is cut

00:18:19

copy paste on the keyboard you know Maps so nice so nicely to characters that are memorable but in linkido I was like how about if you floated the editor your favorite editor you get to pick if you're an emacs person then it's emacs

00:18:32

if you like something else you float your favorite editor above all apps and the only difference between an email and a web page is that one of them says send this to this person and the other one says put this over here that's it that's

00:18:43

it and if we and if we assume if we assume a lingua Franca like HTML plus plus whatever that might have been back in the day then all these things become interchangeable and then any message any nugget becomes promotable demotable

00:18:56

forwardable copyable it suddenly becomes this really Flex little object in the middle of people trying to do work together and I I just just to complete that I think there is this is where we have to

00:19:08

be very careful as designers because if you look if you look at what you said uh there are two ways to interpret that and the two ways um are actually speaking about the same thing the underlying concept is the document oriented

00:19:21

Paradigm which is that instead of calling an app you are calling something to edit something else like a palette to editor text another palette to edit an image and if you look at what Jeff Raskin was saying if you do that in the

00:19:34

same software you are going to create what he calls modality with a Vengeance because you are going to have one software so one thing that is very kind of Monolithic that is going to act in entirely different ways according to the

00:19:47

contents you click on that's kind of how notion operates with Integrations tables text you create a block and the block will call something to edit it and you kind of know once you uh created a way

00:20:01

around the software like once you've mastered it how it works I think that the best way to do that is to actually say here is the data pool here is the object and you choose the viewer you want and and you have a choice of

00:20:13

several viewers like we do not have to do away with everything we do not have to throw Windows away we do not have to throw apps completely away we have to make apps viewers for an underlying data and to add contextual information in

00:20:26

each viewer just to build on what you said briefly and then I'll throw it to you John um one of the things that's been bubbling up a lot in these conversations around tools for thinking is that there are a variety of different tools for thinking they serve different purposes

00:20:38

like kumu is really great for system flow diagrams and I can't do that in the brain but you can't use kumu to do the stuff that I do in the brain and Rome is good for something else so there's this idea of pivot or there's like liminal spaces between

00:20:52

what different tools are good at that we should be able to sort of seamlessly move between so that we could choose our viewer or application on the same data but as as what we're trying to get done

00:21:04

shifts we can shift the tools and then share that view with somebody else or whatever else exactly I mean I think we're jointly on to something really good here and it's not like we're the first people to to come

00:21:16

upon this but the but the idea of inverting the relationship between content and manipulatory tools where we're very much in manipulatory tools equals app mindset

00:21:30

today but if instead as you you started to propose Jerry and I think this is probably a lot of what Cosmic does but I haven't gotten my Beta uh invitation yet so I don't know but like if we imagine arraying you know arraying all of the

00:21:43

stuff the data the goods in a spatial way then then viewers editors applications whatever I want to call them are more like dirigibles because I think we need to get back to blimps in

00:21:56

everything we do that float Above This landscape if we want to be 2D about it and give you kind of a way to peer down on some region of it the particular characteristic of that model that is

00:22:10

especially exciting to me um because I have very narrow range of interests and a lot of and you know and a Big E Day Fix You is that uh is that in this way that window or application

00:22:23

or editor or whatever it is also becomes a metaphor but also a literal representation for attention and in a world where there's collaboration if uh

00:22:34

if we share some region of of you know landscape there's some overlap between our files our data our documents whatever they are then if I pull back far enough I can see that you know Jerry

00:22:48

has his text plus link Plus image editor kind of hovering above this region which is not of any interest to me at the moment but I can see which way it's drifting but I know that Alex is

00:23:02

actually right on top of a thing that I've been working on and so you know this is one of the things that's really easy for digital systems to do and no one's ever done it commercially which is to depict attention depict human

00:23:15

cognition so I I think just to add to that it goes back to one of the notion that uh also Jefferson advocates for which is that he thinks that we are mixing two different uh ways of the mind

00:23:27

to focus his attention one is focus and the other is Locus and the lockers the locus Locus is how you fix your attention on a geographical point and how a UI should embody that concept and

00:23:40

that's exactly what you just what you're describing like we have to today we can focus like we have full screen apps we have stuff like that but it's it's becoming like today window Management in Macos 10 is I would argue way worse than

00:23:53

in Macos 9 like a lot of the productivity work that we have to do is actually becoming more difficult and and harder to to to to create and to maintain and so I completely agree with that Vision but it comes back to that

00:24:06

idea like we have to destroy the idea of a hierarchical file system even in in terms of how we develop software like it's becoming it's going to become a problem if we want to advance there's also weird thing with applesauce

00:24:20

software in general it's like every version of iMovie is worse than the last one every version of iTunes seems worse than the last one I don't use most Apple software because they seem to just be bouncing down the hill

00:24:31

so I I think it comes from the fact that it's an old debate that Apple that almost killed hype account actually hypocal was almost completely killed by Jean Legacy who has the French guy that

00:24:43

took over the RND as Apple when Steve Jobs got kicked out and he was like if we allow I have a workout to be sold with every Mac we are going to completely frighten every third party developer because people will just make

00:24:55

software on their own MacIntosh and they will create their own software and also hypercar is not consistent with the metaphors of the Mac that's another subject but I think that apple is still kind of circling around that problem of

00:25:07

they have to create software as a first party for their platform for the hardware to kind of showcase what it can do but at the same time they just do not have those resources and they do not want to kind of kill the market that should sprung up

00:25:21

naturally from the App Store and if you look at the ilife and iwalk suit that were distributed uh with the mag during I don't know 2002 uh to until maybe 2009 or 12. they were actually

00:25:32

really good and it kind of went downhill as the App Store took more and more importance did you say that on the other resources I I know I know it sounds funny but uh

00:25:44

from I I've spoken with some people in the thing and they told me that yes they have a lot of problems and they just do not have enough developers to maintain everything they do

00:25:56

which seems very very yeah I think like maybe maybe they have maybe it's a problem of process maybe it's a problem of tools like have you used xcode recently like uh really it's it's hard like we do cosmic the native

00:26:09

app for the mac and the iPad and I'm super proud of that but we have to fight the Frameworks to make it work um well my next question was going to be can you tell us more about Wonder OS and sort of the the explorations you're into

00:26:22

in the context of what we've been saying here we've been talking a lot about um kind of the new models uh that we want to apply to kind of existing systems or

00:26:34

um how we might reify those new models in existing systems how we might fill the liminal spaces how we might have this Universal Prof referred text editor that kind of floats up right um and and but what I've done with my

00:26:47

work is um uh uh one of my friends calls it they burn it all down approach which uh once he said that I was I was like wait what do you mean but he was right like it's it's what I've done in my work is

00:26:59

approach it as okay scrap it all start a new pretend like nothing exists and you don't need to integrate with anything that exists today and what would it be like right so that's that's kind of the

00:27:11

context that I implicitly put on the sort of the the design challenge I gave myself um and and continue too so what that ends up looking like is a lot of the things that we've said today uh so far

00:27:25

but without a lot of the um the challenges that it would take to to force them into today's software so for example one of the biggest things that um I've I've explored with uh with my

00:27:39

work and and what Wonder OS is is an itemized OS so the idea is that your operating system is made entire and we're talking about the user environment right the entire uh uh user environment

00:27:51

is made of items and so you you know we we have some pretty clear ones web page email that's an item that's an item items are made of other items so a to-do list item is made of to do items

00:28:04

um and so when you and and items can be in you know other a bunch of other items so then you end up with a graphs you know you've got an itemized graph of things um and items are strictly data that's it

00:28:16

uh they can't like do stuff on their own so it's just data um and then you can and items that data is separate from the services that Supply them if they were supplied rather than created and separate from the

00:28:30

interfaces that render them so when you when you have an item open you can change the you can swap the view that you're using to render it if you prefer to compose your emails in markdown you know you can swap to your

00:28:42

your markdown uh uh rent view right um and then services are the only way to get data in from the outside world uh and so that's how you break the notion of an app if if you could ship one

00:28:55

contained thing that glues the interface with the service with the data then people just are going to ship apps right um but that's not what uh We've collectively decided here is is right

00:29:07

for the user model um and and so in that kind of burn it down approach it it what you burn it down you let go of everything Legacy um it seems like there are plausible answers to the questions that we're

00:29:19

asking um and when you really start to Tinker with them it it makes you think I really want to let go of the computer I have today and uh um pick up pick up something completely new

00:29:31

I I think what you said is very interesting is the the notion of services which is one of the best notion in Unix uh operating systems which is that you can act from wherever you are on the US on on the US on something and

00:29:45

Unix is just a big bag of utilities that happen to to work together politely they you know they are pipeable composable rearrangeable Unix is really just a bag of utilities and nobody sort of

00:29:56

understands that or except for people who love Unix and and I just thought of something which is that the creators of Unix at some point worked on an OS that almost had uh the the metaphors that John was

00:30:08

was uh uh talking about earlier which is plan nine I don't know if you know about that iOS because it's really kind of in the weeds of computing history but plan nine was actually very fun because you could you you basically drew a window

00:30:21

with the cursor and that window could contain anything including another instance of itself so you could create workspace in that way and you could have multiple workspace pointing at multiple layers of data and uh it was actually I

00:30:36

think probably one of the best way to anticipate the future and just to to also lean into those kind of uh metaphors that weren't kind of uh that haven't been very popular it's so

00:30:48

bizarre that we still do not have what that number song called transporting Windows which is a window that can have a link that goes out of the window region into another window to point at a piece of content that is

00:31:01

represented into places it would make our life so much easier and we have aliases on our desktop especially Windows user since like 30 years we should have transparenting Windows like

00:31:12

it should be a basic stuff in the UI of every OS and every app so two trivial items two items of trivia regarding plan nine I think it was named after the the bad science fiction movie Plan 9 from

00:31:24

Outer Space uh back from 19 1957 and it was Ken Thompson and Rob Pike I think at Bell Labs busy busy doing that yes people who like are in the pantheon of of computing pioneers and I think it's

00:31:39

Thompson and Richie who do other things and like there's this Nexus there's this Golden Era of bell Labs that's worth looking back into just like Xerox park for all of the things that were kind of recruiting out of all the people who

00:31:51

were there collab waiting and you know that then jobs is trying to emulate the architectural characteristics of bell Labs so that he can generate more productivity at Apple and all that it's

00:32:04

like um such complicated history as we leave so now back to our regularly scheduled program which is already in progress well I guess you know it occurs to me we've we've generated

00:32:17

a really enticing sack of desirable characteristics or functionality should we talk about data structures that would enable them yes and two things the the enemy here seems to be uh

00:32:31

either managers or developers fear of humans that that people will be terrified as you said with gase and hypercard like there's no way um people will be terrified of this but also fear that humans will go do

00:32:44

something useful and productive together and all the coders that we need to make software for the Mac won't actually code for the Mac because they'll be like just just big hyper card into the OS and just use hyper talk right go crazy with that

00:32:56

and and just before we talk about data structure themselves hype account was originally a project to make it an OS for a Max tablet like it was it was a full self-contained thing to to

00:33:07

to to to be run on on a on on a MacBook uh and Atkinson had diversity he had a black male apple he had to threaten Apple to to bundle it with the Mac he basically said if you don't do this I'm

00:33:21

gonna just put it out openly yeah so he he had a meeting with John Skelly and John scaly loved it and Alan K was at the meeting and really pushed for hypercar to be with the mac and if I'm not mistaken I think Allen was also

00:33:33

one of the main driver uh beyond the whole hyper talk integration to hypercal because we speak about hype account but the magical hyper cat is the language and I think even today we do not have a

00:33:45

language with such a Humane syntax like the the way Harry Potter can be discovered by the user used and to to to go from I drove to I added an interactive uh thing which is a button

00:33:59

or a textual to I script that thing to make it do other things was so Progressive it's it's really like I mean it's really really hard at this moment actually not to invoke Orwell

00:34:11

right the proscription against literacy only ever has one aim and I think it actually I mean hyper cards tenuous life and death circumstances got even more acute when uh when the

00:34:25

iPad came out and I thought it was such a loud message from Apple to say there will be no programming languages I remember scratch had a huge battle early on right with the tablet which was going

00:34:37

to be an important platform for it because you weren't allowed to have languages on the device so that you know that sort of speaks volumes it's crazy too because scratch and small talk all

00:34:49

these things have common ancestry you know all these things kind of lump back together and there's this struggle by brilliant people like uh Dave Reed and a bunch of others who have worked on

00:35:01

these projects forever and get marginalized like like these power tools don't make it into the general sphere and and some piece of this conversation and I don't want to take us in that direction is about how do you strategorize getting this into like full

00:35:15

public use how do you how do you get a community that does and and I don't mean what's the business model that gets this out and successful because many in many cases we don't need a business model I know that many years ago when

00:35:29

Microsoft's laptop tax or Os tax got broken so everybody so computer makers started being able to ship a computer that didn't have to license the Mac OS even if they weren't going to put the Mac OS on it I was talking to the

00:35:43

inventor of the brain saying like if you had an a version of the brain that ran on Linux you could basically take advantage of Linux distros and role roll the brain out as the UI and wouldn't

00:35:55

that be kind of cool because the alternatives are KDE and gnome and those sucked they were like lousy replicas as far as I could tell of the desktop metaphor right and wouldn't it be cool if anybody who went down that path had a

00:36:08

completely different experience of their existence in the OS and the brain does 1 15 or 100th of the kinds of features and functionalities you guys have been describing so it wouldn't have given us

00:36:20

any anything like the all singing all dancing uh uh item world that you're describing Alex but but there seem to be these Market opportunities or movement opportunities and it feels like we're

00:36:32

also at a move at a moment right now where we need to level up this whole thing and and this important thing that's coming out of the tools for thinking conversations is hey we could talk about this cool tool or that cool

00:36:45

tool or this cool tool or that cool tool but really but really if we want to think together and if we want to make this thing work we need to level up our thinking about the OS the databases everything that we're talking about here

00:36:58

sorry to maybe State the obvious but but that's really sort of crystallizing in my head and rising in importance in this whole Quest yeah I agree 100. well I guess we're done then part of the

00:37:14

the the actual manifestation of of such circumstances it seems to me would have to involve a kind of vigorous almost rabid Community feature and I wonder if

00:37:27

bizarre as it is there's anything to be learned from Tick Tock Twitter like sort of massive sort of bell curved shape well I don't know like it's probably some sort of exponential thing instead but like

00:37:40

communities where there is enormous participation a lot of it is in consumption mode but there's never not the opportunity to dive in a little bit and then there are people who are incredibly productive on those platforms

00:37:53

but if instead of creating video content or little textual uh idea mini nuggets it were uh you know bits of data bits of code bits of overlap bits of kind of

00:38:06

knowledge Auto connecting or being guided to connect to start forming enormous rafts of knowledge floating out in the Pacific to get a little Gibson up about it we don't want a gyre though

00:38:18

yeah I agree I think building a community of content creators at the start who really get the system can help shape the system um and kind of kind of help uh you know

00:38:30

advocate for it out I I totally agree I think that would be like the key Cornerstone to get adoption um for Something Completely new so that that's kind of what we're leaning on to to try to kind of uh summarize all of

00:38:43

what you said and including and especially the distribution and the problematic aspect of the business model like when we started designing Cosmic we're like yeah we understood what's the object and we do not want to duplicate those objects we want to do away with

00:38:56

the fine metaphor so everything's going to be stored in a data pool so we use the ipfs and so we cross compiled ipfs we made it run on the iPad the iPhone the Mac Everywhere to have a local node and then on top of that we created the database that does what we think would

00:39:09

be should be Baseline for every new file system of in 2022 which is having the ability to collaborate not on the on on the document but at the object level having authorship tracking having transclusion having cryptography for

00:39:23

each object and the graph separately in different containers and a way to do what we call polymorphous data which is that once you have that kind of superstructure of the file system so you

00:39:36

have immutable data in atfs which means that each time you add a new data you add a new thing in in that file system on top of that you have you add polydata which will allow your data to be different that to me is how we get to

00:39:47

that point where in Cosmic for example you have a text we read the text from apfs in on the canvas we add the poly data to say here is the position on the canvas and series the width and the height and then here is another poly

00:39:59

data which is where is the vault in the text so where it's the underline but let's say we succeeded and Foundation becomes an SDK you take that that data structure because Cosmic already has thousands of users you build on top of

00:40:12

that ecosystem and you build an outliner or a new text app to for to for Jerry to have only one text editor and what you are going to do is you say I'm an independent developer but I want to do that app but I need the community I want

00:40:25

users to better States what you can do is you say I'm going to use Cosmic data structure foundation and I'm also going to use the ID mechanism which means that when you you do introduce your app people can log in with the same credentials as Cosmic but what's going

00:40:38

to happen is that your text editor will just ignore all the police data that it doesn't need like the size and the position of the object but what's super cool is because of transcription what we call Magic copy in in the software when

00:40:50

you modify your text here in the new text editor of auxiliary it's going also going to show up in the canvas and so that's how we hope it's going to work out by saying hey here's the way you can build new apps that have those Baseline

00:41:04

features which is ownership of the data auto ship tracking security with encryption local first offline first multiplayer everything infinite version all of that in there but when you do come out with an app with zero users you

00:41:17

can actually talk to all those users that are already on Cosmic and they may just try it and add it to their Suite of apps so if I may make another kind of 1990s reference it's kind of an open source Claris if you remember Claris by

00:41:30

the way from Apple to to I think it's a clearly understudied model to say we have a platform we need good software for that platform so we are going to edit one some parts of the of that

00:41:42

ecosystem ourselves but if we do stumble upon a very good Indie software we can just bring it in-house and edit it with the logo that customers will trust and recognize that's kind of how we want to do it and so it's it's going to be very

00:41:55

hard very long it has a lot of problems because when you speak with an investor for example how do you tell him that it's coherent for you to start with a spatial canvas for spatial thinking but you are going to end up with a platform

00:42:08

because what you actually needed to build that in order for the canvas to do the things you wanted to do and the best way is probably to maybe separate the two and have kind of a uh

00:42:19

um a non-profit or another type of organization that takes care of that underlying platform I'm excited I just want to figure out how to help roll these things out how to make them work

00:42:32

just to complete very quickly that's the problem that you were speaking about earlier with the current tool for facts if you want to have interoperability you have to use obsidian for example but then it's you will have interoperability

00:42:44

because you use a legacy system which is markdown files which means that you cannot have all the niceties that you have in your own research or that you have in Dyno no ukid on the Block Tana for example but then what if for example

00:42:56

like but then if you do not create a data structure that takes care of all of all the things that just uh announced uh and and spoke about it's very hard to take them into your uh data structure afterwards and so for example that's why

00:43:10

I think Rome has maybe some trouble just shipping and shipping and shipping because that that stuff is really hard to build so for people who are not up on obsidian markdown Rome and all that and without explaining what markdown is let's assume

00:43:23

that away could you slow that down to like 20 miles an hour and just just say that again because it's really important and it's a it's a very specific example of the the problem before any problem that all these tool makers are facing

00:43:39

to me the problem is that so interoperability means that you can do what is called in the webperiums cross-pollination which is basically the idea that you can have a file that you are going to open into Google apps to

00:43:51

enrich it or that because you use a legacy find format like markdown which is a text file format you can just say I'm going to use obsidian but if I become I don't know if I don't like the tooling more I

00:44:04

can just take that very common file file system file structure and open it somewhere else and it's going to work out just fine that works but it means that you restrict yourself as a user to a set of features that is not very Advanced like

00:44:17

doing collaboration is very hard doing encryption is very hard doing synchronization between device is very hard doing version is very hard or you have to rely on the file system which is another set of problems and so if you

00:44:29

look at the other tools that do have very Advanced features like Rome notion antenna and the other one they use completely closed data structures and I I'm certain and I I've spoken with some

00:44:42

of the people in in those companies that they have problems with their data structures because they have to be built on Legacy tools like for example you use to use Firebase to do a transcription based data structure is asking for

00:44:55

problems so at some point we have as designers and developers to say we are going to rip the Band-Aid and we have to start clean with the data structure that will be completely open source that anybody can build upon but that takes

00:45:08

those future that we want as just the Baseline and then each one will be able to fork or build or do some videos and we are not the only ones advocating for that and actually there are a bunch of other projects and systems that are very

00:45:23

well Advanced like what vision is doing for example is their web native find system or what golden render is doing with no experience subconscious like those are they're interesting projects what we hope to do with the cosmic

00:45:34

restrictor is to really make an SDK to build new apps which is a bit different because I do believe in apps as viewers to act on the data that was that was that was awesome thank you

00:45:46

um Alex John with any anywhere you'd like to go well I guess there's a subtle distinction I think if I understand correctly Paul and please redirect me but between

00:45:59

the idea of open sourcing uh access to data uh as a necessary first step to make making things universally interoperable and then the downstream

00:46:13

problem of translating and and rendering interoperable different uh different compound data structures that enable different features so like

00:46:26

there will be plenty of applications even in that world even in the foundation Plus ipfs World Plus Cosmic world that still want to use markdown but you know what about like Ian Neal's

00:46:38

codex OS standoff properties thing right so this there's still I don't want to say an incompatibility but there's still a translation problem even when all of the underlying access mechanisms uh are are

00:46:53

open and maybe there's the opportunity for a kind of meta schema self-description mechanism that can can sort of exactly ease that you know ease

00:47:06

of the translation so it's it's it could be very very technical and I don't want to get access to too long of it so it's actually how it works in the custom database and we did something that's really fun which is

00:47:17

that usually in a graph database you have objects that are connected to other objects by edges and what we did in Cosmic is that the edges themselves are objects so you can inject any kind of new data into those and actually say oh

00:47:31

I want standoff properties let's just add that in the edge itself which also means that you can you could add a code and run code not on the object itself but above which is how for example and this is a completely different issue but

00:47:44

I think it's going to be very very very Troublesome in the next few years today we have a bunch of startups a bunch of companies that are just plugging AI models in the product they don't know what trained the model they don't know how they were trained and they don't

00:47:56

know how to really handle that we could do distributed computing by just injecting a lot of data into a cosmic canvas and have something run into the edge of all of those images to train a new model in a decentralized way so I

00:48:09

think it's not only a new cool way to wear with some problems it's actually going to be needed because if the AI is becomes as prevalent as it looks it's going to be we need as the community of users to be able to train the data

00:48:22

assets and to have models that we know are biased but at least we know why and how they are biased yeah when we take that graph to the system scale if we imagine a cosmic OS if you will you know

00:48:34

I think sometimes uh uh people hear graph and and they get kind of just an abstract notion in their heads but if you take that graph to the system scale

00:48:45

uh it's you know it's it's it's this uh meta system that's running on all your devices um there are some really cool things that happen where um so for example like like today when

00:49:00

um you have a notification show up on the front door of your device on your lock screen right notifications pop up because uh an app has code that was written to say all right put this

00:49:12

notification on this person's home screen you as a user get to decide yes this app is allowed to post all of its notifications or no this app cannot post any of its notifications right that's your total agency in this situation

00:49:25

right um once we have this we've we've got this graph we can we can set up you know notifications as sort of automations that run uh uh triggered by you know say

00:49:37

that the new presence of an item or the change to an item that kind of thing but because of the graph we can do really uh things that will seem ridiculously obvious and simple uh in the future but

00:49:50

now are completely impossible so for example uh emails you don't want to get a notification on your home screen of every single email you receive right those days are gone there's just too much junk coming in it's a fire hose so

00:50:02

you start with your notifications off and you're in your graph right then you you want to set up a notification to say okay when my spouse emails me yes notify me every time immediately I want that on

00:50:13

my home screen I want that no matter what's happening I want to see it right but you can start to do really interesting things because of the graph where you can just say okay if I get an email from someone that I'm currently in a meeting with or about to

00:50:26

be in a meeting with hit me with the notification of that if I get an email from the website that I'm currently on notify me of that right covers all of these use cases someone's saying they're late someone's sending you a meeting agenda right before the meeting or the

00:50:39

zoom link um covers you know the website sending you uh a a you know a discount because you just signed up um your magic sign-in links like all of these things and so then you haven't turned on notifications for all of your

00:50:52

emails um but you get to do these little and your system has all this information in it it knows what meeting you're in it's an event it it's that item is is is associated with the contacts of the

00:51:04

people who are in it so you can set up these automations to for user-defined notifications which is how notifications should work um based on all the items in your graph and it will seem very simple and obvious and people will look back on today and

00:51:16

laugh at how claimed our uh our our rigid systems were you know we used to have to start cars by turning a crank out in front front of the car airplane so it feels so often with software today yeah totally and I I

00:51:30

think there is something else with that with those statistic structures that comes back to the iPad the iPhone the mac and basically any ecosystem today which relies on a lot of devices yeah so so basically I think that we are going to live in a world where you want

00:51:43

to not only switch contacts between devices but you want to use your devices concurrently for several tasks for example one cool use case with special canvas is that you can draw on the tablet while you write text on the Mac

00:51:55

stuff like that very hard to do with a file system that we have today not very it's not it's not well designed for that and you can imagine how having a collection of devices could help you uh in several contexts like you could have

00:52:07

someone in a VR headset and so the poly that extra care would have a coordinates for the 3D system or you could be on a 2d special canvas or an on and on and on so it's also needed to kind of enhance

00:52:20

the experiences we have today because if you look at how a magnet and iPad works today there are nice things but be on the air drop which has kind of that magical feeling of a file flying between the devices nothing is that cool like

00:52:32

Pages look exactly the same uh may all look exactly the same and it's not that like it could it should do much much more because the iPad is a different device than the mac and and or the surface is a different device than my

00:52:45

desktopc and it goes back to another thing which is that if we do that we could have what I would call I don't know if that's the right term in English I kind of this time the computer computer that are made for one tasks like because you would have that common

00:52:57

data structure that would be somewhere on on a central CPU or Central server that could be in your home because it's decentralized you could have a computer to draw with very different brushes and inks and stuff like that and you would

00:53:10

be able to draw at that computer but then you could have a computer for music but then you could have a computer to write and then you could have your laptop because you still need the general purpose computer but I think it's also one of the other problems the computer today does too much things and

00:53:24

maybe we need What Jeffrey skin was also calling information appliances devices that are designed for specific tasks and that are very good at those tasks sometimes you need to have a jack of all trades but I do feel and think

00:53:36

that also there are a lot of contexts where you would benefit from having something that is very designed with a very specific and narrow usage in mind uh because it would it could be better at that desk

00:53:48

that's also the key out of this whole always having to be Tethered to the cloud and your data always going to the cloud problem because if you've got your personal Computing network uh you know your your video camera on your doorbell

00:54:01

is is on that network but it can offload the video processing to say you're you know very expensive huge tethered into the wall um uh you know battery backed up uh uh

00:54:14

sitting idle right now not doing anything exactly so you can you know once you've got this personal Computing Network where where tasks specific to different devices can be kind of offloaded uh amongst uh you

00:54:27

know the things on that Network then then you can you get to have your own personal you know everything you need in your network um and and so that kind of that that's one of those keys that would help untether us from this oh well you have

00:54:40

to sign up and subscribe because only our servers can can do this thing that you need done you know like having your video streaming all the time which is a you know a major privacy uh issue

00:54:54

this call just keeps bringing up things that have been pet peeves or irritations for me for over the years and all the examples and the things we've pointed to as historic examples whether it's Xerox star or Claris or whatnot I'm like wow

00:55:08

and and even the example of I do not know how to make my Android phone send me everything my wife wants to send me and nobody else like I cannot make it do that it isn't working and that's just

00:55:21

like a very simple task you would think right and part of this is is sort of blunders in design at the OS level there's like I think part of what we're saying here is that you need to unpack how many places something went wrong

00:55:34

that make that hard to do hard to code and then you need to leap over and say you know how do we jump over into a green field and do something different and and I'll say that I've been fortunate and a little unlucky in my

00:55:47

life to run into several Geniuses who had invented something in you know in a green field that was never obvious that was just ever going to make it to Market and there was there's a fellow named Sandy klausner who invented a system

00:55:58

called cubicon that basically was Soup To Nuts from Hardware all the way up to Reinventing the web and networking and everything else he designed he was not a coder he was just a brilliant person he designed it all in director and then

00:56:11

there was one point at which the Mac changed the screen resolution from either squares to to rectangles or rectangles to squares I don't remember but if there was this huge shift in how Mac rendered uh you know pixels on

00:56:24

screen director he had drawn 9000 diagrams in director to explain cubicon none of which worked in the new rendering so then for years he had to buy old Macs and keep them

00:56:36

alive in order to do his rendering um right and I'm like wow and he was busy trying to convince Intel or other companies hey if only you do this but but and when you looked at it it was

00:56:48

like he was asking us to have precogs bathed in a nutrient solution with a Jack in the back of their heads in this new system it was so foreign to whatever was existed today right it was just it

00:57:00

was too far too far too far too far ahead and what we're talking about here is not that I'm not I'm not trying to paint it with that I'm trying to say this seems so pragmatic and so doable but we have other market failures which are you know this fear of humans and the

00:57:14

competitive problems and all and and the occasional character who stops a great idea cold um out of one of these one of these kinds of fears so so maybe that's why we have to start small just yes the Mac was

00:57:27

kind of a miniature Xerox star retailing for a quarter of the price and it still had kind of it's some troubles making inroads at the beginning but I I think that's maybe how you can you can try to do it at least you start with something

00:57:39

very very simple that's like oh now you can just pull out an image from the website and that image knows that it comes from the website so your research workflow is better but people don't care about all the Machinery that's down

00:57:50

there but then gradually you can expose it by better features more complex programs and going down that path but it's very hard but I think it's it's doable

00:58:02

agreed and I don't know how much you want to talk about this but but both of you like what paths do you imagine for Cosmic and wonder uh to to get in the world like like then how do we help and how do I become

00:58:15

part of this Posse of people you know building the original Community uh you know one one of my goals in life is to take the data that that I've poured into the brain for 25 years drop it in the Commons and use it as a sourdough

00:58:27

starter or as a mycelial inoculation into the global brain I'm like I'm like I'm dying to do that but I don't know what to inoculate where to put it and how to share it out with other people I I would love to do that yeah you know so

00:58:41

when you when you burn it all down and paint a completely new picture how do you how do you then connect the dots from from the present to that future um I think for something like Wonder OS which is is very much just a research

00:58:53

project certainly at least at this stage um and a very janky one at that uh but a delightfully janky one um you know I with that project in particular uh with Wonder OS it starts

00:59:06

as an alternative user environment one that doesn't do nearly as much as it should but does do some things and does those exceptionally well and and offering you agency and and uh

00:59:18

increasing creativity and your curiosity and your ability your opportunity to engage with the digital things of your world and and map your thoughts you know onto these external digital surfaces which should be better than the physical

00:59:30

services and yet half the time you're going to reach for physical services today because it's just going to be easier which that's horrifying right like these screens are like a completely Limitless canvases we can put anything on them we want and yet we haven't figured out you know these these how to

00:59:43

reify these these Primitives with which we think and speak with one another um you know into into digital world um so uh sorry I you can get me on a tangent on that stuff any day of the week the but so so Wonder OS you know it

00:59:56

you know starts as an alternative user environment for an existing computer um and you you boot up that alternative user environment and you bash around hit the walls a good bit um and when you hit walls that's when you realize okay outer space needs to

01:00:09

expand here in this spot um and you just you keep expanding that'll alternative user environment and then at some point you start to ask kind of the bigger questions of when does the user environment then expand into other

01:00:21

things love that thank you um Paul so I I think we are going to try that that idea uh Wisconsin which is to what

01:00:33

we started to do which is hey here's space from canvas that can do great things like for example you can create a mood board in minutes instead of hours or you can browse the web in a new way which will make your research workflow better and we target it at very

01:00:46

very specific user segments like interior designers designers developers researchers like very very narrow and a few companies are already asking us like can I use that software because I feel that it would be a better alternative than viral or mural or thick jump for

01:01:00

example or other products in that category because we need something that's richer something that can also enable this new way of collaboration that doesn't need everybody at the same time in the same room if you will so I

01:01:13

think like I had a talk with an invisible but that right before the podcast and it was like it's you have to have chapters and chapter one is a single user web browsers for Creative people and that's already quite a lot

01:01:26

then you make it multiplayer and becomes a collaboration tool for visual thinkers and then once you've done that you'll probably have glue and you will have a database of or you will have a user base that's quite large you will have Revenue

01:01:38

then you can open the first chapter which is hey by the way it was built on that since the beginning that thing Foundation is not available to everybody that wants to build on top of that and we can actually help you so the only

01:01:51

thing is that I still believe that I need to um you know get that message out there right now and kind of have that master plan in the open of hey I know we're at chapter one but I'm really working to go

01:02:03

to to chapter three so here's how it's going to work and you can follow the journals here so that's kind of how we hope to do it and honestly I think it's uh it's a 24 to 36 months adventure to

01:02:16

get to the beginning of chapter three wow yeah makes total sense um I'm reminded of um a story about the Wright brothers because I read one of their biographies

01:02:27

and they had a 10-year lead on everybody else they figured out pitched y'all enroll they invented the wind tunnel they like like when they flew their plane other people were basically putting one of these early Motors on top of a wing and shoving it down a ramp and

01:02:40

saying you know hope it's Gonna Fly and a little bit better than that but but the Wright brothers kind of sourced it all out and then the Wright brothers um reporters started coming out and filming them flying they were afraid of

01:02:53

losing their IP and then they wasted 10 years trying to sell this aircraft to armies none of which none of which bought their airplane until Lewis put the rudder in the back and then they bought Lewis's airplane

01:03:06

um but the Wright brothers basically lost this gigantic lead they had on how aerodynamics actually work and so I'd like to I'd like to sort of skip that here and and figure out what

01:03:18

are the fruitful Dynamics and I think they're showing up in the conversation here of getting this in general use starting with chapters starting with something that people can understand but not not dumbing it down so much that

01:03:30

that we've like snipped away what the actual utility is because that's that's what tends to happen here is that we're so afraid people won't understand something new that we just like shave it shave it back to nothingness right so I

01:03:42

I think it comes it comes down to how do you speak to different communities like how do you give Cosmic to designers or developers and then at the same time explaining oh by the way the imagining that special canvas here's how it works

01:03:55

and for you developers here's what you can do with it and another thing that goes back also to what you were thinking about the right browsers is that I think that in startups and Tech in general we are way too worried about competition

01:04:07

and IP like nothing's gonna happen like those ideas are very large they've been around for some of them for 60 years like transplanting windows were designed in 1974 and we still do not have them so

01:04:20

that's also why when someone's an investor or developer someone tells me why would you like to open source your most valuable thing and I'm like because it's going to be more valuable if it's out in the open and other people are

01:04:32

using it than if I just keep it for myself to build my small software that will maybe one day be have be a 10 million user user base like it can be for a billion people for maybe more but

01:04:44

and it will generate value because it is powering other things and and that's how I think it's it's it should work like that we should we should not speak about how to protect that we should speak about how to distribute that like it's

01:04:57

almost a distribution issue and I think that funders don't seem to appreciate is is that and or maybe this is inventors don't seem to appreciate is that the open source strategy is a great defensive strategy for keeping something in the world because as soon as it's

01:05:10

proprietary and something goes wrong somebody else ends up owning and you don't get to work on the thing that you developed your whole life I don't know anybody who's been through that who might be in this room but um and I think that's another answer

01:05:22

another solution part of the solution to the problem that you raised um and and I love that it's it's what what Paul's doing with Cosmic they're starting with even though they're targeting you know specific use cases they're they're building them on those

01:05:36

core fundamental Primitives it's the it's the harder path to take it it certainly takes infinitely longer Paul will admit to that um but they're they're building those specific use cases on the core

01:05:48

Primitives on the database um from the get-go so even though it seems like oh wow this is perfect for this one use case then he'll be able to pull the curtain off of the other you know five doors and say well we these

01:06:00

Primitives apply kind of across the board yeah and I think on top of all of that uh the Fantastic instincts that Paul has about open sourcing critical parts of it

01:06:12

the fact Alex that you're uh you're engaged as an independent researcher not as a startup is that we need to um we need to synthesize new kinds of human organizational structures that

01:06:24

enable promote and reward collaboration across heterogeneous types of substructures whether it's startup whether it's a chunk of Academia whether it's one person who's crowdfunded to you

01:06:37

know undertake critical research so that the whole thing can kind of move in the in roughly the same direction and and that stuff doesn't exist there really is I think room for critically important and exciting invention here like all of

01:06:51

all of the old models are by themselves kind of rotting and demonstrably ineffective yeah I feel like you're a time traveler from the future who is describing at least a future episode of the tools for Thinking podcast and

01:07:05

hopefully hopefully the world that we're walking into sorry Paul go ahead and no no it's just that this is why you can't you can all help because it's exactly exactly what John said the problem is like the tech will get done it's hard as

01:07:19

Alex said like it probably took us two and a half years more to build Cosmic display than it would have taken if we would have just take a file based backend and and use the web Technologies like we choose to have a rest database

01:07:32

cross-compact webassembly with a local ipfs node that works then with interfaces and JavaScript typescript and Swift like it was horrendous and I I suspect that if some uh some investors

01:07:43

had has had a very technical view of what's Happening inside the company that would get frightened but that didn't happen and now we have a true mode the the problem is not the tech it's how do you create paid the structure is it a

01:07:57

foundation is it a non-profit organization is it another company is it a holding company where is it uh is it online does it have an office like who takes care of it who has stakes in it is

01:08:08

it the people is it the other companies like I'm I'm ready to put what we've did at Cosmic Foundation somewhere else that it needs to like for example I know that

01:08:21

the company in itself today is structured to handle how to code that thing like we can do it but it's not structured at all to handle pull requests from the developer Community to handle documentation to handle open

01:08:34

source licensing licensing issues and all of that that's where we need help and that's really what we are trying to um like we don't know what we're doing there yet and one of the problems here

01:08:46

is that the standard model creates centralized behemoths that try to become unicorns and then you get a single supplier and then you get a choke hold in the marketplace and choke holds are not good for the kind of environment

01:08:57

we're describing you don't really want there to be an Elon Musk or a Mark Zuckerberg who control I mean Mark Zuckerberg is basically the king of a country that is larger than the populations of India and China combined

01:09:11

no seriously like like the average monthly usage of Facebook is larger than the full populations of India and China and there's this one guy who's like we'll do this tomorrow

01:09:23

um it's just bizarre well and that translates all the way down to your devices you have all these little silos on your devices which take away your agency bingo bingo so what what we are thinking about

01:09:36

with other Founders is let's say we have this feedback started which is in itself also means that we are not as independent as as we could be but we still are we are thinking of what happens like we we all want that thing

01:09:49

to happen what happens is if we create a foundation but then we dedicate some of our own resources and we pull them together to the foundation and we agree that's what's pulled in the foundation will be uh dedicated uh democratically

01:10:03

like we will say hey here is the part of the data here are the parts of the database that we need for our products and someone else will say hey here's what I mean I've committed two developers with Community two let's vote on which one we do first and we know

01:10:16

that we have a common goal that we are we need to be very Democratic in the way it works um and it's going to be hard but it's also maybe the only way to prove to the outer world that nobody inside that

01:10:28

company or foundation will become that King because that's exactly also one of the most I would say uh block to the adoption of new technology which is to say they say they are open source but are they really

01:10:40

how is what you just described different from an open source Foundation today a regular one when we can point to dozens of them it's not that different it's just on a smaller on the smaller case maybe what's

01:10:52

different is that it would be an early stage or early to late stage startups bending together with competing products in the marketplace on the front end but common objectives on the back end that's

01:11:04

to me would be uh that's fun because what I love about the tool for performance ecosystem in terms of startups is that we all have competing products that are sometimes we're very close but we still exchange ideas very very freely because we're like they've

01:11:18

been here for so many years that we are obviously going to diverge at some point and we aren't going to like no one is going to eat the other one right now not now at least so that would be maybe the the new thing

01:11:30

which is to for-profit companies have very capitalistic models to make money to create value for shareholders bending together and saying yes but in the back end on those very foundational things we

01:11:43

agree that we need them and that we need to create them for computer for personal Computing to advance so we agree to take some proceeds to work on those because all of our products will get better and also on that like have you seen what's

01:11:56

going to happen to notion with Microsoft Loop like Microsoft is a very interesting company now they are doing a lot of work that is quite interesting on the office suit they came up with a framework that basically enables transcription between all of the office

01:12:08

apps and they also took that and created a clone but an exact clone of notion called Loop that's not available yet and you're like that's too bad because even if notion is a 10 billion dollar company

01:12:19

it may well have they may go down the slack route which is that slack which was kind of eaten by teams even if teams was kind of a not as well done product and so you're like but what if the

01:12:33

notion data structure would have been completely open source and they would have been tens and tens and tens of products to build on top of those Technologies to make notion not only a great and beautiful product and startup

01:12:44

but also a new way to build software and so that's how I think we as Founders and as people that have to create value for the shareholders that's how we do it because by agreeing to open source parts

01:12:57

of the thing they pay for we actually reproduct their Investments because we pull those things in other areas where they can be built upon and other companies could emerge and sprung from that ecosystem

01:13:09

so that's also I think our better defense mechanism against the big companies um it's funny I walked into this call not as an architect or anything like that but uh in 1976 Nick lostworth wrote

01:13:21

a semi-famous book called algorithms Plus data structures equals programs right and I had kind of this primitive notion of okay if we can separate the apps from their data um then we sort of get somewhere because

01:13:33

then we get a lot of the the blimp like behavior that that John described and everything else that we've been talking about except I was sort of skinning that down to markdown files on GitHub repos which is what I'm trying to do right this minute

01:13:46

um and that's clearly too simple that that actually it doesn't just properly distribute the data and it it doesn't it fails on so many other kinds of uh kinds of criteria um so I think telling this story

01:13:59

explaining why and how to go through what what we've talked about in this call is is vital and I'd love to be helpful in doing that I think we can pull together different groups of people who would who have different kinds of

01:14:12

skills in in crafting stories telling stories uh animating stories whatever else that would be a very fun thing to devote some some energy to and and then never mind exploring and then setting up

01:14:26

the kind of organization or institution or institutions that you were just describing Paul that I think are absolutely necessary for for this field to thrive and and I'll just add one more thing which is

01:14:38

I think this field needs to thrive not because tools for thinking are kind of cool and uh gosh wouldn't it be great if we had you know some more of those I I think we're talking about civilization here because if we can think together

01:14:51

better we can change how science Works how journalism Works how Education Works how elections and governance work all those things have we sort of if this infrastructure for thinking together is

01:15:04

fruitful then it infects and transforms a little bit like Ice Nine to put in maybe a dark metaphor all the other ways that humans are trying to sort things out to make decisions together and that

01:15:17

to me is is like important work right now and that goes all the way back to Engelbert with the mother of all demos right so mother of all demos engelbart the inventor of the mouse but what did he spend the next several decades you

01:15:31

know what it's it was boot shopping it's how do we come together and do what we do collectively better and that's why you know know the data structures are so important these are the things we think with

01:15:43

right if if we can if our tools can literally help us think better um which better data structures uh help us reflect our thinking and manipulate it in ways that are more intuitive to

01:15:56

our minds um you know then everything we can do we can do better which sounds like we should break in we should break into song right about now but still I I I think our work will not be done as

01:16:10

long as people feel needed to have a notebook with them on top of their laptop like this is the goal like we speak about that a lot of times internally it's this is this is actually

01:16:22

something that's quite mental like we have those devices like you have a 2500 MacBook with you but you still have them all the skin because the Macbook will actually do things less well than the model skin in a lot of case and so that's why we also need new data

01:16:34

structures because it shouldn't be that way um and we've gone pretty much 90 minutes which is sort of the length of these calls I'd love uh and it feels like we're sort of adding kind of uh codas to

01:16:46

the to the conversation but any other thoughts that we want to put into this conversation before we before we wrap it please jump in sometime if we get the chance this would go completely off topic but John I would

01:17:03

love some time to ask you a little bit about the mental models behind some of the things you've designed and the ones that you believe are missing from mainstream software today great let's let's all stay connected and

01:17:16

um yeah maybe we play our cards right uh Jerry will invite us back for part two sounds great I think it sounds great that's really cool yeah same here same here I'm I'm vibing with this whole

01:17:29

thing very wonderfully um so thank you very much um thank you and uh thank you all for listening to tools for thinking a new podcast that just might help you with your thinking if

01:17:42

you're part of a startup in the sector please knock on beta Works door at betaworks.com um and that's it for now thank you foreign

01:17:53

[Music]