Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
2afd69a4979a Kira Howe
846 Plays6 days ago
Guest
- Kira Howe
Kira's web site for all the things
Hosts
- Josh Laid Back Glover
- Ray The Revolutionary McDermott
OTIS
MP3 SHA256
4467e3c8073b7f80d9e14f0756beee1fcb72e30c2c2d0b85e5c4866ebb62752b
UUID v5 from SHA + People
96efa7dd-1a8e-5f1b-94a0-2afd69a4979a
Recommended
Transcript
Introduction and Banter
00:00:17
Speaker
hello we starting shit
00:00:26
Speaker
Okay. We, maybe everyone started now. Okay, Josh, make it official. What? Me? Why? Why me? You're, you're the main, you're the main host. Yeah, Anyway, we, we always just start- I'm the understanding host, but I don't feel, I'm the father of the house. so hello, Kira. Oh, You know, I feel like you're the sort of, uh, I'm the chairman, hello. you're the CEO, you know?
00:00:46
Speaker
okay You just pulled a Putin and changed, uh, changed titles, but you're still, you know, running the show. Well, I think the definite thing is just to start, you know, in, in media rez as it were anyway. Right. Like Valtter's just kind of takes Walter. God damn it. Sorry. I'm like VJ. I recently discovered that I've been mispronouncing Walter's name for eight fucking years. So anyway.
00:01:10
Speaker
but It's with a W. so um But yeah, Chirith, thank you for joining us this um morning for you, I guess.
Introducing Kira and Name Discussion
00:01:20
Speaker
Afternoon. Afternoon, yeah, but it's nice to be here. Thanks so much. Yeah, welcome on Dethan episode, insert hash here.
00:01:29
Speaker
um Yeah, we we love hash on this ah on this cast, that's for sure. Yeah, it's hashing.
00:01:39
Speaker
Why is that funny? Come on. It's not. it's you're childish Yeah, it's funny how childish you are. Exactly. But, uh, so Kira, since, since I invited you on this show, I mean, Ray was against it, but I really pushed. Do you want a woman come back? Oh my goodness. Never again.
00:02:06
Speaker
but
00:02:08
Speaker
do you want yeah this is getting this is getting off to a great start everything is completely alive so far you know I adore Kyra. Yeah. Okay. First truth. Let's go. Yeah, let's go. Yes. Um, but you really want me to start, right? Cause you always start. So you caught me unawares. Okay. You should start. Yeah. I think we're started anyway. It's fine. Anyway, let's, let's give the congratulations first cause Kyra has had kind of like a big year, especially with getting married. So congratulations. Congrats.
00:02:47
Speaker
Thank you. Yeah. So exciting new chapter for sure. It was smart of you to name your website, Kira codes and not insert surname here. Well, I have, I, that did occur to me like a year or two ago when I set up the code with Kira, um, website, mostly for the closures together updates. I was like engaged at the time and I was like still not a hundred percent sure how the naming thing was going to go. But then it was like.
00:03:15
Speaker
Uh, on my radar. Yeah. But I do. I also had cure McLean.com and now I have cure.com too. Um, ah let's talk about this naming thing then. Cause apparently naming is very hard. It's a computer science problem.
00:03:28
Speaker
Yeah. not about daming but about naming yourself Yeah. But exactly. So how does it go? I mean, you know, yeah, it's an interesting thing. what i did right It's not the kind of thing that comes up like for most software developers who are men and like.
00:03:44
Speaker
I don't know. I guess, I don't know. I didn't, it didn't really cause a lot of issues for me or anything like that or hasn't so far. I mean, Yeah. Like there's, there's pros and cons and and whatever. I live in a pretty like rural, small place where it's like very common to be like the Johnsons or the house or the grounds or whatever. And it's like, you know, everybody's kind of the very stereotypical little family. And so I like the idea of kind of having the like family name of the family household thing going on. And there's, yeah, there's like some,
00:04:19
Speaker
some miniscule public presence or whatever out there. or like so Like I have some like workshops and stuff like that and talks published ah like on various places on the internet, but they're like not, you know, super widespread. And I think it's like, there are so few, I mean, there's so few women in the public, like software engineering space in the first place. And there, there are no other ones named Kira. Like I think most people just know me as Kira anyway, and not Kira McClain.
00:04:44
Speaker
So it's like your house, not that different. It's like, oh, it's like the girl who writes closure. um Yeah, it's funny though. It's a cultural thing because like you say, that everyone wants to be the Johnsons or the house or whatever, but we're here ah in Belgium, um where I am. like um Like, my wife has my name, but if you if you were in Belgium, then it wouldn't be like that. All the women retain their own names. So every house has like,
00:05:12
Speaker
they They have like the the the ah hyphenated surnames, but but it's because the man and the woman both retain their names. But is to put the children tend to adopt the father's name. So it tends to be like that. It's a patriarchy in that sense. you know Right. Yeah, I guess that's, must be more kind of European,
Community and Programming Experiences
00:05:32
Speaker
like typical in Europe and in Quebec, which is like the French speaking province in Canada. That's how it works. You, you can change your name, I guess, but it's, it's a much more of a, like ah ordeal a Like here in most of Canada, outside of Quebec, it's like trivially easy once you get married to just
00:05:49
Speaker
assume you, it's, they call it assuming a name. So you're not actually, you haven't actually legally changed your name. Like you don't get a new birth certificate. You don't get a new yeah like like entry in the vital statistics record or whatever as a new person, as you would, if you were legally changing your name, it would, it would like backfill everything previously, but assuming a new name is like your, your birth name is still a legal name. So it's like, kind of, you just have two legal names. It's actually interesting. It's like, you know, talk about like values or, or, like, now I just have two pointers, but they're actually the same. They're an immutable identity.
00:06:26
Speaker
Yeah, well, exactly. It's like values change over time, but you're still the same entity. So it's that kind of concept. Is it like in Sweden, Josh, actually? I mean, I don't know. Do you, in Sweden, do they have the same thing where you change names by sort of like, let's say by convention, you know?
00:06:49
Speaker
I think come most people, I think it's more common to not change your name. And then like quite often kids have like a hyphenated name. um But some people, like i I know like more than one couple who have, um they've both changed their name. So they've picked a new last name and that is the new name of their family.
00:07:16
Speaker
because they wanted their kids to have the same name as them. So they just like, like you said, Kira, it's nice to have the family name. So they're like, cool, we'll just invent a family name for a family. Yeah, that's fun. I've got a couple of couples who have done something like that. It's possible, but it's not as common here. But yeah, you see some of that and you see some of the hyphenating and or like the mom keeps the name, but the kids get the dad's name kind of thing. So it would be, I think that's like the default probably, but you could probably choose to hyphenate it or something. Um, but yeah, it is interesting. Like I hadn't really ever put much thought into it like in my life before. And it's kind of like,
00:07:57
Speaker
Cause like I'm not, I mean, I'm not old. I'm like 32, almost 33. So it's like a little, definitely not old know but you know, you're not like 19. Like it ah totally, like, you know, if you're like 17 getting married, you're like, you don't, nobody knows you. Like you don't have any, you may as well just go by a new name or whatever. But so definite established that old yeah. Right. It's like you have like.
00:08:19
Speaker
you know, like work in your, your former life, but I don't know. But yeah, so far, like it's, it's not been an issue. Like some people were really dramatic about it. Like they're like, Oh, like you're going to have to start your career over and never, nobody's going to recognize you. And I was like, that's, I don't think people are that dumb. Like certain, like none of the people who care about what I work on are like incapable of processing this pretty minuscule update life. I mean, like you said, they're all running immutable databases, so they're fine. They understand the concept of change over time, like it's not a very complicated process. So Tony, in which you were at the conge last year now. It's been a while now. I don't think you've been to the conge, have you Josh? Never. I've never been either.
00:09:13
Speaker
What's it actually like to go there? I've talked to people who've been there, but it's um I don't think it's such a huge event, actually. I mean, I kind of like have this in my imagination in my and in my head. It's like like some sort of like mecca where there's millions of people, but I believe that it's sort of like a relatively modest conference. yeah I think it's, yeah, I don't actually know. I think it's probably um like a few hundred. It's, it's certainly like not thousands, but it's, it's like, if I had to ballpark, I would be like a roughly 500 people maybe go. Um, but it's, it's definitely super fun. Like, I guess I've only gone twice. The first year I went was the year I was speaking there. And then this year, um, I went with the, like I had just joined, um, broad peak, the company I'm working with now. So we, I got to go and kind of tie it into a trip, uh, to meet some of my new coworkers and stuff like that.
00:10:02
Speaker
But, um, both of the, both of the times I went, I had a really great time. It's like a chance to kind of put some faces to names. Like a lot of the big open source maintainers were there and it tends to attract like the kind of core closure community. So like, obviously we're taking egos and Alex Miller goes and, um, some other really familiar faces like, uh, the closure camp people who I had been kind of working with and.
00:10:27
Speaker
So I like, I like that kind of part of it, getting to like chat with the people that you work with online or whatever, um, face to face. And there's a little bit of that, like starstruck, like, you know, you see that for a take you walk in and he's just like a guy in a, like a tweed blazer or whatever. And you're like, Oh my gosh, that's rich Hickey. It's like, Oh, right. Like he's just, he's just a guy. Like he's just a normal.
00:10:49
Speaker
you know But did he have a scarf on? No, cause I mean that degree of kind of, you know, elegance that he's adopted recently has been very sophisticated. I think I met David Nolan at one of the Europlasures. I had no idea who he is, was, sorry. We were just chatting, like, you know, some American guy, we were chatting, you know, it was Berlin or something. He's like, Oh, I need closure script. You're like, wait, what?
00:11:16
Speaker
No, I think he mentioned he had a talk later and I was like, Oh cool. What's it about? And he was like, Oh closer script or I'm done next. That's what it was. oh yeah And I was like, Oh cool. How long have you been using it? He's like, Oh, you know, well, and it wasn't until much later that finally I realized who he was.
00:11:35
Speaker
but but That's that sounds that sounds totally right. Yeah, i've I've encountered David. I don't think we've had like, like much in depth interactions. But I find that it's like a very common posture, like a lot of closure people, like obviously we're saying he's very recognizable because it's like a public figure and guys like Alex Miller and Stu Holloway and stuff like that. But they're all incredibly like humble and kind and nice and like that, like,
00:11:59
Speaker
that type of attitude, I think really permeates the community and and and comes through. And it's true. It's like very common, like that'll happen. You'll be talking, like I met at this conference, I got to meet Colin Fleming, who's the creator of cursive. the show yeah Also just like super like genuine, kind, sincere, and like humble kind of person. Like if you had a conversation, it would never come up that he was like this ID genius. like yeah Yeah. And then you'll be talking to them and be like, Oh, yeah, we like got this talk at this conference. And it's like like, Oh, I'm giving the keynote at like the main closure conference or whatever. And it's like, Oh, whatever, nevermind. But um yeah, I used to work with, um, I used to work with David actually, and Mike, uh, Mike Fikes. Oh, really? Just like you said, they're just like the the sweetest guys, you know? Yeah. Yeah.
00:12:53
Speaker
I mean, when you work, especially with David, you know he's he's like ah he say a really smart guy, you know so he prepares to listen, you know that's for sure. You're so good at listening, Ray. Well, I know. I taught him all I know. but Turns out it turns out that the subset of what was different to what he knew is very small. God damn it.
Closure in Data Science
00:13:20
Speaker
Taught him all you knew in an afternoon. Yeah. That is my favorite part of of working with with guys like that. like is yeah it's It's always fun to you know be kind of like in the presence of people you can really learn learn a lot and and soak up all kinds of stuff. I've been very lucky that way for sure. Yeah. I think actually, I mean, it's a bit it's a it's a
00:13:43
Speaker
You know, I think working with closure people, I've always found that, um, yeah, people are pretty, pretty great, you know, um, being very lucky. And I think. Part of that, maybe I don't know a part of it is that, um, people, people tend to select closure and then that somehow makes them a bit like special in some ways because it is a peculiar language. You know, um, I mean that in a kind of possible way because I'm part of it. So, you know, I'm infected by that sort of like the bracket mind virus. I'm sure I said the friend mind virus. Um, but the, uh, I was listening to your conversation with Eric Norman.
00:14:19
Speaker
I think it was the most recent one or the one before us anyway. yeah you know but ah kind of like There is a little bit of this, I don't want to call it cultiness because it's not quite that bad, but it's yeah, this like oh culture people are awesome. And this is a very special club full of really amazing star people. And it's like, there's pros and cons, but yeah, the pros definitely like the community is is definitely, I enjoy bill being a part of it anyway.
00:14:46
Speaker
Yeah, but I think it's really, I mean, I think the thing that Eric was talking about and I think the thing that we want to promote in general is like being more like welcome to other like newer programmers and programmers that come from like let's say, you know, like the different, ah most programmers arriving in ClosureScript, they're going to come from like JavaScript, Java, Ruby, whatever, you know, they're going to come from like different languages that um but have different ways of thinking about things. um yeah totally I think that was the thing that Eric was kind of talking about, which was this challenge of how do you actually like,
00:15:30
Speaker
get people to understand a way of thinking about stuff that's not object oriented. Um, and it's not like, okay, now you've got to understand what I'm, you know, what a monad is or what an applicative function is. I was like, uh, yeah. Cause I wouldn't pass that interview, you know? so I know. Yeah. Yeah. Yeah. No, totally. and And that's the thing though, too. Like I really admire his approach to, to teaching and community building and stuff like that. Cause that's, he's totally right. It's important to.
00:15:57
Speaker
to also like do that in a way that doesn't alienate people such that like their first encounter with the community is like, okay, everything you know is stupid. And here's the right way to do it. That's not like a productive way to bring new people on board. So, but I think, I think, yeah, we're, we're getting better at that. And there's certainly less of that enclosure than some. Yeah. Like the, the sort of like hardcore less bweeny stereotype is less pervasive here for sure.
00:16:24
Speaker
yeah But what's your experience like that? Because you've you do a lot on the data science um side, right the cycloge. Do you get folks coming in who are like coming from Python, or is it is it more closureists who really want to use closure for data science?
00:16:41
Speaker
There's a bit of both, but yeah, most of the people we encounter are not, uh, closure developers. So, um, really yeah, so it's really interesting for sure. Like there's a lot of, uh, so yeah, a lot of data people use R or Python and R is kind of like, I wouldn't say it's similar to closure, I guess, but it has like some of the similar concepts and it's like a little bit functional, um, in the way that it kind of structures, um,
00:17:10
Speaker
like data processing workflows or whatever by default. So it's a little bit more familiar to people, but yeah, there's, I mean, there's definitely a huge mix. There's, there's some people who are like new to programming and they just, they want to learn programming through like in data science or data analysis, kind of like.
00:17:30
Speaker
Or like or like learn programming for that purpose And so they'll encounter some thing about closure that that piques their interest or there are some people I would say the majority of the people and it's hard to say I guess because at least the most active people are the most visible people who I encounter the most frequently um come to the community because there's like some aspect of what they're working on that They like, they ran into the limitations of the other ecosystems like the Python or our toolkits that they were using or whatever. And so, and then found some like way to do something in Closure or sometimes like Java and then discovered Closure because they didn't like Java um or that kind of route. And then there are some people who are Closure developers who then like
00:18:18
Speaker
choose to do data analysis work with closure. Like that's kind of my situation. I was, I was a software, just like a web developer before, um, and then kind of worked with these sort of like data intensive web apps. And that kind of morphed into working primarily with data and then less with web and still do kind of a little bit of both, but that's kind of a very blurry line these days, this whole like data engineering, software engineering,
00:18:45
Speaker
data science, data analysis, like pool of skill sets is kind of very mushy right now. Data stuff. Yeah, it's like people who work with data. It's like a very constant struggle to try to figure out like how to and describe what we're actually doing. Because it's like part of part of but what's behind that is like, you don't want to be exclusive and be like, Oh, this is for data scientists. Because there's a lot of people who do a lot of really interesting work with data who wouldn't call themselves data scientists or data analysts. But what they're doing is, is
00:19:21
Speaker
Definitely like in that realm and so it's like trying to kind of reach them and and you know include them in in these types of conversations is challenging. The funny thing is to me is like when i when I first got into closure and when people were talking about closure, it was always about, and it still is to some extent about, well to a large extent,
00:19:43
Speaker
but closures data orientied you know that And I think people meant by that, that you take a web request in, it becomes a ring, like but um a map, and then you basically put it into a database, and then you get it out again later.
00:20:01
Speaker
and you put it out into, you know, transfer something transforms ah transforms it out of a map into some web page or whatever. um and And that's what data processing meant, you know. And then we've like moved on to this other thing, which is like, you know, essentially number crunching, you know, which is like, okay, actually, I'm not taking data from the web. I'm taking data from spreadsheets and from public data sources and from other bits and pieces. And I'm doing analysis over that mix of data and producing some aggregates and maybe even putting it into some neural nets or and into some machine learning algorithm or
00:20:50
Speaker
we'll come later on to the dreaded AI. um But will wait but for moment for the moment for the moment, let's you know' us hold our fire. days without talking to om we got to Yeah, let's let's hold our fire on that stuff for the moment. But um yeah, so I mean, kind of like what what what do you think like from from from what you've, from your experience is like that transition to like data like aggregation, number crunching, data, what should we say, yeah the data science stuff like versus data processing.
00:21:30
Speaker
Yeah. Do you think closure is like, is standing up in, in, in that field or is the, I mean, it feels to some people that the trend has gone, you know, that it's too late. Yeah. But, um, but it feels like there are some people who've got this like, no, no, no, no, it's still okay. We can still, we can still jump. can catch up Yeah. That's, that's all super interesting. That's ah something I spent a lot of time thinking about. Cause like, right. Like I think in one sense you're right. Like the we're never, you're never going to like catch up to Python as in like overtake the popularity of Python in doing what Python does well. And it's like, like in that sense, for sure, like the overwhelming majority of like data processing happens in Python right now. And there's just so much infrastructure and tooling and stuff built up around that, that it's like,
00:22:24
Speaker
you'll never reach parity. I would say like we have future parity in terms of what's possible to do, but it's like like the ah higher level stuff that's, or the more sort of like complicated or niche stuff that is less easy to catch up. But had this is one of the conversations I had at the conge this past fall, actually with Rich Hakey, which was like super interesting and enlightening, as you would expect it to be, um but about like Yeah, how the goal, I think shouldn't necessarily be to like, catch up to Python or R, but to like, rather leapfrog them and find like, some solution that is uniquely well suited to closure or well implemented in closure, that is kind of like, like solve some of the problems or fill some of the gaps that
00:23:20
Speaker
Um, kind of can't be filled by any other language or sort of things that are uniquely, that closure is uniquely positioned to solve. Um, so I don't know exactly what that is or what that's going to be or look like is the thing, but I can totally understand that perspective. And I think it's really insightful, um, takeaway to be like, yeah, like the goal isn't to just do what Python does, but better or faster or whatever, because that's a very marginal.
00:23:50
Speaker
improvement to anyone. So like if you're, especially if you're already being productive and already working and making money doing this stuff, it's like, why switch to closure? And if the answer is like, because closure is super cool, like that's not going to convince a lot of people.
00:24:03
Speaker
But if the answer is like, I know. like And that that's part of the problem, because closure people are are kind of fanatics. right like people people with Closure developers are very bare like very into closure. And I totally get it, and I totally am too. But it's also, you know you need to have that awareness that like that's not a super convincing argument to somebody who's not already in that mindset.
00:24:26
Speaker
Closure or die is not convincing.
00:24:30
Speaker
like It's not saying where it's like you can't, it's hard to explain to somebody what they don't know, like you don't know what you don't know. And if you haven't experienced the joy of programming and closure, it's really hard to to like experience that or understand that without a having like, without just getting there. And I don't know if, anyway. um But yeah, so.
00:24:52
Speaker
So I think it's more really useful as a community to think in terms of like, what are those, the pain points or the gaps or the, the particularly challenging parts of the Python ecosystem or our, or whatever that are kind of like ripe for disruption or need a solution. And then which of those might be well solved by, by closure. And it's kind of like unique. um, approach to, to data processing and programming in general. So there, there are a few things and a few threads to pull on, but part of it is like, there's nobody who has like full-time resources to spend on stuff like this right now, really, um, which makes it challenging. But, um, yeah, anyway, there's, I think, I think, I think it's, ah I think that's more of like an interesting thread to pull on than like, Oh, how can we like,
00:25:50
Speaker
build all of these things that Python has because you're right. Like you'll just never catch up. Like there's the ecosystem is evolving so quickly and changing so fast that ah it's, uh, it is kind of like a runaway train situation. But it's funny how like it's, it's well,
00:26:08
Speaker
It feels a bit to me as an outsider in this community, and I admit that that is the case, um is that Python seems to suffer from a similar kind of problem to Node.js in the sense that upgrading things and ah you know sort of package management and that kind of stuff seems to be like a sort of like quite a pain that is, I wouldn't say, well, it feels like it's burning the community quite a bit. I don't know if you have any thoughts about that.
00:26:44
Speaker
That's definitely my impression too. Like I think, I think that is one of the big pain points and one of the, the sort of key problems with the, not just the Python ecosystem, but the sort of like modern data stack. So like, if you think of like the kind of like the flow of information through an organization or whatever, it's like, or whatever the, the sort of life cycle of data is like, it starts out as just like a totally like chaotic mass. And then there's this first step of like grabbing that, that data.
00:27:14
Speaker
the raw data and sort of organizing it and like extracting it from all these various sources. So maybe like streams or databases or files or whatever, and kind of like cleaning it up and standardizing it and processing. it And that's, so that's kind of like becoming known as data engineering, um, in like that the world or whatever. And, and then there's, so now you have this like pile of clean, clean data, and then you need to put it, put it somewhere. So usually like a data warehouse or data lake or.
00:27:44
Speaker
a database or whatever, and that's, that's kind of like data ops. And there's this like ecosystem ah evolving around that. And there's all this tooling and like 57 different AWS services and whatever that you can use. And then you have this like data warehouse or whatever. And now, and then it's, and then it goes to the analysts. So there's like data analytics, which you can do like in that database or data warehouse.
00:28:10
Speaker
and um or like kind of moving towards data science which is like kind of a blurry like overlap with data analytics but I think it's like the data science is sort of maybe the more like one way to think about it is the more uh like sort of like so not I don't want to say sophisticated as in the other parts aren't but like that's the more like kind of mathy statistical analysis type of stuff ah people do, but you can only do that on once you have the clean data in the first place, because it's actually like, like a tremendously difficult thing to acquire in any organization. And then after the data scientists, there's like the business analysts or like business intelligence or that kind of family of jobs. And those are people who are like taking these insights from the statisticians and the analysts and like, being like, what does this mean? Like, like as a I could show you a billion in graphs about like, here's your sales per month, or here's your customer profiles or whatever. But
00:29:02
Speaker
you need that overlap with the insights of the domain or the business or whatever to to kind of like extract any meaningful insights out of that. So anyway, all that to say, like right now, that whole like flow of information is like like six different roles and takes like about 1000 different tools. And it's a very fragmented um and complex like ecosystem. And so I think that's one um potential area where closure has a lot to offer because there's
00:29:39
Speaker
There's like a lot of parts of that, that can be kind of unified or simplified or at the very least done like with closure. So it depends on the scale too. Cause like, if you're just working with a little bit of data, you can, you could do all of that with like Python scripts. Um, but usually you're not like, usually in any sort of, in any organization of any size, you're working with pretty large amounts of data.
00:30:06
Speaker
Um, and then you're into the, like the big data world, which is mostly Apache stuff that's written in Java or Scala. And so like all of that is actually much easier to work with from closure than Python.
00:30:18
Speaker
Um, and you very quickly bump into like performance issues, which I guess, yeah, as another tangent note, we're on like tangents of tangents, but like it's fine tangent away, you know, it's math. We're allowed to tangent. yeah Yeah. I would say like another main kind of like pain points, uh, that I've been thinking about, or that may be an interesting fact to pull on. Um,
00:30:44
Speaker
is the sort of like performance issues. So like pretty quickly you run into the limits of what you can accomplish with Python. Once you are working with larger than memory data sets, like like data sets that don't fit in your your RAM. And there are like some solutions for that. And there are new tools coming out all the time. But the thing is like those tools, usually the solution is like get out of Python. So like that's, you see like polars, which is actually implemented in Rust.
00:31:12
Speaker
And there's all of these libraries that are like Rust or Go or C under the hood with this like very thin Python interaction layer. um But that like, that makes the ops overhead like very complex. um So it's not necessarily like, like you can, you can only hide the the underlying language for so long, like at a certain point, you you have to face it. And so, and sometimes these Python libraries are wrapping, yeah, like Spark or Kafka or whatever, um which you can just use their native Java interact like integrations or whatever, um and and then use Closure. So anyway, I think, yeah, that that kind of the current like best practices for for managing this like modern data workflow, in which is like sort of pretty typical in most organizations,
00:32:09
Speaker
is like kind of a mess. Um, and then yeah, nevermind, like the parts, the parts of that flow that are, that are code, like not all of that is code, right? Like some of that is like database administration and like orchestration and like ops stuff, but the parts that are code, it's not that uncommon to see like piles of Python scripts, like, like not really organized, not very like well maintained, like certainly not optimized. Um, and so,
00:32:38
Speaker
Yeah, like I think, yeah, I don't know. It's not an easy problem. It's not like there's like a silver bullet. But Closure has a lot to offer because, like yeah, first of all, the dependencies are much more stable, and it's much easier to manage like these complicated types of um like relationships between libraries in Closure because you can sort of count on like you can count on things more or less not breaking. And like I know people are going to be like, yeah, that's not always true. like Of course things break. It's not like Closure programs have no bugs. But like it's a cultural thing. it's not
00:33:12
Speaker
it's the culture of the community is to like not screw people over it. Like you don't push breaking changes to your libraries and then be like, well, they'll deal with it. Like nobody like significant in the community really does that. And that's, that's kind of, um, where the stability comes from. And so yeah, there's that. And then obviously closure has like, doesn't have the performance constraints of Python being hosted on the JBM. So there's, there's a lot of,
00:33:41
Speaker
There's a lot there to work with. It's just like, is that obvious to me anyway yet? What the like, what the solution is. It's like where there's enough kind of pain to yeah justify that kind of investment in the change. I guess that's what, that's what it's really down to, isn't it? It's like someone somewhere has to think that this is hurting too much. And yeah there is an alternative that I'm aware of that's better.
00:34:09
Speaker
Um, so a lot of it is probably awareness, like a lot of things, you know, um, like, uh, yeah, but which was saying when he was talking about why he even wrote closure was because he wanted to write something in the list. but Um, but he couldn't do it because people.
00:34:32
Speaker
wanted C++ plus plus or they wanted C, yeah or they wanted a Java. you know But he realized that if he could live essentially you Lisp stuff on Java, then guess what? You'd kind of get all the development benefits, but they'd have all the run runtime things and all the kind of like,
00:34:52
Speaker
objections might float away. Now, obviously, closure hasn't went like as big as Python, so it's not been... Those kind of objections still exist, I guess, because you know but a lisp is necessarily a kind of... um Somehow, I don't know why. It seems like very natural to me, but it seems like it's somehow it's kind of counter-cultural for people who've like Learn C or Java, which a lot of people in universities have done in the last 10, 20 years. um And now they're learning Python. yeah So it feels like, kind of like from a, like if your default is Java or Python or JavaScript these days, I guess as well, then it's kind of kind of cultural to go to a lisp. So maybe that's an issue as well.
00:35:49
Speaker
It's definitely not the norm and like just the whole way of thinking about programming. Like I think object oriented programming is like so embedded in the industry, and like so thoroughly now it can be like, like, you know, like I, well, not anymore. I guess now I'm, I'm working full time with broad peak, but before this I was self-employed and doing various consulting things and stuff. And like, yeah, like just even going back to writing like Python or or Java or something like that,
00:36:17
Speaker
It's, it's a kind of jarring, it's like, there's a totally different way about thinking about information and how information flows through a program, right? And so it can be, it's, it's one thing if you go back and forth and you're kind of familiar with both, but if you're like, if you've never worked with a functional language and it's like, that's kind of an ambiguous term, but like, you know, if you don't, if you don't have state, like people, most programmers just like instinctively make classes for everything. And then these classes become dumping grounds for like information. And then this, and then,
00:36:46
Speaker
the information is all wrapped up and it's like i just want like it's just a list of numbers you usually or something like this that you want and that you need to pass around but it's like instead i have to instantiate this thing and call this like these three methods on it to get it into the right state. And then you got to remember what state it's in. And it's like, people are so used to thinking that way. They can't, they don't even understand when you say like, I just need the list. Like I don't care about any of this other stuff. Like it's just, it's just a list of numbers. And it's like a really foreign way of thinking about information. But, but I think that's, that's part of what makes closure such a natural fit for these like data heavy workflows, because like it's inherently like
00:37:29
Speaker
Closure code is data and data is code and that kind of mentality. And it just flows so much more naturally through like, like a functional kind of program where you, you have like your data comes in here and that goes through all these functions and it's just data and data out every step of the way. And you never have to keep track of this intermediate state of things. You never have to worry about how things got into this or that state, but I don't know. It's a.
00:37:59
Speaker
It's not intuitive, I would say to like an average, like, like normal, whatever person, however you think about it. ah Normal person. What's that like? Yeah, I have never been normal. I couldn't tell you.
00:38:16
Speaker
I think there's there's definitely something to what you're saying about like this data data first approach of closure being well-suited to data processing problems, surprise, surprise.
00:38:28
Speaker
but i think like It struck me, I don't know, 10 or however many years ago that as a programmer these days, all we're really doing is we're like getting some data from a web request, massaging it in some way, putting it on some wire or in some database as Ray said. so like and all really like The core of what most of us working programmers are doing is data transformation, right? so we're not like english We're not doing what programmers 25 years ago did where they were you know writing C or whatever and really cared a lot about state. and you know the
00:39:07
Speaker
state the machine was in and having to be really efficient with all of this stuff. Like these days, you know we got we got cycles to burn and thinking thinking about problems or sorry, programs is a series of data. Jesus. Sorry, we've got a world to burn. Yeah, yeah yeah but yeah no, I think that's a really insightful point. like in ah In a lot of ways, the kind of dominant paradigms in the industry, I think, are very much a product of a former era where like storage was very expensive, compute was very expensive. And so you have all this ah stuff around. And you know I think that's, I was talking to a sort of colleague about like,
00:39:53
Speaker
Yeah. Data engineering stuff the other day. And it came up in that context too. Like just, just the way a lot of code works these days is, is a product of that time where storage was very expensive. And so you're constantly like mutating things and trying to very conserved space. And it's like storage is effectively infinite now. It's like, it's very cheap and very easy to start things. And it's, and like, there's still, you want it, you know, their compute can cost money and it depends what you're working with, but.
00:40:19
Speaker
Like you're saying, like relatively speaking, like it's not worth, compared to the cost of engineers, like especially your compute bill is probably negligible. And so it's like, if you can save one software engineer one hour of time, it's not worth worrying about like your extra whatever, $100 of S3 storage per year or whatever.
00:40:42
Speaker
Um, or like your database replica that's, that's taking up a little bit of space. And so, but it's like this, these patterns and these ways of thinking that were developed in this era where you had to be like extremely conservative about, about resources. Like it's not, it's not a bad thing. You want to be space efficient and like compute efficient, but that's, that's yet another thing. I'm like, I know this is a closure podcast and I'm a total closure fan girl, but it's like.
00:41:06
Speaker
Closure is good for that. like um Chris Nuremberg talks about that in a lot of his talks. um yeah like All of this high performance computing stuff. Closure legitimately is best in class and a big reason to use it is that you're you burning less energy to get through the same amount of processing.
00:41:22
Speaker
and um And that's a huge benefit. But yeah, conversely, it's like, it's also, but that's, yeah. Anyway, this is another thing just across time is like the immutable data structures. Like, because often I just think, you know, often conversations with people when you talk about immutable data and you got to not mutate things, they're like, Oh, but it's so wasteful. It's all this space. And it's like,
00:41:43
Speaker
It's because they've never heard of trees or tries or how you pronounce them. And like this concept of like these, these persistent immutable data structures are actually pretty efficient. It's, they're not, you're not storing a million copies of all of your data all over the place and wasting, you know, like wasting space or memory.
00:42:02
Speaker
Um, and yeah, and it's, it's obviously not as like compact as, as humanly possible, but, but yeah, like we have these very unique, um, language features that are just like, like that especially is not available. Like the way immutability is possible in other languages, but it's implemented very differently yeah and, and does it just doesn't fundamentally work the same way. And so it's like, we can, we can really take advantage of that to to accomplish some of these goals.
00:42:32
Speaker
Yeah, I mean, I think algorithms and data structures seem to be you like the meat and gravy of what we're doing here. So talking of which, because um again, if I understand it right, um a lot of the, like like you were saying before, a lot of the fundamental kind of algorithms and data structures for a lot of this stuff is written in C or it's written in C++ plus plus or Rust. So in other words, there are kind of like a lot of these people in the end, they drop to kind of low level um algorithms and data structures. um So it's kind of peculiar how Python has become so popular.
00:43:16
Speaker
other than the fact that I guess it's just more like the cobalt of today that it's somehow become, ah yeah, I mean, I'm kind of joking, but I mean, it's sort of like, it's become, it's like approachable, you know, that it's sort of feels like it's fairly easy to write and it's, yeah, I think that's a big part of that. you know Yeah. Yeah. Well, I mean, I think that's, that may be more or less the story is like, it is, it is very kind of like ergonomic and.
00:43:46
Speaker
like yeah whatever developer-friendly and learner-friendly. And you can imagine, because this is the other thing, is that most data people, however you think of them, data engineers or data scientists or data analysts or business intelligence people, are not software engineers primarily. And so they're right they are writing code, but they're like this is, as a software engineer, iss like this is a category of people you just often don't even think of it's like what do you mean there are people who write code who are in software engineers and it's like there's actually a lot of people who write code who are not software engineers and so they like maybe don't even know about get they don't know about
00:44:19
Speaker
like file structuring best practices or dependency management or all these things that are just table stakes and software engineering that we take for granted. I mean, now, like 20 years ago, you could find people who were like emailing files back and forth and, you know, renaming them and like final, final, especially final version, final, final, final for real. This is actually the final one. Please don't delete this. Seriously. Last one. Don't delete this. But like, luckily, mercifully, we have,
00:44:49
Speaker
mostly moved past that in our industry. But yeah, like, there are if you if you just dropped like a statistician into like a C program environment, they would it would just be like a huge mountain to climb compared to a much higher level language like Python, where it's like pretty straightforward to get started.
00:45:10
Speaker
And you can, you can legitimately be quite productive without really knowing that much about like the fundamentals of software engineering. Um, but I think, you know, that's, that's something we have in closures like closures is a, it's, it's a different language. It's a different way of thinking about programming, like we mentioned before, but it's still a very ergonomic language. Like it's very high level. You don't have to care about like.
00:45:34
Speaker
memory management or data structures or whatever, and you can be productive. And that's, and that's part of what the Cyclos community is working a lot on right now is like, it's kind of trying to like build another layer.
Advancements in Data Visualization
00:45:47
Speaker
So we have a lot of the very, like the fundamentals are there, they're like high performance computing libraries and the numerics and.
00:45:54
Speaker
the math and the machine learning and stuff like that. Yeah, that's Ray just wrote, R speaks to statisticians. How can we speak to them? Oh, you're me up. Thanks a lot. Oh, sorry, I blew it. I wasn't supposed to hide the chat. world, Kira, come on. Hello. You're supposed to let them know that we're- I will fix it in a post. Yeah, yeah. But anyway- Ours speaks to statisticians. you know How do we speak to them? Yeah. yeah i mean yeah So this is this is what this is what it's all about is in in whatever the work that but Daniel and I and the others are doing, trying to build these tools in these libraries in the Cyclos community. And sometimes this is ah this is one of the challenging things is like software people don't always understand. If you show a software engineer a certain library or whatever, they'll be like, yeah, that's totally intuitive. like Why do we need to build this? This is a huge waste of time.
00:46:52
Speaker
And it's like, no, but it's not for you. Right. It's not for you. Like, and Hadley Wickham has written a lot about this. Well, wrote a lot about this like 10 years ago, he was saying, um, Hadley, so Hadley Wickham is kind of like one of the very prolific, um, authors of a lot of our packages and that, in that ecosystem are for statistics and data science and data visualization and stuff like that. He's like the.
00:47:17
Speaker
Yeah, like the bork dude of their world over just like constantly producing so many very useful things. Um, and anyway, yeah, he was about 10 years ago wrote in a paper, like there are all these tools and they're all these really great, like ways of doing things and and libraries to accomplish these, these tasks, but they're very oriented towards software engineers. Like these very kind of like.
00:47:42
Speaker
low level APIs and sort of complicated config and stuff like that. And what we really need is to, to simplify all then abstract all of that away. And I think. that That's sometimes a tough sell. Like I was saying, like software people don't always understand the value of making things like not just simple, but also easy. Um, i'm going take a drink. All of you listening at home, you know, if you want to, if that's, um,
00:48:16
Speaker
I think that's, that's one, and anyway, that's one of the things, but there, is there is some element of like, It's, it's really hard because there's, there's very limited resources in the community. There's very limited time and energy and very few people who are like actively working on this stuff. And so there's this question of like, what should we be pursuing and what is the priority? And that's, that's something I really struggle with because like my, the the temptation is very strong to just like mimic.
00:48:43
Speaker
what's out there, right because it's easier. It's like as much very obvious to look, if you're if you've ever worked with our Python, it's like immediately obvious what's missing. If you you pick up a Closure data science project and you're like, oh, I can't do this. And so you're like, oh, I'm going to implement that in Closure. And now I can do this thing in Closure. um And that's one like path to pursue.
00:49:05
Speaker
for for developing the ecosystem. And I do think it's valuable. And I certainly don't want to like want to like knock all of the like amazing people, including myself, who are working on that type of stuff. um But there is that thing where it's like the goal isn't necessarily or maybe shouldn't be exactly only feature parity with the R and Python ecosystems. But then that raises the much more difficult question to answer, which is like, so what is the goal? And like, what should we be working on? And that's the whole, like, how do we leapfrog that entire ecosystem? Kind of like, kind of like Datomic, like, it's like Datomic wasn't just like a better Postgres. It's just like a fundamentally different way of thinking about storing data. It's like, you can't, if you're using Datomic, it's like, you can't use anything else. It's, it's uniquely suited to solve this, this type of, this type of data storage need. And so it's like,
00:49:59
Speaker
That's kind of what I would like to figure out is like, what is the like analogous type of like killer app or killer feature in like data engineering or data science that we could implement for, for our ecosystem.
00:50:16
Speaker
that it wouldn't be just like a better option. It would just be like people who need to do X can only do it with, with this awesome new closure tool or this awesome new tool that happens to be built in closure, which like most people who use it won't even care about because they will not be like looking into the implementation or whatever. It's just like, uh,
00:50:37
Speaker
a side effect or whatever consequently provided. Oh, we don't. Oh, no, we do. That's right. we're We're not Haskell people. Closer people are fine. Yeah. no side effects allowed yeah It was, um, storm was like that, right? Like, um, storm was written in closure, but like the author, I can't remember the person's name now, but like they didn't. Oh, was it Nathan? Yeah. Okay. Well that explains it. There's a through line there. Okay. Yeah. Yeah. But like, I don't think he really like made a big deal of the fact that it was written in closure at all. Like, uh, it had Java, you know, SDK and everything. So only closure people knew that it was written in closure.
00:51:18
Speaker
And I think like that that's kind of you know i think that exactly the right way to go about it if you're trying to convince people to use your stuff. It's just like to to find a way to solve a problem that there isn't a good solution to today. right it's like what and it like that's That's the million dollar question. right is like That's my grand vision or whatever for all of this. is like like make something useful basically is like that's just like all I want to do like I feel like most of the software I've ever written is just like in the garbage now and it's like but just kind of a bummer. That's success though isn't it right? like
00:52:03
Speaker
sad sure how you just How are you like a calibrating success here, Josh? Well, because i I no longer need that stuff. You see, I have i've transcended. Oh, it's a Buddhist way. No, I love throwing software away. It's my favorite thing. It is true. It's very satisfying. Yeah. But yeah, that's, that's a good point. Like I think ultimately is what you want is like something useful and that's like certain people will be you know like you can nerd out about it at the conge or whatever and talk about it but but the majority of users won't care what it's written and it'll just be like really useful to them and anyway that's that's the goal i think just a quickie on that one then sorry josh no go ahead re no please
00:52:50
Speaker
This is the, we have a Canadian guest on, so we're getting like super polite here in Devon. This is, you know. like circle of politeness this this I've been in situations where this goes on for like five minutes and everyone's like, no, no, no, you. Or like at a four-way stop, is sometimes you just get gridlocked for like two minutes. And it's like, no, you go, no, you go. And then a pedestrian shows up and then what? It's like, we're all just waiting for each other. Somebody's got to go.
00:53:15
Speaker
Yeah. Well, people find me useful in Sweden because I'm from the US and they're like, okay, but this guy will be a dickhead and just go ahead and you know just do the thing that nobody else. Go ahead, Josh. It's not rude. It's my birthright.
00:53:33
Speaker
But we were talking exactly. so
00:53:38
Speaker
Oh, Lord. Okay. Where were we? and Yes, the killer app. No, I was just wondering because one of the things you wrote about after the conge was this grammar for graphics, right? and You were super excited about SVGs, which I would have found very strange, except surprisingly enough, I too am super excited about SVGs when I realized they're just data. And that's pretty cool. But so talk about the talk about this grammar for graphics. like like What is it? Why do we need it? Why are you so excited about it, et cetera?
00:54:12
Speaker
Sure. Yeah. I mean, part of it is just kind of like, it's my personal hobby horse or whatever, and has been for a very long time. I've been interested in data visualization and like, um, kind of how to, yeah, how to represent information visually in like a meaningful way. But I think, I also think it's one of the pieces that we're kind of missing in the ecosystem is like a, a robust, um,
00:54:37
Speaker
kind of closurey way to visualize graphics. Um, cause it's, it's such an essential part of the data workflow. Like as, as you're working with data, part, like part of the initial steps almost always involved this like kind of exploration phase or whatever, where you're just like poking around and like, how does this relate to that? And what does this look like plotted against that? And if you, if you observe like a data scientist,
00:55:02
Speaker
doing their job or a statistician, they're just constantly like writing little things, checking, see what it looks like throwing it away, writing little things, seeing what it looks like throwing it away. And um anyway, yeah. So we there are there are ways to do that in Closure, obviously, but they all leverage javascript various JavaScript libraries on the front end. And Daniel has been implementing a lot of really amazing um stuff, kind of kind of leveraging those, but doing the ah grammar part in a Closure-y way.
00:55:30
Speaker
But anyway, then yeah, Timothy Pratly and Chris Hauser at the last conch had this super fun like whimsical talk about SVGs, among other things. And it made me realize like maybe those, maybe SVG could be a really good solution for the graphics part of it because I think, yeah, we're we're making a lot of progress. I say we as in as if I'm contributing anything, but it's it's it's all Daniel. and We have some conversations about it sometimes, but but he's done ah just a tremendous amount of work there. um Among other people, certainly lots of other people have helped us here, but
00:56:07
Speaker
But yeah, the so the grammar part I think is is more settled. And so, yeah, so I guess if anyway for those who aren't familiar, the grammar of graphics is like a sort of way of thinking about data visualization in terms of like gradually, semantically building up graphics or like plots or images, like a visual thing out of data and that is like layered and incremental as opposed to like describing a ah plot.
00:56:38
Speaker
like picking a plot from a library. So a grammar of graphics library is not just like a collection of charts. Like if most visualization libraries are basically a collection of charts, it's like, this is how you make a pie chart. This is how you make a bar chart. This is how you make a scatter plot. This is how you make a whatever plot. And then if you want to do something that's not one of those plots, you basically can't. But the grammar of graphics is like a different approach. It's like saying add like a dot layer and then add a line layer and add a,
00:57:08
Speaker
whatever type of scale and a whatever type of legend. And so you can build it up piece by piece in kind of like a DSL almost. That is like, ah like kind of a language if depending how you think of languages, but it's almost like a language for describing graphics and there are there are like analogous things to verbs, analogous components to like nouns and, and that kind of thing. So anyway, so that is all, there's, there's a book called the grammar of graphics by Leland Wilkinson. And there's a bunch of, um, prior art by Hadley Wickham, who has also the kind of canonical implementation, which is called, t flat two as we know yeah yeah yeah yeah. Um, so that, that part of it, the kind of like, how do you describe the graphic is, is like,
00:57:55
Speaker
There's a lot of prior art there, and there's a lot of inspiration to draw from, and and already a pretty good closure implementation. But then there's this thing of like once you have the plot described, you have to actually render it, like turn it into pixels on the screen. And so that's the part right now. Basically, we're just like shipping that all off to JavaScript, like either Plotly or VegaLite or Echarts or these various um frontend libraries that are supported. And it's at that point, it's just like a data munging thing. It's like you can turn this, this day, this specification that you generated with closure into
00:58:31
Speaker
the shape of data that any of those libraries accepts. But like what I think would be cool is to have a pure JVM implementation of that. Part of it is, part of it's just, again, like there's, this is what I struggle with is like, I don't know if there's a good reason to do this. And this is, this is my one project that I let myself work on anyway. It's like, like at the last conch, Rich had a ah little brief speech about like, do things that are fun. Like don't just only code for money or work. Cause then you'll start hitting it.
00:59:01
Speaker
um which I think is really good advice. And so it's like, this is okay. Like this is maybe totally useless, but it's actually just really fun and enjoyable. So I'm just gonna do it anyway. but um But anyway, I do think there would be benefits to having a JVM implementation, mainly like you're you don't you remove this dependency on on like the browser or any specific JavaScript library. And so you can generate these like to put them in a report or on your own computer or whatever.
00:59:30
Speaker
And then you're also, um, like you have the the full JBM at your disposal, so you can do kind of like more heavy lifting, more like statistical analysis, or make use of any of these, these statistical libraries that we have for data enclosure, um, in the graphics rendering thing. So there are like certain visualizations that require some specific ah statistical on them that like you can do in advance. You can you can like obviously process your data however you want before you pass it to the biz library. But sometimes it's like, if you're just playing around and you want to see like a histogram, for example, you have there you have to do some like aggregations and stuff there. And right now, that will happen in JavaScript, which can be really slow for very large data sets. And there's like other downsides. So anyway, yeah, thinking about how to
01:00:25
Speaker
how to do all of that. Like, I mean, the other thing is if you want to, you want to render an image, right? You're limited to like image formats. And so that's like vector raster. And then which basically means like SVGs or like canvas. And I started going on the canvas kind of rabbit hole, but then it's, it's, it's truly a rabbit hole. It's kind of kind of a mess. And there's, I think like, yeah, a lot of other benefits to, to,
01:00:55
Speaker
pursuing SVGs. So yeah, because, because like you said, they're just, they're just data and they can also browsers can render them. So you, it would be, I think much easier to make them like accessible um with like adding all of the relevant tags and and stuff like that. You could probably add, like have JavaScript free ones and then augment them with Java or JavaScript to make them interactive in a much simpler way than like anything that would be possible with a canvas.
01:01:25
Speaker
So there are some avenues there that I think would be really interesting to pursue. And that's that is one area that's kind of like, ah whatever, still under active development. Like on this the sort of like cutting edge of of data visualization right now is like interactive graphics, I think are not a solved problem. Like I would say like graphics rendering is probably arguably solved and it's like maybe futile to pursue yet another implementation of that, but it's like one, one could do it if, if one were like crazy, but, um, but I think arguably interactive graphics are not like, there's not like a universally accepted best way to do it at least. Um, even though there are solutions, like you can obviously do and interactive graphics. We've all seen them, but it's like, yeah there's, there's not like an ideal
01:02:18
Speaker
ah ideally necessarily like super easy way to do them. So then I think we just found your killer app, right? Like that's a just do it. karro like you yeah Yeah, no problem.
01:02:33
Speaker
You know, um the the thing about graphics is that obviously, well, they're optimized for different things these days, it seems, but um they were originally optimized for like game rendering and stuff like that, um which is basically a visualization problem, um you know, with people shooting each other and all these great things. What's the last game you've played, Ray?
01:02:59
Speaker
Murder. Yeah, I don't know. ah Really, I don't know. i've I've not played any games, almost. So, yeah I don't play games. But um but I've seen people on... on i've seen The damn kids on the damn internets. Yeah, my kids, they play games. I could say League of Legends, you know. That probably is already a bit kind of old. I wouldn't say Fortnite, because I know that's douche, but yeah.
01:03:27
Speaker
ah but You've just offended half of our listenership now. yeah true Yeah, but I think what I was going to say was that I wonder whether there is a sort of Like are these statistical packages all kind of like fairly flat, um, like renditions in 2d space or are you looking at, you know, 3d spaces, uh, explorable spaces, you know, with VR ah and stuff like that, even, you know, that's super interesting. I basically, no, I had not.
01:04:07
Speaker
really thought about 3D like there are some you can like simulate like you can have three-dimensional graphics or whatever like ones with an x, y, and z or something like that but yeah no in terms of actually rendering graphics in a 3D world that reminds me of uh this is a this is a side survey but it's kind of funny one ah in a very former life, like three jobs ago, I was working on, uh, like a 3d rendering engine for, um, LiDAR data annotation. So it was like one of these companies that would farm data out to like third world countries and, um, like hire people to circle faces of people or cats or whatever. So you could like annotate the data for training, um, for vehicles or yep, yep, yep. So we had this thing where you could like, you can upload a photo and then you could either like draw things on it or you could
01:04:54
Speaker
like draw color pixels or whatever. And then it was like, well, now we want videos. And then we want LiDAR videos. Anyway, so um but in one of the ah product planning meetings. You know, people were like, what about this? And what if we want this? What if we want that? And one of the other engineers was like, we can build you whatever you want. Like if you want a 3d world, if you want a 3d world where you wade through and take your data off a shelf, we can build that. You just need to decide what you want. Anyway, I thought that was, I was like, I'm picturing this. Does it have legs though is the question. Yeah. 3d world. That's a meta joke. I'm sorry. That's a metaverse joke. Yes.
01:05:32
Speaker
you're waitinging through your data. But yeah, that it would be interesting. But anyway, basically, yeah, no, I have not considered pursuing that. That was a real pain. At that time, we were using 3JS as like the 3D rendering thing, because it was all in ah in a browser. So it's, it's, ah it's super powerful. But probably these days, there's, there's something better i G.L. has become a bit more standardized and a bit more you know gives a bit more access to the underlying hardware. and it was all Because it was a whole G.L. war, wasn't there? right and There probably is still. i mean you i'm I'm not really aware of the G.L. war situation.
01:06:11
Speaker
um But I know there was a sort of standardization effort around um WebGL. That's super interesting. I don't know how far it is yet, but but this concept of being able to you know render into a 3D environment via very, very capable hardware seems quite interesting. It does totally make sense to leverage like the purpose-built hardware for graphics rendering to do that type of intense like visualization because yeah, for sure that was the main problem is you just, you very quickly bump up against the limitations of, of like what a browser can process. Um, if we crashed many, many browser tabs, try to make make that work and.
01:06:56
Speaker
Because the funny thing about it is it's kind of like top and tail to me, because a lot of the the bottom of data and data and art analytics that Chris is talking about, for instance, is using CUDA or something similar to essentially do matrix multiplication, right which is all sitting on a graphics card. And then if you want to get to the top level, you know, with the visualization, well, it's a graphics card, you know?
01:07:20
Speaker
there oh yeah yeah the thing that it is supposedly for before people started using them to mine bitcoins or whatever for this.
01:07:31
Speaker
Yeah. We're not going to talk about LLMs, right? Okay. yeah These things are really fast at doing that. yeah Yeah. We've already been going for over an hour. I think, you know, we we could, we, maybe what we should do is like, I don't know if you're up to, to up to a chat about LLMs, but, uh, cause I'm guessing that a lot of people are excited about getting into LLMs. He says, well, I know they are. So, um, does c closure have like a compelling LLM story?
01:08:01
Speaker
That's a, yeah, no, that's totally fair question. I think, I mean, the answer is like, sort of there. are So there are, I guess it depends on you might getting into LLMs. Like, I think there's a lot of people who have these, like, like,
01:08:13
Speaker
what probably most people interact with LLMs, like use them for things. I'm thinking about startups, you know, who want to like use an LLM to train a model and, you know, maybe they want to use an existing model or they want to, you know, I don't know enough about the, the, the kind of state of the art, but I know there's, there's lots of startups basically trying to look at LLMs that are using, they're trying to use things in a different way or a more efficient way or a more, uh, like same way, like the vegan LLM type idea, you know, where you kind of like, you know, you consume, uh, ethical data or something, or at least you consume the business owns, you know, and maybe it's like, uh, the way the software stuff. You mean like training, training your own large language model.
01:09:13
Speaker
But like, or like using the same approach, but I like, I would be like a much smaller thing or something like that. Yeah. I mean, I guess to be honest, I haven't really been following, there is a conversation going on in, in the. Closure in Zulip instance about, um, LLMs and there are a few libraries coming up that I think.
01:09:30
Speaker
facilitate that type of work. And there are certainly ways to like interact with existing LLMs. So if like if you have an external model you want to use from Clojure, that's for sure possible. um But yeah, i would I would think it's probably pretty early days in terms of like training an LLM. And yeah, that's an interesting one, because it's like that's it's probably not too late to jump on that hype wagon. The other thing, though, is a lot of people in the Clojure community are kind of like averse to hype.
01:09:59
Speaker
like by nature and so it's like LLMs like that's pretty much the standard response and so um There are definitely people there are definitely people working on it interested in it but I would say it's not as like my general sense in the community is like There it's not that they're dismissive like I think people people recognize that they're very powerful but people are more interested in like how can we like leverage this to be more productive and useful and not necessarily like not not as as bullish on the LLM train as as the average person in the industry might be, I would say. and Or maybe that's maybe I'm just projecting. That's kind of how I feel, I guess. I definitely find them useful. And I think like it's I have come around. like ah like Two years ago, I was like, these are a huge waste of time and energy, and this is ridiculous. But now like I can see the utility. but um
01:10:53
Speaker
Yeah, I have not like played around with you know training any or or developing any like anything like that. So what made you come around? I think, I mean, they just got really good. Like, I guess this is, this is sort of related to like one of the other things I wrote about in that, like conjureflection was this thing about like software engine. I think the role of software engineers is, is really already changing, but definitely will change more and more as these LLMs get better and better. Cause it's like, at this point they can kind of write code, like not super well and not like perfectly, but there's, there's like,
01:11:29
Speaker
less value in being able to just like churn out boilerplate now than there was before, for sure. But like what they can't do and probably never will be able to do is the like thinking part and the designing part. And so this whole thing of like like enabling software engineers to just like think about problems and design solutions and architect systems, I think is like it's difficult because it's kind of like what we were talking about before the show. like it's Sometimes that doesn't look like work. or it doesn't like it doesn't seem productive, but like coming up with the right solution to a problem can be incredibly valuable, but there's it's it's kind of like inherently at odds with, like I don't know, the
01:12:18
Speaker
like the nature of the industry or whatever, or like the, the sort of expectations of, of an organization. Like, I mean, I don't know. I've been, I've been pretty lucky. Like I don't feel like I'm under a ton of pressure constantly to like produce code or whatever, or like churn out lines. But there's definitely this like sense that the like thinking part is like.
01:12:37
Speaker
supposed to be pretty quick, and then you should start writing code. And it's like, you know, there's no hammocks for in this, in this world. Yeah, not, not in America. Anyway, Americans aren't big on hammock time. I think it's like, they're very like productivity minded.
01:12:53
Speaker
Yeah. um that but i mean I am wincing because I agree with you so much, but i mean that that is the stupidest way to think about productivity because like productivity should be about like solving problems, not right know not generating actions, whether those actions are lines of code or You know, whatever. So it's just it's just infuriating that um people get so hung up on things like I mean, everybody knows that like lines of code is not a good way to measure program or productivity. And yet people still fucking do that.
01:13:31
Speaker
Like maybe not, they maybe they don't say it like that, right? That's the thing. Like, even if they're not literally counting lines of code, there's definitely a vibe in every organization. Like so-and-so opens so many pull requests, they're so productive. And it's like, yeah that's not necessarily true. Like I've definitely been on teams where like the person writing the most code was just causing the most problems. It was like, just stop. like That's so almost always true though, right? Code causes problems.
01:13:57
Speaker
I know, that's the thing is like less code is less code is less problems. You want less code. The thing is, that like when I first started, back in like 1453.
01:14:08
Speaker
yeah
01:14:11
Speaker
I was going to say 1952, but you know, whatever. yeah generally Well, you know, it was pre-capitalism. um a good time. It was a good times good time. with if How do you know he's the king?
01:14:27
Speaker
yeah Yeah, I was flip-flopping between Catholicism and Protestantism on a regular basis. a size
01:14:36
Speaker
Right. Okay. So, but yeah, when I first started, there was this sort of like, you know, you were talking before about like this workflow of things in the data science world where you have you know the the data cleansing and then the ops and then the engineering. of them I've already forgotten what half the half half of them were. Data cleansing doesn't sound like yeah you need a priest for it though, right? yeah know Or an exorcist. something yeah Or it's a beautician type activity. you know There you go. But anyway, what I was going to say it was that when when I was starting, there was a sort of like this concept of
01:15:14
Speaker
ah like a project manager who wasn't a project manager like we have now. He was like an architect, essentially, who defined the project. And then there was what was called an analyst. Business analyst, yeah.
01:15:26
Speaker
yeah Well, yeah, a business analyst. Or a systems analyst. Sorry, that's what they were called. So there was, again, this division of labor, this Taylorist view of a software factory where you have the people doing the blueprints and all this kind of stuff. And then the programming was people that were pejoratively called codemonkeys.
01:15:47
Speaker
who were you know it's like The funny thing is, that this stochastic parrot model of AI where you know you just basically put a bunch of LLMs there and ah given a spec or so-called prompt, they will just basically write all the code, mostly naty right know But I remember there was, I'm kind of off a tangent to some extent here, but yeah forgive me. There was this band in the, ah you know about this ah this concept of the infinite monkeys, like we'll eventually write the works of Shakespeare.
Role of LLMs in Development
01:16:21
Speaker
And this is like dream that is clearly bullshit. It was just a concept of infinity. But there's just there's this dream in me in the sort of LLM world, it feels to me like, Oh, well, actually, given enough LLMs, we are going to write the works of Shakespeare and solve all physics according to some ultimate, you know? oh god So, you know, it's just absolute nonsense.
01:16:41
Speaker
you know but But this idea again that kind of like you you put an LLM where a coder is, you know and they're just going to be productive yeah because because all of that all that thinking stuff is in the prompt.
01:16:56
Speaker
And all of the boring stuff is in writing the code. And there's someone who writes code for a living. i I find that mildly offensive, but also, but I also see that I cause I write closure code. I see that other languages have more boilerplate that there could be that the sort of like, and again, I'm not shitting on those little other other languages, but I see the opportunity for kind of autocomplete and other languages to be more powerful. It looks more impressive as a demo.
01:17:25
Speaker
if you can kind of fill in like you know um the the bits of Java boilerplate or the bits of JavaScript boilerplate or whatever that that come as a must-have. So it's all to say that that I feel like now in this world where the roles are being munished together. And I think most developers actually are kind of architects and analysts and stuff like that, because you this idea of having a separate role that does all the thinking um is right proven to be absolutely bullshit. you know It's like communist Russia or something, where all the ideas come from Moscow. you know well it just know Not that I'm hatching on communism, but that particular version of it
01:18:09
Speaker
um That particular version of it where all the ideas come from one place is obviously. right And just when that when the person designing the solution is so disconnected from the implementation, it just like invariably breaks down. like yeah I've been on teams like that before too, where there was like there's like a tech lead and this person architects the system, and then it's like, you just go build this. And it's like, first of all, being the person who has to go build, that really sucks. It's like super demoralizing and soul-crushing to be like given a very specific spec down to like how to organize your files.
01:18:39
Speaker
And then go do it. Like that's just not a job that any intelligent programmer wants. And, but also like it mostly just doesn't work. Like it's this, this whole thing of like, like that's, that goes back to the sort of like management styles or whatever too is, is like you, you, you have to be okay. Like people, there are always many ways to solve a problem, right? Like there's always multiple implementations that will get you to the right answer. And it's like,
01:19:05
Speaker
Just like getting everybody to coalesce around like a single style or it's not like style guides are useless. Like I get that there's some value in having a coherent code base or whatever, but like just like bike shedding about those types of nitty gritty details is like just a total waste of time. Like if you at the end of the day, right? You got to stay focused on like, does it work? And if so, let's move on with our lives. Like let's not spend two weeks rewriting it so that it it looks better to my eyes or whatever. Like.
01:19:33
Speaker
Yeah. And that's, anyway, that's what I find like a lot of times those like architect roles tend to end up being much more like, like anyone who's capable of implementing a system is, is more or less capable of like designing it and thinking about it and reasoning about it too. And so it's like, all you're doing is like, like forcing me to write my code in your like idiosyncratic style, which is just not a value add. Like it doesn't make any sense.
01:19:58
Speaker
And it's just a, je ah it's basically an argument generating factory as well, because oh yeah totally it's just a personality clash guarantee, isn't it? Where someone is saying, this is how you should write your code. And you're like, well, I'm doing it. yeah And, uh, it doesn't seem, it seems like, it seems like it's not the best way. And then you're just arguing for a few days about, well,
01:20:19
Speaker
Yeah. Why isn't it the best way? Look, I'm a more senior person, do it this way. And then yeah yeah exactly then you're kind of like, then you're in the autocracy, then you're in the autocracy, aren't you? You know, it's just like, you know, this is not a nice place to be. That's the thing. That's not a good, it's not a good team dynamic at all for sure. Well,
01:20:39
Speaker
don't worry, we can just have the LLMs argue with each other about that stuff. so No problem. We've automated. No, but I don't want to dive too deep into this, but um I don't know if you all read the Closureverse RSS feed or but Planet Closures, sorry. um But there was a pretty interesting one and and Ray, you and I listened to the Juxtcast about this. um where ah they were talking about LLMs being used for programming like this. And like one interesting thing in there was about exploring options spaces with LLMs, which you know I actually, but that is one of the most compelling arguments that I've heard for any of this stuff. It's just like, I want to quickly understand what doing this in multiple different ways would actually look like.
01:21:33
Speaker
And, um, you know, I think like, I, sorry, I completely have no idea why I said that, but, oh yeah, having they all, I'm sorry to argue with each other. Yeah. Yeah. Yeah. No, I mean, I think at that point, Josh, Oh yeah, I was. Yes. Yes. lost the thread yeah Sorry, tangent to tangent to tangent. I was on the arc tangent, obviously. i Yeah, that sounds interesting. I should listen to ah to that conversation. But yeah, I mean, that checks out. That's been my experience so far too is like, I find them really useful for sort of like brainstorming or like kind of bouncing ideas around or like, How can we do this? How can we do that? And then like i it's it's true in my experience, and I totally sympathize with people who are like, these are useless because they can't write code that works. And it's like ah think that like it's true that most of the code they write like doesn't work, but that also doesn't make them useless. like You're saying like you can tell right as a senior engineer, if you glance at some code, I can tell i could tell you it's not going to work. But I can also be like, oh, that's an interesting way like to maybe pursue or like an an interesting kind of like path to to follow. And so I think
01:22:39
Speaker
that's kind of how i use them is like well like i mean at work so far, they were they're still ah the the kind of verdicts at whether we're allowed to use them for work. But on my side projects, I am, yeah, we'll kind of like, be like, well, you know, I'm trying to do like, whatever, a grammar, graphics, enclosure, and here's like a billion examples from other languages. And like, you know, what might be some ways to to like, think about it in in a more functional way or something like that. And just like, on those types of things where you're working alone, and you're like, just kind of like brainstorming or bouncing ideas back and forth. I think it's just like,
01:23:13
Speaker
It just like accelerates things. It's not like it really is doing anything for you necessarily. It's just like, it's like, it's like processing faster or something like that. Like, like it just, I feel like it just makes me like much more productive and faster and that.
01:23:27
Speaker
But it's it's not like it it can do what I do like at the end of the day. so And sometimes too, it can be frustrating because you like you waste like an hour and you're like, okay, all these ideas are stupid. That was a huge waste of time. But ah then in those times, I wonder if like, would it have taken me like 20 hours to get to that same conclusion? and Like if I had to think through all of this myself, but um yeah, that's it's interesting.
01:23:50
Speaker
This is interesting. I mean, I definitely have the same feeling, you know, when I'm pairing or, you know, programming in a group with people, you know, just that idea of like, Oh my God, this is like a superpower. So it is, ah yeah it is very interesting. Like when you are forced to work by yourself, like, right I mean, all of the stupid stuff that I write for myself, I just feel like, Oh, this is taking me so long. Why is this taking me so long? yeah And so i just like I really like programming as a social activity. ah like I think the the outcomes are better, like the the code is better because you've thought through it from multiple angles, and but also it's just you know it's ah it's just more fun to to work with another human being. so this is like
01:24:37
Speaker
Yeah. I don't know.
Productivity and Learning Critique
01:24:38
Speaker
This is something I kind of think about as a remote worker. Oh yeah. I mean, I think a lot of us are in that book because yeah, I work a hundred percent remotely. I live in like in the middle of nowhere. And that's the thing like at work, it's like you have that real time chat or like often you'll just hop on a video call or whatever. And that experience of yeah, like thinking through something together or just even pairing and doing something together is super satisfying to me too. like it's And it does feel like way more productive, even but even though it's like now it feels like you're taking, like or whatever, from the business perspective, you might be taking twice as long because you've got two people on it. But it's like it would have taken me, for sure, like way more than twice as long to do like what you can do with two people at once.
01:25:21
Speaker
right um do think thats me slightly is a sort of like I don't know. I've got a sort of, um, like back in the day, back in the 15th century, we read a lot of books, you know, and we couldn't have any, well, actually 15th century, we didn't, because the got the Gutenberg press wasn't invented then. But let's say 18th century, you know, we used to read a lot of books and, uh, but, um, the thing I find from reading books as well is that You kind of get insights or thoughts that you can't get just from absorbing summaries. You know, you can't have that experience. And that does worry me these days that, yeah, you can grok a lot of knowledge in summary form.
01:26:11
Speaker
But it doesn't give you the deep level of insights that you gain or you would necessarily you need to gain by spending time doing the work. And I feel like productivity these days is some sort of a God that we're bowing down to. yeah And I want to tell that God to fuck off. Sorry. But, um you know, I don't think that productivity is the be all end all of things.
01:26:38
Speaker
um I think oh for sure the there is a sort of necessity, I would say. I mean, I think there's a sort of joy in it as well. There's ah there's a slog in it sometimes, but there's a sort of necessity for us to actually do the work. you know No, for sure. And sometimes it's it's not even like the work, it's it's like you're saying like the difference between reading a book and reading the summary. like It takes longer, not just because there's more text, but because you're thinking through like the content and you're understanding it and your your brain is like making those connections and stuff like that. like I think there's there's that's the thing. like There's a certain type of of understanding that can't just be like absorbed
01:27:19
Speaker
Like from a a summary or something like that, like it requires like human brain cycles to kind of process and think through it. Like that's where ideas come from. And, yeah and that's, that's the thing that drives me crazy about the people who talk about Alan's like, oh, they're going to like save the world or come up with the next thing. It's like they they fundamentally can't like right at the end of the day, it's just, it's a, it's a random number generator. Like it's, it's just a bunch of, it's just a bunch of numbers that are spitting out the their best guests. It's best guess about what's the next most likely number to be in the sequence. And it's like.
01:27:49
Speaker
it's It's not that it's not cool and it's not that it's not powerful, but it's like it fundamentally incapable of like new discovery. That just doesn't make any sense. You know what I mean?
01:28:03
Speaker
Yeah, that's exactly right. And it's just the thing, I mean, you ask an LM, I don't know, to give you a summary of Moby Dick or something, or like a Thomas Hardy novel, you know, like Jude, you know, Jude the Obscure or Test the Durbeville and both of which, you know, you know, I shook Thomas Hardy's hand at the time and I thought it was a great job. You know, but especially when I test the Durbevilles, it's sort of like an emotional experience. um You can watch the film.
01:28:31
Speaker
But it's not the same as the book. right you know And now you can like you've got these like different levels of experiences. You can get a summary of chat GPT or some other LLMs are available. um but um you but yeah But reading the book and reading Hardy's expressions and having an emotional attachment to the real detail is is something different. and Obviously, science and facts are somewhat more, they're then they're more pliable in in that in that sense to summary. um you know But I feel still like you we're losing something in not going through the details. um know So i'm I'm all for like
01:29:20
Speaker
accelerating our learning and having some way to, to absorb new information. But on the other hand, I feel like Cheating is still cheating, you know, um, and, uh, I think it's like the teachers often said, you know, it's your own time you're wasting, you know, it's just like, it's your one time you're getting back. I don't know. I don't know where I'm going with this, but I'm not happy. Well, I think there's some thread to pull on there for sure. Like this kind of like call to productivity thing is a real problem. I really like ah Oliver Berkman's thinking on this. Like if it. He has a new, newish, anyway, book called ah Meditations for Mortals. Anyway, but it's like, like, that's the thing. The fundamental thing is like this, the stream of information to process or like the list of tasks to do is not just very long. It is infinite. yeah And as soon as you accept that, it totally changes.
01:30:09
Speaker
your perspective on life it's like getting going through like processing an infinite stream of tasks faster is not going to get you to the end any faster it's just you're just going to waste more of your life processing things and so it's like it's like what is right like when you're reading something and when you're listening to something like what is the goal right it's like the goal can't be to get through the list the to-do list or or like Read everything on your reading list like there is a functionally infinite amount of content to consume yeah and you are never going to get through it and so it's it's like coming to terms with your Like a finite nature as a human and accepting that like there's only so many hours in the day and I maybe that's part of it is like Yeah, I don't know
01:30:59
Speaker
Like, like, yeah, there, there are contexts where it's like in a certain thing, in a certain, in certain situation, I just want the gist of this. I just want the instruction notes and then I'm going to move on with my life. But yeah, like the the goal of the point of life, isn't to like get through your task list or like read all of your, all of your reading books as fast as possible. It's like the goal is to actually learn something or like make some meaningful connections. I'm already like, nevermind, like doing other things in the real world, but like, yeah, I don't know. Maybe something about that runs me the wrong way too. Right. It's like, like my.
01:31:29
Speaker
my goal when I'm reading something new isn't to get through it as fast as possible. It's to like really absorb it and process it and like internalize it. and That's just at odds with like this whole culture of like do everything as fast as possible and do everything as efficiently as possible. and yeah That's not always the point. Exactly. I think there's a lot to be said for um a deceleration, to be honest, like slowing down and and not in the um kind of lifestyle coach sort of way. And I'm not trying to be, sorry, I'm not trying to be flippant here. I'm i'm really saying like, yeah you know, like, ah I don't know, there there is, and kind of going back to what we were talking about at the beginning, I don't remember whether it's on the recording or not, or about um
01:32:17
Speaker
you know, that working in an environment where you kind of decide when it's ready, rather than being told you have to finish it by next Tuesday. You know, like there is definitely value in that. um And, you know, like taking the time to actually develop these thoughts, I think, like we're talking about. Will you get your solution out there, the fastest doing that? um No.
01:32:44
Speaker
But, you know, Closure took whatever 10 years to write. Datomic took 10 years to write, whatever, you know? yeah yeah So like good things actually do take time. And I think there's a lot to be said for taking your time. and and yeah yeah doing stuff Yeah. I think the other thing is the the the flip side of that is that um mean maybe it's somewhat too obvious to say, but the the cult of productivity is a completely dehumanizing thing.
01:33:16
Speaker
yes it's so it's It's completely saying that all you are is a cognitive machine, just a part, you're just a completely replaceable part that can be swapped in and swapped out.
01:33:28
Speaker
And but that's unfortunately where we're kind of going with these LMs, you know where people are comparing the output of LMs to humans. And it's like, yeah we don't we won't need lawyers, we won't need doctors, we won't need teachers, we won't need this, we won't need that, we won't need programmers.
01:33:45
Speaker
you know um And I feel like I haven't got a problem with if this thing genuinely was a kind of one-to-one replacement for us, then I get it. Fair enough. But this concept that this, like you say, this essentially random number generator that happens to be very good at generating a decent guess can like replace a human being is offensive. yeah um yeah tremendous And yet it's very culturally
Technology's Role in Society
01:34:20
Speaker
acceptable. know yeah And that's that's a kind of like a worrying narrative to me. I know we're kind of not talking about closure anymore, but but no that we're not um we're allowed to speak about other things on this podcast. Now that Josh is here, we're being given permission.
01:34:35
Speaker
brought in the scope. Yeah. It's so good. Emily Bender, Emily Bender and Timmy Gibrew and a few people in that kind of circle have some really interesting thoughts along those lines. But yeah, I mean, I totally agree with you. Like this whole, I mean, it's not, it's not, I don't think it's a coincidence that that, that narrative and that mentality kind of prevails from like Silicon Valley and yeah that very like techno utopian or dystopian, depending how you see it, like Dis. Ridiculous. Yeah. like That's the thing. like there There are a large number of people who like basically like have their entire lives in Silicon Valley and like more or less don't interact with the rest of the world. and it's like It's a very weird place. and I think it's not a coincidence that like some of the ideas that come out of there are just like, to like a normal person who doesn't live there like seem completely absurd, but yet they talk about it like it's totally normal. And it's like, do you like do you hear yourself? like that's That's ridiculous. Am I crazy? like every This is crazy, right guys? And it's like, then you talk to other people and they're like, yeah, that's crazy. But then you talk to like a tech person and they're like, no, man, it's the future. And you're like, it's very it's very disorienting for sure. and Talking is disorienting. yeah You just rendered us speechless, which has never happened before.
01:35:47
Speaker
ah Yeah, no, I think honestly we we could, we could ah this is really good and I feel like we should um we should have more of these conversations, you know, maybe it's in a bit more depth on some of these podcasts because, but we're already in our 46th, which is probably, I think most people have dropped off by now, i would be married which is it which is a damn shame. yeah yeah yeah but um i honest i see I feel like with that there is a sort of need let's Let's just do this for five minutes at least, just to sort of like, let's get the Luddism out there, you know, because I'm very much, you know, um, a kind of like, uh, with the Luddites in the sense that technology is fantastic, but it's for human good, you know? Um, and we should, we should definitely not worry about technology or we shouldn't stop developing technology, but we should have some.
01:36:45
Speaker
concept of, like like you were saying about Silicon Valley, that it's not just about people being enriched. It's about enriching the its ah but enriching humanity, um we you know the kind of laudable ideals that we might have around technology.
01:37:01
Speaker
seem like that's what we that's probably a lot of people could, the the vast majority of people could agree on. know if they want to bring and you know And of course, you can hide a lot of this um this nefarious stuff underneath the sort of shield of medicine or saving the planet or whatever.
01:37:22
Speaker
But in fact, it's just really about encouragement. And I feel like that that is really the problem that we're that we're experiencing. and We're being gaslit yeah and we all know it, but we can't do anything about it. And this is the sort of frustration of this stuff that's going on right now, in my opinion.
01:37:40
Speaker
I totally agree though. and I mean, yeah, I'm not just saying this cause it's, cause it's you guys, but like, that it's just, that's the fundamental, capitalist dystopia we live in, right? It's like everything is for profit. And if it's not profitable, it's not worth doing. And that it's just like that really wants.
01:37:58
Speaker
Like everything like it totally like ruins the essence of being human, right? Like the point of being alive isn't to make these rich guys richer. Like that is clearly not the point, but like that's what all of our entire society is oriented around. It's like, if you're, if you're a minute to minute activity is not enriching somebody who is very likely already beyond wealthy, like wealthier than anybody ever needs to be.
01:38:23
Speaker
It's like deemed unworthy. It's like, it's like your, your work is not valued in society and it's like, yeah, or you're or you're not productive or you're wasting your time. Like if you're, if you're just reading for the sake of reading or we're learning an instrument or playing with a child or going for a walk with your dog, it's like you're wasting time and.
01:38:42
Speaker
And yeah, it's, it is, I think it's and deeply frustrating and it's, everybody hates it. And yet we all are powerless to change it because we're all at the mercy of the system. Because if you stop working, you will die. Like we live in a society where the stakes are so high. Like if you don't, if you don't earn money, we are like, we, have we have decided as a society that if you don't earn money, you don't deserve to live. Like effectively, like and at least in Canada and the U S like,
01:39:10
Speaker
If you're poor, you are just completely fucked. Like you are, you will be homeless. Like in no time, you will have health issues. You won't have access to food or shelter. And it's like to me, yeah, it's just, it's completely absurd and and ridiculous that we live in like the wealthiest.
01:39:25
Speaker
like society and the history of the planet. And yet we have like people who don't have enough food, but anyway, whatever. That's like a big soapbox. Like, yeah. thing yes Yeah. It's this low violence of neoliberalism, but it is like, we're, it When you lay it out like that, we are all held a hostage, right? Or die. It's as simple as that, work or die. And I get to decide what you work on, by the way. Yeah, yeah yeah exactly. You don't get to just work on things you find interesting. I know you live in Sweden, for God's sake. like That's supposed to be the the way out. you know it's like that Even in Sweden. There is no way out.
01:40:05
Speaker
It's true cuz like I'll often tell people like I mean I'm I'm definitely and it is a little bit rich
Living Beyond Capitalism
01:40:10
Speaker
I get that it's a little bit rich for three like software engineers who are like very wealthy in the scheme of things to be complaining about all this but it's like You know, I'll talk to people and be like oh I really miss working on this or that thing and people are like why don't you do it? And it's like I need a job like I need to work full-time to survive like I literally cannot survive unless I work full-time And that requires that I do the things that my employer demands. like That's the deal we have. I do the things they want. They give me money. like it would be It would be disingenuous for me to to do otherwise. And it's like you know I have a little bit of time on my evenings and weekends and whatever to do the things I enjoy. but And and you know I'm relatively lucky that to be able to often find work that I enjoy. But and ah yeah, it's a crappy situation because like a lot of people are in very precarious
01:40:55
Speaker
situations through no fault of their own. They're working extremely hard and doing their best and still like living on the margins of society. and it's like But you're right, like it's it's it's not like one of the things that you know, people say like, we need to fix the system or whatever, but it's like not really, like it's actually working exactly as it's intended to. It's not broken. It's not broken at all. This is exactly the the design. It's like, if you don't bend and you don't you don't oblige, you will be marginalized and you'll pay the price. Like, you know, ultimately with your life and anyway, whatever, this is getting very grim, but you know, it's, it is it's interesting to think about and talk about because it's to me, it's like, what, what can you do about that? Like,
01:41:38
Speaker
like One of the things I've done is like I live in like the middle i live a very small town. We bought two and a half acres of like empty land, and the goal is to eventually like try to slowly opt out. and it's like I don't think that's super realistic. like you can't You can never like fully dissociate from like the system or whatever, but it's like I think you can like by focusing on like your community and your local environment. like could start like growing some of your own food and then you don't have to buy it from the grocery monopolies. You can like like you know raise your kids in a small town where they know their neighbors and they're safe to walk on the streets and and like go to their classroom instead of like like whatever, walking through this like chaos. and I don't know. it's it is very It's very disempowering though because it's like, what can you actually do about it? I don't know. Wow. I think the way to do things about it
01:42:26
Speaker
Well, i I think first of all, I don't think it's a problem for the three software engineers who are relatively wealthy to complain about it. I think we should complain about it because we have a platform, we've got something, you know, where we're fortunate and i think I think it's good to hear voices like this that are ah kind contending with the the the the sort of generic narrative, um I think it's important to get these things out here. you know Even if we're you know in a tiny corner, I think it's okay. um But I think the way to do things about it, I think the way to do things about it is to organize in the end.
01:43:06
Speaker
Um, not to put too kind of like much of a, uh, uh, what should we say, like a reference to other things that myself and Josh have done, but we have this other podcast called organizing tech in Sweden, which is all about, uh, tech workers have organized and become union members. And it is becoming a thing, you know, that, uh, that is happening. Obviously.
01:43:31
Speaker
We're in a weaker position um in many firms than we were like 40, 50, 60 years ago prior to the kind of neoliberalism of the modern age. But but it's still something we can do. you know And it's something that we should see that's on the table.
01:43:49
Speaker
You know, um, we shouldn't take it off the table ourselves. You know, we should put it on the table and let other people kick it off. They probably will. They probably will. You know, if you, if you're in, uh, if you're in a cobalt mine in, um,
01:44:04
Speaker
in the Democratic Republic of Congo, heavy heavily like quotes on Democratic Republic of there, um then you'll get shot. you know that That's how high the stakes are there. but um But over here in Techland, we can at least talk about it, so we should do that. you know yeah We have got options. We have got things we can do.
01:44:28
Speaker
We can organize with our communities, we can organize with um the people that we're allied to. So but let's do that. Let's end on a more hopeful note where you know we did go grim, but I think we can recover by getting involved and getting organized. Indeed.
01:44:48
Speaker
And on today's podcast about closure, we talked about existential despair. Uh, yeah, no, there's a lot to be, a quick reset to essential. I didn't really do much magic. Sorry about that. as one Don't worry. Yes, exactly. What was the point?
01:45:11
Speaker
We'll receive it on that then first. Yeah, amazing. Thanks, Kira, for ah for showing up and talking about exciting stuff for a long time and then five minutes of just, like, real cut-wrenching. Yeah, yeah.
01:45:27
Speaker
be real the conversation. It's like the unintended consequence of mentioning like poverty or something. It's like, Oh, well, oops. That's why we don't talk about that. just I think we brought up LLMs. You know, I think that was it. That's what we were talking about. That's how it led here. Yeah. Anyway, LLMs. Very cool. it was Very cool. Amazing. ah um They brought value to their shareholders, you know, so that's great, isn't it?
01:45:57
Speaker
Yeah, yeah. On the happy note, we'll stop the recording here. Thank you for listening to this episode of DeafN and the awesome vegetarian music on the track is Melon Hamburger by Pizzeri and the show's audio is mixed by Wouter Dullert. I'm pretty sure I butchered his name. um Maybe you should insert your own name here, Dullert.
01:46:23
Speaker
voter if you'd like to support us please do check out our patreon page and you can show your appreciation to all the hard work or the lack of hard work that we're doing and um you can also catch up with either Ray with me for some unexplainable reason ah you won't interact with us then do check us out on slack closure in slack or closure verse or on Zulep or just at us at deafened podcast on Twitter Enjoy your day and see you in the next episode.
01:47:30
Speaker
Yeah, that's exactly the reaction I was looking for, yeah. but
01:47:40
Speaker
yeah for for for audio For audio listeners, she was yawning they all the way through that part.