- Emergent Behavior
- Posts
- EB-4: Dingboard - Transcript - Part 2
EB-4: Dingboard - Transcript - Part 2
Ate A Pi
April 14, 2024
Transcript: Dingboard - Part 2
🧑🦰 yacine: It's all like how corporations use Canada as a test bed because it's very like America, but the regional controls make it very easy to just flip a flag and send the test out to Canada.
👩🎤 Ate-A-Pi: Yeah. And, you know, a lot of the Chinese firms, what they do is they have multiple teams working on features at the same time. There's no like, okay, we're just going to have one team working on, you know, segmentation or whatever. They're like, all right, we're going to have like three teams. And then you each get like one region. And whoever gets the best, best feature out that feature is going to get, you know, you're going to get promoted, you're going to get bonuses. The rest of you guys, you know, you're going to get absorbed back in the labor pool.
Right. So, uh, and, and that's, you know, 996 culture doesn't come from like having like FaceTime having FaceTime of 996. It comes from like, you know, getting fired. If your feature doesn't, you know, doesn't go live. Right.
Right. So, uh, and, and that's, you know, 996 culture doesn't come from like having like FaceTime having FaceTime of 996. It comes from like, you know, getting fired. If your feature doesn't, you know, doesn't go live. Right.
🧑🦰 yacine: It's kind of fascinating how much competition is a motivating factor or like a driving motivator. I think like a lot of really great accomplishments in American corporations is driven by competition between different people, like organizations. I think that you can like stoke it healthily and keep it like in a, you can do it to like a healthy degree. The TikTok approach is, I think,
they realize that the cost code is actually very cheap to produce and very replaceable. The thing that isn't cheap to find is really sick talent that can just do that and see things and execute super quickly. So that's actually really fast. I did not know that. So that's super interesting.
they realize that the cost code is actually very cheap to produce and very replaceable. The thing that isn't cheap to find is really sick talent that can just do that and see things and execute super quickly. So that's actually really fast. I did not know that. So that's super interesting.
👩🎤 Ate-A-Pi: Yeah, I mean, I think one of the reasons why it's very hard for them to imagine getting sold is that I suspect in order to run the business outside of China, they would need to export maybe like 10,000 engineers from China. And I think it's just not possible. So
🧑🦰 yacine: Yeah. I think the people don't really understand how much of a technical achievement social media applications are. The throughput that they spit out is like, it's like an order of magnitude that's difficult for a human to imagine.
👩🎤 Ate-A-Pi: Yeah, every time every time Zuck gets blamed for like, some something escaping the moderation, I'm like, do you do you understand how many hits, you know, how many times how effective their filters were, they were, their filters were 99.999% effective. And then you had like 100, like incidences of things escape, right. But because there's no there's no 100% filter, right. So
🧑🦰 yacine: Yeah, people.
even humans, I think like people.
get a bit confused when it comes to scale. They kind of imagine Facebook like this mom and pop shop, but I think a better way to imagine it is that they have like a rocket engine inside a basement and the rocket engine is at full throttle all the time. And then there's probably a thousand of those. And the fact that like, it's basically a guaranteed inevitability that there's going to be a harmful use of the platform. And I think like, it's actually pretty incredible what Facebook has done. So Facebook, without Facebook, DingBoard would not exist because they open sourced, right? Like a lot of their software.
which then turns into more research and other folks producing better and better models, which I rely on. But not only that, they also produce a lot of really great models that understand images, which actually helps other corporations, hopefully at some point governments to use AI models to help get more work done faster. So I think Facebook
the amount of positive impact that they have by creating their models and giving them out for free.
even humans, I think like people.
get a bit confused when it comes to scale. They kind of imagine Facebook like this mom and pop shop, but I think a better way to imagine it is that they have like a rocket engine inside a basement and the rocket engine is at full throttle all the time. And then there's probably a thousand of those. And the fact that like, it's basically a guaranteed inevitability that there's going to be a harmful use of the platform. And I think like, it's actually pretty incredible what Facebook has done. So Facebook, without Facebook, DingBoard would not exist because they open sourced, right? Like a lot of their software.
which then turns into more research and other folks producing better and better models, which I rely on. But not only that, they also produce a lot of really great models that understand images, which actually helps other corporations, hopefully at some point governments to use AI models to help get more work done faster. So I think Facebook
the amount of positive impact that they have by creating their models and giving them out for free.
👩🎤 Ate-A-Pi: Yeah, I think when I look at Zuck's strategy, it is basically he is so in that they're going to capture, they're going to improve the way that humans communicate with each other, right? And as long as they, because they own that, like the web, the network in between two human beings, they own, you know, three or four different ways now.
🧑🦰 yacine: Yes. They actually own the infrastructure as well, which is very important to keep in mind. They literally own the data centers and the edge networks and stuff. They're running their own, they're their own server provider.
👩🎤 Ate-A-Pi: Yeah, it's always struck me as interesting why he didn't try to pull an AWS and try to make some more money off of like outsourcing, allowing other people to use their servers.
🧑🦰 yacine: Dude, if I had to guess, if I had to guess they probably don't have enough servers for themselves. How could they sell? Right? Like how could they sell their servers if they don't have enough for themselves?
👩🎤 Ate-A-Pi: They had a service called Parse, which they bought at some point. That was providing a back end, kind of like Firebase, kind of like a memory back end. And the moment they bought it, Parse lost all the users because people were like, oh, you know, Zach's going to find out which social app that I have is hot on Parse. And he's going to like, they're going to.
🧑🦰 yacine: Fascinating. I generally trust that corporations don't do that. Like, there's so many controls. Like, if you've ever worked with a large corporation, the amount of due diligence that folks do and are encouraged to do, like, the sort of the Western systems of corporate governance in general are actually quite good. And I really wouldn't worry about that. Like, that would be crazy. I would be excited about the fact that my data could be spied on and I could get to sue Facebook for a lot of money.
👩🎤 Ate-A-Pi: Yeah, that, you know, there's this interesting thing where during the whole privacy scares of the 2010s, you know, people are pushing on, oh, you know, privacy is important, blah, blah. And it didn't make sense for a lot of voters. And I think it was either John Oliver or someone else. And he goes up to people in New York and he asked them like, oh, what do you think about Facebook and privacy?
Et cetera. And they're like, oh, you know, it doesn't matter. Et cetera. And then he says like, they can see your dick pics and then like, Oh, Oh my gosh, privacy is important.
Et cetera. And they're like, oh, you know, it doesn't matter. Et cetera. And then he says like, they can see your dick pics and then like, Oh, Oh my gosh, privacy is important.
🧑🦰 yacine: Yeah, exactly. Yes. I think generally education is pretty important for these folks who, if it's not your computer, then it's not your data, right? It's a baseline assumption to have, especially if your threat model is someone sees a picture of my penis. And generally just don't upload pictures of your penis into cloud services. That is pretty, but I think it's kind of easy for people like us who are pretty good with computers.
to expect or have a higher expectation of gen pop and think that you should. I actually think it's kind of like, you need to be careful to not do that. The average human, there's a responsibility for corporations to educate or make it very easy for people to understand where their data actually is.
to expect or have a higher expectation of gen pop and think that you should. I actually think it's kind of like, you need to be careful to not do that. The average human, there's a responsibility for corporations to educate or make it very easy for people to understand where their data actually is.
👩🎤 Ate-A-Pi: I see that, but I don't believe in that it's the user's responsibility to take care of their data that closely. I think users can have an expectation that things will happen as they should, and no one's going to do bad things, regardless of where that data is sitting. I think that's a fair expectation for users to have. And I think society should be able to do that.
🧑🦰 yacine: Yep. Yes.
👩🎤 Ate-A-Pi: you know, can and should reshape itself to enable those expectations to be real, right? Of course, you also have free speech and everything else, and you want to enable the other side of that too. But I think user safety, I think, is something that users have a right to, and companies need to figure out how to enable those, right?
🧑🦰 yacine: Yes, it's incredibly important. It's just like a general, excuse me, some general ethics thing, generally good ethics, where you need to set really good expectations of your users and handle their data appropriately and do all the necessary work. The threat models that I deal with are a bit different from, let's say, average gen pop, where for me, there's certain things where we're like, OK, that should be on a piece of paper.
buried under, you know, like there's things that are like that level of threat models. So like, for example, like don't like zip your keys and like upload on Dropbox because Dropbox might get pooned. So yeah.
buried under, you know, like there's things that are like that level of threat models. So like, for example, like don't like zip your keys and like upload on Dropbox because Dropbox might get pooned. So yeah.
👩🎤 Ate-A-Pi: Let's talk a little bit about competition. So right now, I would say Dingboard is known, right? Like on Twitter, it's known. You must have a bunch of VCs asking, knocking on your door. And you should have a bunch of people like, hey, offering checks, et cetera. You should also have like people trying to copy your features, et cetera. Like, so how does that work? Do you see?
Do you see people like, you know, fundraising based on similar ideas? Um, and, and how would you, you know, I, there's this tweet from you, me watching a VC backed image editing startup with 300 employees get a $500 million valuation. And then there's a, there's a, there's a McDonald's anime, uh, character, uh, with a nah, I'd win. Uh, you know, uh,
Do you see people like, you know, fundraising based on similar ideas? Um, and, and how would you, you know, I, there's this tweet from you, me watching a VC backed image editing startup with 300 employees get a $500 million valuation. And then there's a, there's a, there's a McDonald's anime, uh, character, uh, with a nah, I'd win. Uh, you know, uh,
🧑🦰 yacine: Yeah, exactly. That's the kind of, I think the most important thing for the baseline for startups is that you need to be able to survive. And in so far that is satisfied, I can go nuclear. So my motivation behind posting that picture of me in McDonald's is that I am saying that I can go get a job at McDonald's to keep Diggin Board running.
at the limit, I can do that. So effectively I'm playing a different game than my competitors are in the first place, which fundamentally means that like, they fundamentally cannot compete with me because they're in the business of building venture-backed startups and greasing the wheels and then eventually dumping the stock on the public.
at the limit, I can do that. So effectively I'm playing a different game than my competitors are in the first place, which fundamentally means that like, they fundamentally cannot compete with me because they're in the business of building venture-backed startups and greasing the wheels and then eventually dumping the stock on the public.
👩🎤 Ate-A-Pi: Right.
🧑🦰 yacine: I'm not in that business. So they literally cannot compete with me. What can happen, though, is that I take their market share on their users, which is a lot easier than I thought it would be, mostly because I think that when your company grows too large, you start optimizing for the you will have trouble optimizing for the right things. So the right thing here being just user satisfaction and steady growth and survivability.
I think you'll start optimizing for.
I went on a whole rant about this today on a Twitter space, but it's kind of frustrating to me to not see a venture-backed ding board because it's so important to exist. But it kind of makes sense too because it's really, I really don't get why this doesn't exist. I actually don't understand why this has never existed before.
And I don't get why I don't feel like a bigger threat from the VC-backed startups or the large incumbents. I actually don't understand why they aren't threatening me, because they aren't. And they should. In a well-functioning market, I think they should. I think there's a very big market demand. I've definitely proven it to myself. But also like,
I think you'll start optimizing for.
I went on a whole rant about this today on a Twitter space, but it's kind of frustrating to me to not see a venture-backed ding board because it's so important to exist. But it kind of makes sense too because it's really, I really don't get why this doesn't exist. I actually don't understand why this has never existed before.
And I don't get why I don't feel like a bigger threat from the VC-backed startups or the large incumbents. I actually don't understand why they aren't threatening me, because they aren't. And they should. In a well-functioning market, I think they should. I think there's a very big market demand. I've definitely proven it to myself. But also like,
🧑🦰 yacine: I have things that they don't have. Like I've got a marketing channel and I've got the David versus Goliath meme. And I think it encourages people to like become true. Like it becomes easier for people to become true fans and they're like actually talking to the person who's actually made it. And they're not like some like, you know, database record of some user that's, you know, getting aggregated up into some sheet that people are using to like raise, you know, be series B and three C.
🧑🦰 yacine: It's really hard. I mean, here's a, I think this is an salient point to make. I think that I could probably build Dingboard in two weeks if I just started from scratch again. Just straight up, just empty VS Code. I'd probably build it a bit differently. I'd probably use a game engine that runs natively on more platforms. But like.
I could write it in two weeks from scratch. So why hasn't anyone done it yet? Right, like that's like, I guess that's like the million dollar question. It's like, why hasn't anyone actually done it? I think I also have the propensity to like overestimate the like median skill of software developers. Basically anyone with, you know, the skills to build DingBoard is probably, you know, a consultant and making a lot more money and in a lot safer manner.
And the reason I'm mostly doing this, the reason I'm really doing this is mostly for amusement. I'm actually not trying to make a lot of money. I should have put up, for example, there's a paywall on digboard.com right now, so you can access the AI features, which are really great, by the way. And you should totally do it. But now there's a paywall. After being up for six months, there has been no paywall, which in retrospect, I should have done a long time ago, because it does definitely help you scale. But I'm not, I don't know, my motivations aren't.
Like I'm not even competing with them, right? Like I've been rambling for a while, but it's, I've been very confused about this myself. Like, why hasn't anyone actually become a threat? Um, I think a lot of it too is like, I just know what to optimize for. Like I know how to, like, I know what actually matters. Um, and it's not something that VC companies can have a moat around. So they're inherently not motivated to do that thing. It's something that was like very functionally easy is like build an app that doesn't suck ass.
make it easily accessible for people. If they can't afford it, just flag them in for free. But...
I could write it in two weeks from scratch. So why hasn't anyone done it yet? Right, like that's like, I guess that's like the million dollar question. It's like, why hasn't anyone actually done it? I think I also have the propensity to like overestimate the like median skill of software developers. Basically anyone with, you know, the skills to build DingBoard is probably, you know, a consultant and making a lot more money and in a lot safer manner.
And the reason I'm mostly doing this, the reason I'm really doing this is mostly for amusement. I'm actually not trying to make a lot of money. I should have put up, for example, there's a paywall on digboard.com right now, so you can access the AI features, which are really great, by the way. And you should totally do it. But now there's a paywall. After being up for six months, there has been no paywall, which in retrospect, I should have done a long time ago, because it does definitely help you scale. But I'm not, I don't know, my motivations aren't.
Like I'm not even competing with them, right? Like I've been rambling for a while, but it's, I've been very confused about this myself. Like, why hasn't anyone actually become a threat? Um, I think a lot of it too is like, I just know what to optimize for. Like I know how to, like, I know what actually matters. Um, and it's not something that VC companies can have a moat around. So they're inherently not motivated to do that thing. It's something that was like very functionally easy is like build an app that doesn't suck ass.
make it easily accessible for people. If they can't afford it, just flag them in for free. But...
🧑🦰 yacine: The problem with a company like that as a VC backed company is that someone like me can just come in and like scoop up your meal. So they won't find it in the first place. Maybe it's just that, you know, image editors are not good for VC fund. Like, you know, I've talked, I've spoken to a lot of VCs. So you said recently, VCs are reaching out to you and stuff. And what I'll often say to them is, hey, like, thank you for reaching out. I appreciate the vote confidence, but I do not think that this is a good venture. You know, this is a, I don't think this is a good business to like become, like to become an investor in.
moonshot because this thing has a hard cap of how much money it's going to make. It's not going to get an Adobe. Someone else will come up and clean our lunch. Like a really good example is the Photopea guy who's basically just rebuilt Photoshop from scratch in JavaScript and WASM. Why would you invest in a Photoshop when someone can just do that? Maybe it's actually just like VCs are just not incentivized to fund these things. And
are reallocating their resources into somewhere where capital actually does have a big advantage, like hard tech. Venture capitalists, you actually need to lock capital to get that started up. And so they have actual alpha on that. They have opportunity that other people don't have that are locked out of so that they end up getting a bigger piece of the pie. I think that's probably what it is. I think it's just some functional thing.
moonshot because this thing has a hard cap of how much money it's going to make. It's not going to get an Adobe. Someone else will come up and clean our lunch. Like a really good example is the Photopea guy who's basically just rebuilt Photoshop from scratch in JavaScript and WASM. Why would you invest in a Photoshop when someone can just do that? Maybe it's actually just like VCs are just not incentivized to fund these things. And
are reallocating their resources into somewhere where capital actually does have a big advantage, like hard tech. Venture capitalists, you actually need to lock capital to get that started up. And so they have actual alpha on that. They have opportunity that other people don't have that are locked out of so that they end up getting a bigger piece of the pie. I think that's probably what it is. I think it's just some functional thing.
👩🎤 Ate-A-Pi: So I would maybe, so what often happens for a VC kind of play is that you need a narrow niche to show that you have initial user product market fit. And then over time, you need to keep expanding the market. So if you look at, let's say, Elon.
Elon does this, started off Tesla with, hey, we're going to make this car. And then he went on to, okay, it's also going to be self-driving. Then he went on to, okay, we're going to do robots because that's this enormous market. So he keeps expanding the market each time because he's not going to be able to hit the next valuation hurdle if he doesn't expand the market. The only way that he can... Because...
Elon does this, started off Tesla with, hey, we're going to make this car. And then he went on to, okay, it's also going to be self-driving. Then he went on to, okay, we're going to do robots because that's this enormous market. So he keeps expanding the market each time because he's not going to be able to hit the next valuation hurdle if he doesn't expand the market. The only way that he can... Because...
🧑🦰 yacine: Right.
👩🎤 Ate-A-Pi: And a lot of the thing about him in every era is that right now people are like, okay, great, you're an EV car company. Right? How big is the EV car company going to be? So he has to push forward this idea that no, we're going to replace all labor on earth. Right? He has to push this idea of bigger markets. And I think when I look at Dingboard, it reminds me of the early days of Figma. Right? When... because...
🧑🦰 yacine: Yes.
🧑🦰 yacine: Right.
👩🎤 Ate-A-Pi: Figma was basically collaborative design. And the reason why Adobe had to buy Figma in the end was because Adobe was in this single site kind of design, and they hadn't figured out how to transition to kind of cloud design because it's completely different ways of interacting with the design stage, right?
🧑🦰 yacine: Right.
👩🎤 Ate-A-Pi: And what was remarkable for me in looking at Dingboard was that the fact that you'd unpacked the inference layer into basically a partial on the client side and a partial on the server side. And you were having this interaction. And it struck me that this is impossible for, this is basically something that starts to look a little bit different.
And it's going to be hard for, let's say, even, let's say, an Adobe or Figma, which has to provide a consistent user experience of multiple platforms. You have to have a consistent user experience of the iOS and the Windows machine. And it has to all work on the browser properly. And you can't have this thing where, perhaps, with a different browser, it doesn't work. They're not going to be able to do that off the get-go. And so,
And it's going to be hard for, let's say, even, let's say, an Adobe or Figma, which has to provide a consistent user experience of multiple platforms. You have to have a consistent user experience of the iOS and the Windows machine. And it has to all work on the browser properly. And you can't have this thing where, perhaps, with a different browser, it doesn't work. They're not going to be able to do that off the get-go. And so,
🧑🦰 yacine: Uh...
👩🎤 Ate-A-Pi: That to me starts to like, okay, now you have this kind of thing where they can't really do it without reformatting a bunch of stuff, which in a big company takes a really, really long time. And you can kind of start off with the core features and expand outwards.
🧑🦰 yacine: Ah, oh.
🧑🦰 yacine: I would challenge that assertion. So the ONIX, I've tested on phones that are 80% pre-optimal. So the inference actually works on all phones. The UX is just not designed very well. But you said earlier that founders who show initial traction, they need that initial traction to show up to venture capitalists and raise funds.
But I would actually challenge that assertion because, for me personally, it doesn't make sense. If you hit success and then you have user traction and then you have cash flow, it actually doesn't make sense to me why you would raise when you already own the whole bag and the bag is basically just gonna keep on growing. It's kind of hard for me to rationalize why I would ever...
For me, it's actually the thing that actually cost me a lot more than, you know, I don't give a shit about my company and like, like the value of my company. I actually care about how much work it is for me to like do all the paperwork and like understand all the legalese to make sure that I'm not getting screwed. So for me, I'm just like, actually, no, I just won't do that. And it's way too, it's way too painful and it's annoying and I don't want to do that. I don't want to do it. And yeah, like it's, it's kind of, it's kind of fascinating to see like this, like.
need for growth as well, where you have to return that investment, that investment process where the company itself has these growth opportunities and are constantly reaching for those growth opportunities. But I would challenge the need for that and I would have people consider a lifestyle business that is self-sustaining and they build what their users want and just do it very naturally
But I would actually challenge that assertion because, for me personally, it doesn't make sense. If you hit success and then you have user traction and then you have cash flow, it actually doesn't make sense to me why you would raise when you already own the whole bag and the bag is basically just gonna keep on growing. It's kind of hard for me to rationalize why I would ever...
For me, it's actually the thing that actually cost me a lot more than, you know, I don't give a shit about my company and like, like the value of my company. I actually care about how much work it is for me to like do all the paperwork and like understand all the legalese to make sure that I'm not getting screwed. So for me, I'm just like, actually, no, I just won't do that. And it's way too, it's way too painful and it's annoying and I don't want to do that. I don't want to do it. And yeah, like it's, it's kind of, it's kind of fascinating to see like this, like.
need for growth as well, where you have to return that investment, that investment process where the company itself has these growth opportunities and are constantly reaching for those growth opportunities. But I would challenge the need for that and I would have people consider a lifestyle business that is self-sustaining and they build what their users want and just do it very naturally
🧑🦰 yacine: Because at the end of the day, once you have a cash flow business, a business that's creating some cash flow, it frees up, in my mind, the only capital that really matters, and it's time. And you can start reinvesting that time into other projects and hitting those growth opportunities as well. The reason I think that Figma and I think Figma probably crushed me, they tried, to be honest. The reason I don't think Adobe will do something like this is because.
They already have a cash cow. And if they build a competitor that's as cheap as dingboard, people will just stop paying for Photoshop. So they'll just end up competing with themselves. Yeah, so hopefully Figma doesn't crush me. Ha ha ha.
They already have a cash cow. And if they build a competitor that's as cheap as dingboard, people will just stop paying for Photoshop. So they'll just end up competing with themselves. Yeah, so hopefully Figma doesn't crush me. Ha ha ha.
👩🎤 Ate-A-Pi: On another tack.
👩🎤 Ate-A-Pi: What is your favorite Japanese anime and why?
🧑🦰 yacine: Um, so my favorite anime is space dandy. One of my favorite animators space dandy. Um, I have a lot of favorite anime like Steins gate and stuff, but space dandy. I like space dandy a lot because it's very relatable. I can relate to the main character.
And it's very, very funny. It reminds me of old adult swim cartoons. I really, really like that. And if you're looking for an anime to watch and you're just trying to hang out some episodic anime, it's really good. But man, there's so much great anime. What's really cool about all this diffusion stuff and tools, like even DingBoard, I plan to make it so that you can very easily create images. And using, so basically the problem with these image generators is that
you're outsourcing all the creativity. But in reality, what I think should happen is that we should build the tools so that you can very specifically control these diffusion tools so that the creativity is still coming from you and from the models, all we get is automation of labor, like the uninteresting non-fun part. I think like...
As we build tools, and I really do plan for, to make Dingboard better at creating anime images that you specifically control and design yourself, being able to draw things on the canvas and then have it automatically spawn something. I think that we're going to have more and more anime actually get created. You already see this happening with Blender and really cool people, people are coming up with cool Blender workflows and comfy UI and stuff. But unfortunately, not everyone can wrangle ARC install and CUDA, so hopefully I can ship something that can help bridge the gap for a lot of people.
Uh, very exciting, very, very exciting. There's so much, there's so much manga that needs to be animated. Oh my God.
And it's very, very funny. It reminds me of old adult swim cartoons. I really, really like that. And if you're looking for an anime to watch and you're just trying to hang out some episodic anime, it's really good. But man, there's so much great anime. What's really cool about all this diffusion stuff and tools, like even DingBoard, I plan to make it so that you can very easily create images. And using, so basically the problem with these image generators is that
you're outsourcing all the creativity. But in reality, what I think should happen is that we should build the tools so that you can very specifically control these diffusion tools so that the creativity is still coming from you and from the models, all we get is automation of labor, like the uninteresting non-fun part. I think like...
As we build tools, and I really do plan for, to make Dingboard better at creating anime images that you specifically control and design yourself, being able to draw things on the canvas and then have it automatically spawn something. I think that we're going to have more and more anime actually get created. You already see this happening with Blender and really cool people, people are coming up with cool Blender workflows and comfy UI and stuff. But unfortunately, not everyone can wrangle ARC install and CUDA, so hopefully I can ship something that can help bridge the gap for a lot of people.
Uh, very exciting, very, very exciting. There's so much, there's so much manga that needs to be animated. Oh my God.
👩🎤 Ate-A-Pi: Any, uh...
Any Japanese ding board users?
Any Japanese ding board users?
🧑🦰 yacine: I actually don't know, so I don't really keep it. I actually need to set that up because of I'm hitting that point where I need to start reporting on GST. So I need to know who's Canadian and who's not.
👩🎤 Ate-A-Pi: I always wonder, because in Japan, you know, anime artists are actually very lowly paid. It's kind of like sub $1,000 a month kind of, you know, and a lot of piecework and absolutely abused, like absolutely abused by the studios. In fact, when Netflix started its big anime push,
🧑🦰 yacine: Yes.
Yes.
Yes.
👩🎤 Ate-A-Pi: It was even worse because the studios had a lot of contracts, but they wouldn't pay anything above the going piece rate. And people enter into it with all of this passion. And then they find themselves in their mid thirties and they're living with their parents and working 18 hour days for like 500 bucks a month. It is really, really tragic. And I wonder, what is the bottleneck? Is the bottleneck...
🧑🦰 yacine: Yes.
👩🎤 Ate-A-Pi: I mean, if they had better creation tools, would they be able to create better? Or is it like, is it the entire packaging and distribution? Is that, is that really going to be the problem? Like even if you had better tools and the, because, and the sad part is a lot of the artists, they're the, the part of the AI that we're putting out right now are the pieces that will automate the, uh, the, the hand-drawn, uh, artistry.
while the storylines are still created by the main creator or the studios or whatever. So I wonder how that transitions, whether we actually get these independents kind of being able to distribute, to make and distribute on their own using AI.
while the storylines are still created by the main creator or the studios or whatever. So I wonder how that transitions, whether we actually get these independents kind of being able to distribute, to make and distribute on their own using AI.
🧑🦰 yacine: I think that's what I really hope happens, is I hope that the crash in cost helps enable people to create their own anime and be able to own the whole production process, such that they capture, like the actual creatives themselves, capture more of the value, rather than the engineered corporate system that is designed to extract value from these creative folks. A really good example, like a really good analogy of this is, Dingboard itself would not exist
for GPT-4. So if GPT-4 did not exist, I would not be able to learn how to write shaders, to build an image editor in a game engine from scratch on a web browser. It is not normal that I am able to claim that I can do that in two weeks. So this new version of DingBoard, I rewrote the whole user interface in the user interface library. In November, it was my third rewrite, I think. And the fact that I'm able to do that as a single developer, I am able to write. So I have these Twitter scripts.
I'm not automating and advertising. What I do is to, I use it to automatically accept message requests, and then it launches me directly into the message so I can directly respond to people. So I've been able to use GPT-4 to help me write code that can help me do support faster.
Um, so what I can do then is I can bootstrap my company, whereas otherwise I would have had to call up these VCs and tell them, Hey, I'm going to turn ding board into a robots and then we're going to take over the world and it's going to be $1 trillion Tam. And instead of me having to do that, I can just like, you know, get the power of five, you know, software engineers by augmenting myself with GPT four. And, um, then I get to capture all of the value as a creator of the, of the software. And I get to have a closer personal relationship and capital is less of a
follows the list of a problem. So people just generally overall end up becoming happier. So that's what I actually hope happens. And I think a lot of the really early anime creators really did it because they.
for GPT-4. So if GPT-4 did not exist, I would not be able to learn how to write shaders, to build an image editor in a game engine from scratch on a web browser. It is not normal that I am able to claim that I can do that in two weeks. So this new version of DingBoard, I rewrote the whole user interface in the user interface library. In November, it was my third rewrite, I think. And the fact that I'm able to do that as a single developer, I am able to write. So I have these Twitter scripts.
I'm not automating and advertising. What I do is to, I use it to automatically accept message requests, and then it launches me directly into the message so I can directly respond to people. So I've been able to use GPT-4 to help me write code that can help me do support faster.
Um, so what I can do then is I can bootstrap my company, whereas otherwise I would have had to call up these VCs and tell them, Hey, I'm going to turn ding board into a robots and then we're going to take over the world and it's going to be $1 trillion Tam. And instead of me having to do that, I can just like, you know, get the power of five, you know, software engineers by augmenting myself with GPT four. And, um, then I get to capture all of the value as a creator of the, of the software. And I get to have a closer personal relationship and capital is less of a
follows the list of a problem. So people just generally overall end up becoming happier. So that's what I actually hope happens. And I think a lot of the really early anime creators really did it because they.
🧑🦰 yacine: loved the craft itself. I don't think that the first ones did it because they knew it was a cash cow. It turned out to become a cash cow, which then had like these, you know, these large organizations for what it's worth. I'm very grateful for because they've created many of the my very favorite animes that I've very much enjoyed.
I think like the, yeah, so that's like, you kind of see this already with reality TV shows. I watch a lot of Minecraft YouTubers, like a lot of what, like that's what I'm gonna do after this podcast. I'm going to go make dinner and watch Minecraft YouTubers.
And it's this, and they own the whole, they own the relationship with their customer. And it allows them to like, it's so much more capital efficient and allows them to like create so much more value with so much less resources. It's actually quite incredible. We already see this distribution happen with music and stuff. It's got these interesting, weird like knock-on effects. I think it's the cause of what we call stuck culture where, you know.
The last common song that me and you know are probably is probably Drake or something because he was like one of the last artists before the decentralization of like media distribution. Now it's like I listen to weird hyper pop that no one's listened to. And you probably listen to like some weird shit. No one else is listening to because you like been put into like weird little cluster. Yeah, there you go. Right. It's like that's the last you know shared piece of cultures, we're getting like super decentralized but like it's like this distributive effect of technology. I'm very, you know,
I think like the, yeah, so that's like, you kind of see this already with reality TV shows. I watch a lot of Minecraft YouTubers, like a lot of what, like that's what I'm gonna do after this podcast. I'm going to go make dinner and watch Minecraft YouTubers.
And it's this, and they own the whole, they own the relationship with their customer. And it allows them to like, it's so much more capital efficient and allows them to like create so much more value with so much less resources. It's actually quite incredible. We already see this distribution happen with music and stuff. It's got these interesting, weird like knock-on effects. I think it's the cause of what we call stuck culture where, you know.
The last common song that me and you know are probably is probably Drake or something because he was like one of the last artists before the decentralization of like media distribution. Now it's like I listen to weird hyper pop that no one's listened to. And you probably listen to like some weird shit. No one else is listening to because you like been put into like weird little cluster. Yeah, there you go. Right. It's like that's the last you know shared piece of cultures, we're getting like super decentralized but like it's like this distributive effect of technology. I'm very, you know,
👩🎤 Ate-A-Pi: My last good song was the Spongebob Glorb song. That was...
🧑🦰 yacine: I have a friend who like always like jokes like, oh, like you're so, you're so techno optimistic as a person who works in tech and derives a lot of value from it. But I am, I am pretty like techno optimistic. So hopefully that like relation to Dingboard is helpful.
👩🎤 Ate-A-Pi: On another note, do you think a lot of people in tech are actually techno pessimist?
🧑🦰 yacine: Um, I think so. I've met a lot of people. It's really funny, dude. Um, you meet these people who are like incredibly smart and then you start talking, I mean, generally like pretty much everyone's very into ML and AI and the direction that it's going. And when it ends up happening is like, I see this like thousand mile stare. They like just like, they like opened their eyes like super wide and you're just like staring off into distance. Um, because they realize about the implications of like AI.
So it's like pessimistic in a way. I think that like, it's kind of hard for me to like judge the distributions, but I think it's easy to be optimistic when you're like making $2 million a year and you're like in your Waymo and your Waymo is like effortlessly dodging all of the poor folks in San Francisco who are like addicted to drugs and you're like going through a 2 million tech job and you're like, oh my God, I live in like a post-scarcity world, hell yeah. It's really easy to feel that way when you're like a very clean Waymo. But I think like generally.
So it's like pessimistic in a way. I think that like, it's kind of hard for me to like judge the distributions, but I think it's easy to be optimistic when you're like making $2 million a year and you're like in your Waymo and your Waymo is like effortlessly dodging all of the poor folks in San Francisco who are like addicted to drugs and you're like going through a 2 million tech job and you're like, oh my God, I live in like a post-scarcity world, hell yeah. It's really easy to feel that way when you're like a very clean Waymo. But I think like generally.
👩🎤 Ate-A-Pi: Yeah, I noted that on your bros, is this post-scarcity? I asked my friends as the Waymo effortlessly navigates through the rough part of SF where people are dying of sepsis towards my $2 million comp AI researcher job.
🧑🦰 yacine: Yeah, exactly. Like, like bros. Dude.
🧑🦰 yacine: Dude.
🧑🦰 yacine: It's actually like pretty funny. Cause it's like funny cause it's true, right? Like it's like, I was like, so I mean, I mean, that tweet was inspired because I was hanging out with my friend and we were going to meet other friends at a bar in San Francisco.
And we just like walk through the rough part of town. We're like two, two burly guys. So like it was pretty fairly, fairly like safe. I didn't feel that sketched out, but like the juxtaposition is just so funny. Like it's very, very interesting. Like if you walk through like the tenderloin or like the edge of tenderloin into Hayes Valley, it's like the most, one of the top tenant experiences you can have because you go from those like horrible, you know, hellish landscape of squalor and death. And then you go into like this pretty beautiful, serene, hipster neighborhood, and you get like Greek food to your friends.
fascinating. San Francisco is fascinating.
And we just like walk through the rough part of town. We're like two, two burly guys. So like it was pretty fairly, fairly like safe. I didn't feel that sketched out, but like the juxtaposition is just so funny. Like it's very, very interesting. Like if you walk through like the tenderloin or like the edge of tenderloin into Hayes Valley, it's like the most, one of the top tenant experiences you can have because you go from those like horrible, you know, hellish landscape of squalor and death. And then you go into like this pretty beautiful, serene, hipster neighborhood, and you get like Greek food to your friends.
fascinating. San Francisco is fascinating.
👩🎤 Ate-A-Pi: uh, yeah, person who is automating jobs. Yes. Fuck. Yeah. Let's go techno-optism. Going to auto my Waymo. The same person who got automated one year later. What the fuck? This fucking sucks. Where's my union?
🧑🦰 yacine: Yeah, I mean, I made that tweet after like the whole like Devon thing popped off, which for what it's worth is like very impressive marketing from their point. Here I am talking about on the podcast. But in general, people who like got like really upset about like, oh, like, you know, automation for programming is coming. Oh, no, I'm going to die and not be able to pay my bills. I have a general advice for people who are like really afraid of automation. If you make your job fixing problems.
You will never be automated. And if you are automated, if we've automated the thing, if you've automated the job of fixing problems, then we can automatically fix the problem of you not having a job. So just focus on fixing problems for people.
You will never be automated. And if you are automated, if we've automated the thing, if you've automated the job of fixing problems, then we can automatically fix the problem of you not having a job. So just focus on fixing problems for people.
👩🎤 Ate-A-Pi: Yeah, I think, you know, I saw that as well. If you're worried about Devon automating your programming job, maybe it's time to make your job fixing problems instead of rent seeking on your ability to code.
🧑🦰 yacine: Yes, I generally think that sometimes programmers have a little bit too much hubris and overestimate their self-value and connect it too much to their ability to program. And I think that programming inherently is like, like the actual act of picking a programming language and writing out code and stuff. It's...
I don't know, man, like don't be too attached to it. Like I think the actual thing that matters is like fixing a problem. So like I could have been building Dingboard in isolation and like being like, oh, like cool ML models and stuff. But if I ship an ML model and I don't like actually make a feature that makes my users happy, then it's actually just like a waste of my time. And I'm basically just like having fun and I shouldn't expect to derive any value from it.
I don't know, man, like don't be too attached to it. Like I think the actual thing that matters is like fixing a problem. So like I could have been building Dingboard in isolation and like being like, oh, like cool ML models and stuff. But if I ship an ML model and I don't like actually make a feature that makes my users happy, then it's actually just like a waste of my time. And I'm basically just like having fun and I shouldn't expect to derive any value from it.
👩🎤 Ate-A-Pi: Right on. I'm gonna ask you, there's a Peter Thiel comment question. What do you believe that no one else does?
🧑🦰 yacine: What do I believe that no one else does? Uh... What do I believe that no one else does? Uh, I...
🧑🦰 yacine: Um
🧑🦰 yacine: I try not to have too many beliefs. So I actually don't know if there's anything that I actually believe that no one else does. I believe some stupid things because there's a lot of utility in having that belief. For example, if you actually corner me and ask me if GPT-4 or any language model, for that matter, is conscious, I'll say yes. But if you also ask me if my computer right now not running any models is conscious, I will also say yes. And if you ask me, is a banana conscious, I will say yes. But a banana is more conscious than a rock. It seems like that, right?
one thing I believe that no one else believes. It's very funny to like actually, it's like ish, it's panpsychism ish. It's like the whole consciousness thing and the sentience thing confuses me a lot. And the only way I can reconcile that confusion is by basically believing that anything that processes information is sentient. So.
one thing I believe that no one else believes. It's very funny to like actually, it's like ish, it's panpsychism ish. It's like the whole consciousness thing and the sentience thing confuses me a lot. And the only way I can reconcile that confusion is by basically believing that anything that processes information is sentient. So.
👩🎤 Ate-A-Pi: Panpsychism.
👩🎤 Ate-A-Pi: Anything that processes information is sentient.
🧑🦰 yacine: It has to be, dude. That's all I'm doing. Ha ha ha.
👩🎤 Ate-A-Pi: I'm kind of with you there. OK.
👩🎤 Ate-A-Pi: I saw that Elisa Yudzkowski had blocked you. What happened there? Have you had any interactions with him? What has been the interactions with the Doomers in the last year, I guess?
🧑🦰 yacine: Yes.
🧑🦰 yacine: So I think the first thing, this was very long, very much long time ago, so Rocco is pretty, like Rocco is pretty, is kind of like a, I'll call him, who's the lights year's
antagonist. Who's the protagonist of the lights here antagonist? I don't know. But anyways, like Rocco is like Rocco is like the mini lit here. And me and him had a Twitter debate about, you know, yak versus, you know, E slash stop, wherever they're calling themselves these days. And the whole
antagonist. Who's the protagonist of the lights here antagonist? I don't know. But anyways, like Rocco is like Rocco is like the mini lit here. And me and him had a Twitter debate about, you know, yak versus, you know, E slash stop, wherever they're calling themselves these days. And the whole
👩🎤 Ate-A-Pi: So I'm going to just interrupt you for a second. And we're talking about Rocco of Rocco's Basilisk. And I guess he had a forum post about 10 years ago, where he said, well, you better be good, because if not, in the future, an AGI is going to come and abuse you if you're talking bad stuff about AGI. This is my perception of what it was. So he said that. And that was Rocco's Basilisk.
🧑🦰 yacine: Yep.
Yeah, that guy.
Yeah, that guy.
🧑🦰 yacine: Yes.
👩🎤 Ate-A-Pi: which is that an AGI in the future would come back looking for you. And so you should be well-behaved towards AIs and AGI's. And that was basically how like Elon met Grimes by talking about Rokos Basilisk at some point. So, and Elisa is kind of the, one of the leading lights of the EA movement, effective altruist movement.
🧑🦰 yacine: Yes.
🧑🦰 yacine: Yes.
🧑🦰 yacine: Fascinating. Interesting.
👩🎤 Ate-A-Pi: And they started to get worried about out of control AGI about 15, 20 years ago. And he's been, uh, he's a character. He writes a lot about it, I guess. Uh, he's written a lot about it. And now that it's finally in the public eye, he's got a much bigger platform. Uh, he, the surprising thing for me is Aliza is actually a transhumanist. He's actually very, very pro technology, but he's so he's not a Luddite, but he's very worried about.
AGI in particular, I guess.
AGI in particular, I guess.
🧑🦰 yacine: It's like, yeah, it's like pretty, he's a, I like it, let's hear, he's like super, super consistent. It's kind of easy for me to model him and I like that.
And I think people take this stuff way too seriously. So I think the reason Leicester probably blocked me is because I was being a dipshit on the internet, which is pretty often. But then I apologized and he unblocked me. So I actually like him. I think he's kind of cool and funny. I think that you shouldn't take his views too seriously. It's very easy to not act optimally if you get paralyzed by fear. I think that we're probably all gonna be fine, to be honest, I'm not too worried. I'm a believer in this little takeoff. But I like these folks, they're really fun. They're hilarious.
If you kind of think about it, I want to call them effective altruists, by the way. I would call them the doomers. But the doomers are basically...
part of the same coin as the EAC folks. They're just internet nerds who are internet addicted and they're hypothesizing about the future and getting really excited about it, either excited or afraid, but definitely interested about it. So that's kind of how I view that stuff, where it's like fun, it's all fun and games. I think some people take it a little bit too far. I think sometimes corporations co-opt it because it's good for their forever bottom line.
Um, but like, I think like the, the folks are just like, you know, doing that stuff in isolation or mostly just having fun. Um, like I'll give you an example. Like I, I basically did this debate with Rocco on Twitter space and you see this like, you know, debate happened all the time with like, you know, Holtz, like, you know, George Hotz and, uh, not, not Holtz, George Hotz. And, um, you know, Connor, yeah. And Lazier, Connor, whatever, like one of those, you know, like nerds talking to each other and like smoking weed or something. And, um, I think like.
And I think people take this stuff way too seriously. So I think the reason Leicester probably blocked me is because I was being a dipshit on the internet, which is pretty often. But then I apologized and he unblocked me. So I actually like him. I think he's kind of cool and funny. I think that you shouldn't take his views too seriously. It's very easy to not act optimally if you get paralyzed by fear. I think that we're probably all gonna be fine, to be honest, I'm not too worried. I'm a believer in this little takeoff. But I like these folks, they're really fun. They're hilarious.
If you kind of think about it, I want to call them effective altruists, by the way. I would call them the doomers. But the doomers are basically...
part of the same coin as the EAC folks. They're just internet nerds who are internet addicted and they're hypothesizing about the future and getting really excited about it, either excited or afraid, but definitely interested about it. So that's kind of how I view that stuff, where it's like fun, it's all fun and games. I think some people take it a little bit too far. I think sometimes corporations co-opt it because it's good for their forever bottom line.
Um, but like, I think like the, the folks are just like, you know, doing that stuff in isolation or mostly just having fun. Um, like I'll give you an example. Like I, I basically did this debate with Rocco on Twitter space and you see this like, you know, debate happened all the time with like, you know, Holtz, like, you know, George Hotz and, uh, not, not Holtz, George Hotz. And, um, you know, Connor, yeah. And Lazier, Connor, whatever, like one of those, you know, like nerds talking to each other and like smoking weed or something. And, um, I think like.
👩🎤 Ate-A-Pi: Yeah, George Hutz and Eliezer. Yeah. Connor. Yeah.
Reply