Astro and SST, Humans Are Amazing and Awful, and Superconductor Optimism
Well, I finally figured out that link doesn't change. So like
Speaker 2:It's just the same link every time. Okay. So for the
Speaker 1:past 45 episodes, the process has been, I get on and I wait for you to DM me the link. And I haven't noticed in 45 times that the link doesn't change.
Speaker 2:That's awesome. I love
Speaker 1:it. And today, I figured it out.
Speaker 2:Cool. So you can just bookmark that thing
Speaker 1:Mhmm. And just
Speaker 2:show up. That's one less step. I don't have to copy the link and send it to you now. I love that actually.
Speaker 1:We're gonna be so much more efficient now.
Speaker 2:Oh, look at us. We'll have more time to talk about such interesting topics such as Astro. Yes. Oh, that's a great one, actually. Yeah.
Speaker 2:I was just wondering if you had thought about what we talk about because I have not.
Speaker 1:Well, share share what you're tell me more about what you're
Speaker 2:thinking and exploring. Yes. So, I mean, been using Astro a lot and SST. And there's just there's some features that Astro they're they're moving fast over there at Astro. And you guys you guys are also moving fast, but you have a lot of different things you worry about that are not just Astro.
Speaker 2:Yeah. So I think it hit me this week that, like, I'm in a good position to push some stuff because I care about this stuff. It's in my critical path. Yeah. So Astro, you pointed out just this morning, has already launched the code splitting stuff.
Speaker 2:So instead of Astro shipping your Astro site as one big function, now you can split it up by page, by route, or whatever. And that means there are cool opportunities now in in in SST to take advantage of that and bust it out into separate functions. So that's cool. Also, Lambda has streaming now. Mhmm.
Speaker 2:Well, that's been a while, I guess. But that's some low hanging fruit that I'd love to add because ASTRO supports streaming. So SST plus ASTRO streaming on Lambda. It's like all these little things that I feel like just chipping away at these people who are like, look how awesome our thing is. I'm not gonna name names.
Speaker 2:We do streaming and stuff. Yeah. Well, so do we over here in AstroLand and SST land. What was the other thing? There's one other thing.
Speaker 2:Go ahead.
Speaker 1:You said streaming, code splitting, and that's a good idea. Oh, redirects. Redirects.
Speaker 2:Redirects. Yes. So two Astro two dot nine has this redirect config where you can it looks a lot like Next. Js where you can configure redirects,
Speaker 1:which is
Speaker 2:a lot better than just creating files that have, like, the fences and a redirect. Yeah. Which is what we have right now, a bunch of files for every single redirect we want. We can move that into this config. So, yeah, there's another opportunity.
Speaker 2:There's just like these three little low hanging fruit I feel like I could come in and just like, boom. One PR, maybe three PRs. I don't know. Whatever you think. Hit all this stuff at once.
Speaker 1:Yeah. I mean, this is exactly the model that we're aiming for. We can get the initial thing working but because we're not heavy users of any of these frameworks, we can only like, it's just not aligned. Like, we can't Yeah. Do it as well as someone actually using it.
Speaker 1:So so exactly what we hope for that there eventually is someone that is really motivated to get this stuff working and then they get it working for themselves and in that process for everyone else. And I think that's why like this is a great example of a place where it's just a crazy advantage for this kind of thing to be open source. If you look at, like, equivalent places to Hostess, like, Vercel or Netlify or whatever. Whenever whenever Astro releases a new feature, suddenly some PM on their team is like, oh no, like we gotta support this, like where do we fit this in the road map? Like other people building iterating on their framework just creates work for them that they directly have to solve and they might Yep.
Speaker 1:They have the same problem as us. They might internally not be heavy users of these things. So they might not even know about these features and they probably don't even hear about it till someone expects us to work, etcetera. So yeah, this is exactly exactly what we want. So definitely excited for you to try that.
Speaker 1:The the one thing I'll point out though is
Speaker 2:I'm not smart enough. Sorry. Jumping jumping ahead. Oh, something else you're gonna
Speaker 1:You're point not supposed to do the insults for me. That's that's taking away
Speaker 2:the critical thing that I
Speaker 1:that I provide here. Let rely on you to insult you. You don't need
Speaker 2:to
Speaker 1:insult I'll I'll wait for it. Yeah. The route splitting thing. So just to for the for the audience just to explain what it is again.
Speaker 2:Right now Not me. I totally know. Go ahead.
Speaker 1:I mean, you I I do think you know.
Speaker 2:I think I do. Yeah. That was a stupid self deprecating
Speaker 1:Yeah. Stop. Stop insulting yourself. I will handle that. You are gonna be
Speaker 2:bad insulting Okay. I am. It's true.
Speaker 1:Good deal for you. So right now, Asho ships as a single function, which is fine for most situations. But as your application gets larger with lots of dependencies, it's just a lot to put into a single function. So they announced they launched code splitting, which I actually didn't even notice. I I've been like eager for that feature.
Speaker 1:But, yeah, it seems like they launched it and and I missed it. So now it can split one bundle per route, so you can deploy separate functions. The issue that though is, CloudFront by default only allows 10 behaviors per distribution.
Speaker 2:25.
Speaker 1:Was it 25? Okay.
Speaker 2:Yeah. I just had it raised to 50. That's how I know.
Speaker 1:Okay. Was gonna ask. So one, you can get it raised and two, was the amount they raised enough for you?
Speaker 2:It well, not when we had the redirects in. Somehow the redirects did create more behaviors. I didn't look at them closely. But it was still we were still hitting the limit and then we had to cut the redirects out
Speaker 1:and Oh, okay. But so ignoring the redirect, just a normal route splitting for the number of pages you have, is 50 enough?
Speaker 2:So it's it's not even just the pages. It's like every file in your public every file in the root of your public directory. So all the static assets you wanna ship, if you got images or fonts or whatever, all of those, the top level. So, like, if you have individual files in that public directory, each of those gets its own cache behavior. So it blows up really fast.
Speaker 2:And then folders also like, you could put everything in a subfolder within public, it's just one cache behavior. Just the nature of the way it's it's a CloudFront thing. It's stupid. I I never really thought about this that hard until I ran into this. But, like, you can't define a path that's like some complicated glob where it can handle, like, regex style.
Speaker 2:Like, if it ends in these, it's like every if you wanna have, like, an extension, like every PNG and JPEG and and GIF or whatever, you've gotta do, like, separate behaviors for each of those. So it does it blows out really fast.
Speaker 1:Wait. But why why do static assets need to be split in that way? Aren't they all served under the same policy?
Speaker 2:I have not actually looked at the behaviors that SST is deploying into my my CloudFront distro. I haven't looked on that closely, but I thought I assume that's what it was. I I read something I thought in the SST docs even that said the number it was like a little warning that, like, the number of files in your public directory will you'll you'll hit this cache behavior limit.
Speaker 1:Oh, in the public direct okay.
Speaker 2:In the public directory. So static assets that you you serve, I guess those just get put in the bucket and exposed directly and not going to like the ASTRO node server.
Speaker 1:The clean way to limit the number of behaviors you need is you make a subfolder for all like assets, right? So if you do like assets slash all of your stuff, that's one behavior that applies to all your static assets because a pattern can just be the assets. Yeah. But if you have a bunch of random files in the roots of the public folder, then there's no like clean way to specify a pattern. So, yeah, that might be why and that's that that is kind of annoying.
Speaker 2:We we moved them around. Like we had them all in the root and then we realized, yeah, we just put them all in subfolders. Right. We already had a bunch of subfolders too and each of those gets a cache behavior. So initially, yeah, we blew past the 20 or 25 whatever it
Speaker 1:was. Interesting. Yeah. I wonder yeah. You might be able to come up with just like a better way to think about some of this stuff as as you like as you go through it.
Speaker 1:But yeah, so I I was actually referring to the stuff that actually needs SSR. So each SSR route. Like, even then, like, are you under 50?
Speaker 2:Are you sure that each so each route actually gets a new cache behavior? Are you sure? No. Those all just like hit the default behavior and they flow through to the ASTRO node server. Right?
Speaker 1:Well, then that then they would all go to the same function.
Speaker 2:Oh, you're saying in this new world with splitting?
Speaker 1:Yeah. Yeah. Yeah.
Speaker 2:Oh, I hadn't thought
Speaker 1:about it. What the blows Oh, interesting. Yeah.
Speaker 2:So CloudFront routing is the the function?
Speaker 1:Here's the annoying thing. So, I think for the first bit, CloudFront routing might help. So, again, just to clarify, we're getting into a lot of details here but hopefully, someone finds this interesting.
Speaker 2:Someone on the CloudFront team is like, oh, this is good. Yeah.
Speaker 1:CloudFront is something called CloudFront functions. These functions can run for one millisecond. That's their limit, their runtime limit. They're just meant for like quick logic and they can kinda when a request comes in, they can change certain properties
Speaker 2:of limit? How many milliseconds?
Speaker 1:One millisecond.
Speaker 2:Oh. Oh, no.
Speaker 1:And it's not even JavaScript. It's like like a extremely limited subset. It's it's like an embedded scripting language I can put into your CDN to change what happens when a request comes in. They can't change the origin that it goes to. So if they could change the origin, could I basically build a router there that just says, okay, this request goes to this function URL, this request goes to another function URL, etcetera.
Speaker 1:But it doesn't do that, unfortunately. You can't
Speaker 2:do So then how could it even do the redirects? In my mind
Speaker 1:I think you can't change the origin, but with the redirect, you're not changing the origin, you're changing the path. That's a good URL.
Speaker 2:The path and the origin. Oh, yeah. Yeah. Okay. Yeah.
Speaker 2:It's still same you save
Speaker 1:right away with the 302 or whatever 30 whatever it is. So I think you can do that. So it's there's just like a weird limitation there. So So
Speaker 2:what's the answer, Dax? Just raise your limit of cache behaviors? I mean, it comes with a it comes with a performance implication. Right? Like, those cache behaviors, the more of them, presumably, that's, like, hurting performance by itself.
Speaker 2:Right? That's what the CloudFront, like, docs seem to suggest. Right. It's like yeah. It's like, don't raise this limit too high.
Speaker 2:Like, you can increase it, but, like, there's a real world cost to that. And I guess that makes sense.
Speaker 1:Well, because the like, the routing pattern matching is like a whole deep, like, world of making that really efficient. So I imagine Yeah. And people always complain like people like, know, Jared from Bun always complains about like because he got into like implementing that for Bun. He's always like, oh, these certain features like certain wildcard feature like blows up the performance of certain things. Yeah.
Speaker 1:So I don't know if there's like that much of a I think from like the micro optimization standpoint, like they're probably for those people it's like
Speaker 2:But not a huge a
Speaker 1:100 x worse with certain features but I I would be surprised if like, you felt it as as someone's
Speaker 2:been maybe this about. Seems to be a higher default limit. I I actually have the ear of someone on the CloudFront team who's in my DMs right now and was asking for some feedback. So I might suggest, hey, maybe that limit could be a 100 by default. What are the real world implications?
Speaker 1:Yeah. So we were asking for the as suggesting the same thing because because of that. But yeah, I I'm wondering if there's like some clever way to structure your astro project that limits the need for this. I also wonder, like, with route splitting, your performance might be worse because there's just more cold starts overall.
Speaker 2:So you
Speaker 1:have to really understand whether route splitting if it lowers each of your functions by, like, 10% in size, it's probably not worth it. If it lowers No, right. By like a lot, then it's very worth it. So It's
Speaker 2:gotta be like you've got a really heavy dependency on one certain route
Speaker 1:Yeah. Or And it makes sense just to split that that one route out and Can't we do that. But we were thinking that in the SST adapter for ASTRO, you could potentially specify what you actually want split. But I haven't looked too deeply into that again just because this is all theoretical for me. I'm not actually building Yeah.
Speaker 1:Anything with this.
Speaker 2:Yeah. Because you could still well, I guess assuming Astro in the build process when you build with splitting turned on, assuming they still give you the original, like
Speaker 1:Mhmm.
Speaker 2:Fat function, then you could SST could opt in to some of the new functions. You know what I'm saying?
Speaker 1:Yeah. I don't yeah. Exactly. I don't it's there's a question of can you do this from the adapter? Like, does Azure give you everything you need to
Speaker 2:do from the adapter? It might it might not. So
Speaker 1:yeah. I mean, you can also rebundle. You can, like, build your own, like, router
Speaker 2:and rebundle was gonna say,
Speaker 1:could you just say
Speaker 2:bundle up the multiple functions? Wow. Yeah.
Speaker 1:Yeah.
Speaker 2:We're getting really in the weeds here. This our audience for this is like Fred and someone from the CloudFront.
Speaker 1:Yeah. Exactly. I heard you're talking to the Astro team tomorrow.
Speaker 2:I heard you're coming too.
Speaker 1:Yeah. I'm gonna be there was like,
Speaker 2:not that guy. Please not that guy. I love it.
Speaker 1:I have two meetings with them tomorrow. Isn't that crazy?
Speaker 2:Oh, wow. Look at you. Yeah. It's becoming part of the ASTRO core team. This just make you an honorary member of the team.
Speaker 1:Yeah. It's like I again, just I never have I never have meetings. And I had, like, three meetings in one day the other day. Oh, jeez. This Friday.
Speaker 2:You got your Sunday meetings.
Speaker 1:My Sunday meetings. Toxic. Meetings. It's like a it's like a to me, I think of it as some kind of like blood sacrifice. Like, do a blood sacrifice where I do one meeting on Sunday for no meetings the rest of the week.
Speaker 2:Oh, that's a
Speaker 1:good idea. Yeah. It's But worth
Speaker 2:then you actually have a lot of meetings. Sorry.
Speaker 1:Yeah. Occasionally, like very rarely. My it was just like all external people. My previous role, I had minimum three hours of meetings every single day.
Speaker 2:Daily? Yeah. Oof. Minimum. I mean, I've been I've been at places where it's the meeting culture.
Speaker 1:There's a lot of that. Yeah. I had nothing else to How
Speaker 2:has that not died yet? Like, how is that not like these companies that like half your day is in meetings if you're like an IC, like you're contributing code. How how does anyone think that's a good idea? I don't understand how
Speaker 1:this I wasn't an IC at that company. Like, was Okay. I wrote zero code. I wrote like basically zero code. Yeah.
Speaker 1:So half my day was meetings, which I think for a manager of entirely and purely management role
Speaker 2:Yeah. That makes more sense,
Speaker 1:I guess. It's like not terrible.
Speaker 2:You're in one on ones and you're in well, I I don't know what your role was.
Speaker 1:No. I Direct Basically your one on one roles. One on one roles. One on one meetings with everyone on my team every week. Yeah.
Speaker 1:So I spread them around. Those were fun.
Speaker 2:Leadership meetings.
Speaker 1:The one on ones are fun because they're just like getting to fuck around with people. Yeah. So those are not bad.
Speaker 2:Lower pressure.
Speaker 1:Yeah. And then there were the other kinds of meetings
Speaker 2:Also much more active though. Sorry. Like, you can't just like chill and do your work while you're on those calls.
Speaker 1:Yes. You're actually having to talk. Yeah. To be honest, I think almost all my meetings were fairly active. Because again, it's only work that I did.
Speaker 1:So I
Speaker 2:was not a director. So Director. My meetings were mostly coding and half listening. They're kinda listening. Yeah.
Speaker 2:And then they're like, what do you think, Adam? Or do did you wanna story point this one? Like, I do not, actually. Let's go with five. Probably like five.
Speaker 2:I don't know. Roughly five story points, please.
Speaker 1:Yeah. Yeah. Then the other category meeting was and this was what I was fantastic at and this is why I think I was good at this role. It was, meeting with other people that led other departments and making them feel like, okay, engineering knows what they're doing even though everything is going wrong. Just just like making one feel positive and confident and like
Speaker 2:Yeah.
Speaker 1:Oh, yeah. Like, totally.
Speaker 2:Like I think I'm good at that too, by the
Speaker 1:way. Yeah.
Speaker 2:Maybe that's why we
Speaker 1:have a podcast here. Yeah.
Speaker 2:That's that's mostly your job. Is making it Everything is
Speaker 1:fine. Everything Yeah. No matter what We
Speaker 2:are competent. We know what we're doing.
Speaker 1:Yeah. I mean, it like manifests into reality because like
Speaker 2:Oh, yeah.
Speaker 1:The when people stop feeling that way, they start to need to stick their fingers into stuff and that either makes stuff go even worse.
Speaker 2:Yeah. There's there's a reason for it. Can I ask you a question, on the record? Like, you deal with all these, web frameworks, ASTRO, Next. Js, Solid Start, SvelteKit, whatever.
Speaker 2:You guys support them all. And you deal with all those people, a lot of external meetings I hear. Do you have a favorite? Who's your favorite? Like, who do you like working with the most?
Speaker 1:And it's
Speaker 2:like building with? Yeah. Like, just like as a team, like, really like interacting with that team. They're great.
Speaker 1:It's probably gonna be Astro mostly because they are the ones that actually have a company behind it. So there's there's like like that's everyone's day to day job. That's the thing that they're focused on. So it's very easy just on a base level, it's very easy to work with them because we're also working on this as like a real job.
Speaker 2:Yeah, that's your job. Yeah.
Speaker 1:Yeah. So it's like a very clean collaboration. And beyond that, they're just just personally all great. Like always really excited to help
Speaker 2:and Yeah.
Speaker 1:And it doesn't really
Speaker 2:Why yeah. Why do you hate Next. Js? Why do you hate those people so much?
Speaker 1:Don't even we don't even work So with
Speaker 2:Can I Oh, I guess you don't? Open Next. You just kinda do your own
Speaker 1:I mean, we'd like to, but it's just
Speaker 2:Oh, well, put the word out. I'm sure somebody one of them, someone who works at Vercel has heard this podcast. We've had, like, episodes titled Vercel. So I'm sure someone in our listenership works there. And if, you know, this is an Olive Branch.
Speaker 2:Dax is reaching out. They'd like to work with you.
Speaker 1:It's not an Olive Branch. It's actually something quite
Speaker 2:different. It's an
Speaker 1:Olive Branch. It is please take this project off of our hands and just do it instead of us. Oh. Please just do it. Please, like, just take it.
Speaker 1:Like, we don't want this project. You take it. You take it
Speaker 2:and mean, would they take would they take OpenNext or would they are you saying, like, they should open source the equivalent of OpenNext? Is that what you're saying? Or no, take over OpenNext and just make it
Speaker 1:Just first conceptually make it so that OpenNext doesn't need to exist or needs to exist less.
Speaker 2:Yes. I gotcha.
Speaker 1:Our relationship with every other framework is is pretty different. We do very little in the adapter. We do like a little bit of work in the adapter and we just get it working AWS. With Next. Js, it's like and the thing I always say is if you combine our efforts on every other framework and combine it all together, it's still a fraction of the amount of time we spend on Next.
Speaker 1:Js. So Yeah. It's just it's just pretty different.
Speaker 2:I mean, it's telling that, like, you look at the SST organization in GitHub and you guys have the SST repo, which has, like, ASTRO SST and all the adapters for all the other ones. And then OpenNEXT is a separate repo. Yeah. It's like a whole big thing.
Speaker 1:A separate repo I never look at. I never never look at.
Speaker 2:Does is it Frank? Does Frank live in there?
Speaker 1:Yeah. Frank dies in He goes there and he dies and then we have to resuscitate him and he comes back and then we send him back in there. Oh, man. Yeah. It's I'm
Speaker 2:very aware of all that.
Speaker 1:Yeah. Asher is the one we've worked the most with. I think, the other frameworks also, everyone has been really just everyone's like really good to work with. I think this is my one of my favorite parts about open source. It's when and I talk about this all the time where it's not meant to be like random people submitting PRs that you don't know.
Speaker 1:Like that's that happens and that's fine. Yeah. But the best part of it is when you like establish a relationship and then you can kinda work really efficiently together because you both know each other and everyone likes each other and works well together. And it's a lot more streamlined than just submitting a PR out into the void. Yeah.
Speaker 1:So we've gone to that place with, like, with the solid team, of course. And then I mean, even with SvelteKit, like, Racharis has, like, been taught, like, replied to our PRs and everything. So it's nice to establish these relationships because whenever we need something, we're not going through, like, random person door we're going through that we have a relationship with your door that's, like, kind of frees out in faster way. So, yeah, pretty much every every other framework we have that with.
Speaker 2:How hard is it for like a just like a product company, just like company that makes some product and has a bunch of customers or users or whatever. Is it hard to sort of adopt an open source mantra where you're gonna, like, make your stuff open and and available for people who are passionate about the product to contribute? Because you just said, like, not about, like, random people submitting PRs. What if it would only be random people because there's no one really aligned that cares about your product? You're kind of like you're like an edge of the graph.
Speaker 2:You're not like, you know, SST, Astro, you guys are like in the middle of that graph. Yeah. If you're like building products for specific things you know what I'm saying? You know what I'm getting at? Yeah.
Speaker 2:Theoretically.
Speaker 1:I think yeah. I don't think that really works. And even within SST, we have portions of it that are like that. Like the core of SST, we don't accept any contributions to because it's very unlikely that some person is gonna get on the same page with us and like know exactly where we're trying to go. Yeah.
Speaker 1:If you look at the console, which is a very typical SaaS product, It is open source. Extremely unlikely that we're ever gonna get a meaningful contribution on that. Yeah. Because there's just so much. Like, you have to be aligned with, like, the way we think about product, the way we design things, the way we do whatever.
Speaker 1:We still like to keep it open source because it is a great code example of a real life
Speaker 2:It's learning.
Speaker 1:Yeah. Exactly.
Speaker 2:It's a good teaching example.
Speaker 1:Yeah. And I think that's the main reason why people should make their stuff open source. And there I guess there is an offhand chance that someone does contribute something useful, like they like take care of something tedious or like Yeah. Something that is a small cut but not high enough priority for anyone on the core team to work on. Like there's always gonna be random stuff like that that people can take care But yeah, as a learning exam, it's been actually been insane.
Speaker 1:Joked the other day, I think we have more users of our code base than our product because a ton of people have just been, like, taking that setup and, like, trying to build their own stuff.
Speaker 2:I've referenced it a lot. I mean, I've already looked at the Just console like for just best practices with SSD, like, just as I'm working through stuff and I realize I'm not thinking I still have, like, CDK brain, and I realized that, like, there's things that haven't transferred over for me in terms of how to do it best in SSD. So I just look at the console repo, and that's, like, my gold standard. Know it's the last thing you guys worked on. Yeah.
Speaker 2:I love it.
Speaker 1:Yeah. And that's the thing I I was looking for when I was first getting into serverless stuff. So I'm glad it in like a real app that because for me, the questions with serverless that I had in the beginning were never like, what's the grand theory behind this? It was more like, how do you like check for the current user? Is it like what the works for?
Speaker 1:They just like really specific things like that.
Speaker 2:And then there's the whole local first thing, like that's a great repo for learning all that stuff. I haven't even touched any of that but
Speaker 1:Yeah, exactly. I think I think that's a part I've been surprised that a lot of people because I because I we talked about this a bunch that I think more people should to do that but I acknowledge that's a big step in learning curve and a lot of upfront upfront costs of building stuff that way. But I I've seen a bunch of peep posts on Twitter of people being like, oh, look at this thing I'm messing around with, like kind of based off of our setup.
Speaker 2:Some of it looks better than our console. I've see I think I've seen some of these conversations like Sock talking about he did it with Next. Js and like Yeah. Just people it's so cool how people will take that, learn from your example, and then apply it in other ways. And, yeah, I love open source.
Speaker 1:Super cool. Yeah. And and in some way, that's gonna come back to us. So whether that's they build a product that's really good because they're using a similar setup that leads to good products and we get to use it or they come back and they're like, this pattern actually works better. I'm just kinda seeing other places testing it.
Speaker 1:So, yeah, I think more companies should just consider open sourcing their code. It's usually not your competitive advantage as much as everyone likes to tie up all this stuff with IP and and all of all of that.
Speaker 2:Like Right.
Speaker 1:When I see someone's product, I like can guess how it works for like 99% of cases. Like, I don't need to see their code.
Speaker 2:Yeah. Sure.
Speaker 1:But yeah, I think more companies should should should definitely consider that. I I am seeing that a little bit more too.
Speaker 2:As a a product company, you've never dealt with like you don't have the internal muscles for, like, dealing with a community if if it does build up. Like, if there there's a group of people that do wanna help, is that, like, a challenge too just to, like, build up the the mentality of an open source organization where you have to handle that community and triage and I don't know. I just imagine a lot of stuff involved with that.
Speaker 1:Yeah. So we like, we are an open source company and we, like, very intensely want a community and we, like, try to grow that as much as possible. Even then, like, we I don't think we do, like, that great of a job
Speaker 2:at doing all of you. Mean, there's literally three of you.
Speaker 1:Yeah. That that's the thing that suffers the most when we get busy with other stuff. Yeah. So I can definitely see that if you're, like, you're not really an open source company that, you probably aren't gonna spend any time there. Again, I think it's okay.
Speaker 1:I just don't think it matters that much because even for us, we, like, don't do an amazing job and it's still, like, fine. So
Speaker 2:It's still serving a lot of other purposes. Like you said, like, the learning aspects of it is like a baseline. That'll always be there regardless of if you're taking contributions or not or if you're is that I guess, how do you communicate it, though? Like, do people ever get upset? Like, hey.
Speaker 2:I put a PR out a week ago. Why I haven't gotten any eyes on it. What's going on? Does that ever happen?
Speaker 1:Yeah. I know. All the time because we are just
Speaker 2:Oh, yeah. I'm gonna push one today and be like, Dax, it's been two hours.
Speaker 1:Come on. Okay. This brings me to another topic I wanted to talk about. Okay? So, as what I said earlier, me and you have a relationship.
Speaker 1:So when you push a PR, I am gonna look at it more than I would look at a random PR. A lot of people might hear that and they're like, that's unfair, that's not a good system because you have like, people are getting unfair advantages, which is a 100% true. I'm not gonna say that is not true. Yeah. But for a lot of people, their ideal world and I talk me and Liz talk about this a lot with apartment hunting.
Speaker 1:Like, in New York, when you're looking for an apartment, there every building, insanely standardized process of like, you fill out this form, you tell me your credit score, you tell me your income, we have this exact way of computing whether you get it or not, and then that's the process. No one skirts that process. There's no there's no like clicking with the lamp, none of that. Just a standardized process. And that is technically fair, like prevents something extremely unfair from happening.
Speaker 1:But it also kills any opportunity for like two people to really just connect and be like, okay, this is a good situation that we wanna we wanna grow. Yeah. Whereas when we moved here in Miami, we saw this house and we talked to the landlord and like, very clearly the type of people we can because they live two doors down, like exactly the type of people we wanted to live near. They really clicked with us as well. Yeah.
Speaker 1:It was a really great like personal connection and then when we left, they called us and being like, hey, we'll lock we'll knock down the rent by $500 if you if you guys agree.
Speaker 2:That's awesome.
Speaker 1:So it's like it eliminates opportunities for like, for that to happen. And I get that like that's technically unfair in in a way of looking at it but I just don't think any extreme is is the best in that case. Like, you do wanna allow for these really personal local localized things to happen like that where something like better than like the average outcome can happen. Whereas they would have just gotten an average neighbor, now they got like a more ideal neighbor and more ideal tenant. So same with the PRs.
Speaker 1:It's kind of it works the exact same way.
Speaker 2:Yeah. But is that different? Is that a wholly different thing than like nepotism? Like where, you know, who you know kinda gets you, I don't know, a career or something else. Is that a wholly different thing than what you're talking about?
Speaker 1:No. That that I think it's the same it's the same problem. Like, another extreme is someone is, like, racist and, like, doesn't they have full control over who they make decisions around and they, like, ignore, like, they have some like crazy bias or like, you know, they just give it to someone that's that got there an unfair thing. I I like get the other extreme.
Speaker 2:So those are the extremes.
Speaker 1:Yeah. Where I feel like in New York, it's on the opposite extreme and I think a lot of cities, it's it's on the opposite extreme where it's like so overly standardized, any concept of individuality just like disappears entirely. It's hard to strike that balance when I think people react to those negative situations by taking it too far in the other way.
Speaker 2:Yeah. I I'm thinking of an example of because I've always thought about the like, I don't know, the who you know thing. Is that even bad? Is it kind of the way the world works and that's okay? There is a Jack no.
Speaker 2:Joe Buck. He's a, like, a baseball announcer. Mhmm. Not there was. He's still an announcer, I guess.
Speaker 2:I don't know. He does football too. I don't know. He's an announcer. And his dad was Jack Buck, who's like this legendary announcer for the Cardinals here in Missouri.
Speaker 2:And he he makes a lot of self deprecating jokes about nepotism because he very much got his career from his dad being this legendary announcer. But also, like, I don't know, he's got a great voice, like, genetically from his dad. Like, there's, like, a reason that they had this great announcer voice, and he's good at what he does. Like, the fact that his dad was really good at it and that's why he got his job. And I know, like, there was well documented the way he got into his career was very, like I mean, it was nepotism, like, to its core.
Speaker 2:But, like, is the world worse for it? No. He's a good announcer. So I'm I'm sure that there are extremes, like you're saying, like, cases where it plays out very badly, but it's not all bad. Like nepotism doesn't have to be a bad word.
Speaker 1:Yeah. Yeah. Exactly. It's yeah. It is tricky.
Speaker 1:Like with everything, the world is really complex and any kind of like rigid rule like that usually is not is not the right way to go.
Speaker 2:That speaking of that, because this has been very top of mind for me. When you said the world is very complex, I realized I have this worldview. And I don't know if I've communicated to you or if I've ever put enough thought to it to communicate it to anybody. But I have this worldview where, like, I don't blame anybody for literally anything.
Speaker 1:Yeah.
Speaker 2:Like, I view everything that happens, every, you know, action that someone takes or reaction. It's just like this product of, like, their entire life and every circumstance and every whatever. And, like, not that they don't have, like, free will to, like, do whatever they want, but just that, like, everyone's sort of a product of their environment, their their genetics, their upbringing, all those things. And it's just sort of, like, yeah, they were gonna take that action. So it's very hard for me to, like, place blame in any situation on a person or a group of people because I do just view it through this lens of like, I don't know.
Speaker 2:I haven't lived their life, and I don't know I don't know what led them to that point to do that thing, but it it probably all makes sense. Like the math of it in the universe probably all works out. Does that make sense? Is that way too far removed from our conversation? No.
Speaker 1:I I feel the exact same way about everything. And I think the way this manifests a lot is when I think a good example and maybe a very extreme example is, who's the Theranos woman, Elizabeth Holmes? Yeah. So obviously you see someone like that that did a bunch of stuff bad wrong. But a part of me, like, much sees how I could have ended up in that situation, like, just being myself and just being in a certain environment and, like, set up in a certain way.
Speaker 1:So I also have a tough time being, like, getting angry or like upset because I'm like, I like very like, I'm like, way you're saying it, like, you can just see how they became that way and you can see how like any like human put in that situation would have become that and then says, the other side is, I like, can see myself in that situation. I'm like, wow, that would suck. Like, I would suck to be in that situation. And like, I I I don't most things, even things that are bad, I can like see how I would have just done the same thing. Yeah.
Speaker 1:Liz talks about this a lot too, like whenever there's situations like that, she feels the exact same way. Like, she has a hard time like getting angry or upset for those reasons. So I feel the same way. It's like, don't I would have done the same thing or like
Speaker 2:I thought I didn't know how much of it was like, I I was a middle child. Like, I've got this very this personality that wants to, like, diffuse any conflict and, like, work it out between parties. I didn't know if it was that or if it was just some kind of other worldview, but I I feel like there's a name for it, I feel like you probably know it and you're just hiding it. Like, don't you wanna sound too smart, but like, oh, well, that's the crow's eye principle. Everyone knows that.
Speaker 1:No. Okay. I think it's it comes from feeling like there are the things are deterministic and there isn't really, like, free will.
Speaker 2:Is that what it is? Yeah.
Speaker 1:I mean, think that's where it's for me, it stems I can from see
Speaker 2:where you can play it out to that degree. And I don't know how I hold on to my belief that people get to choose what they do. I I still believe that, I guess.
Speaker 1:Yeah. You like I agree that it's definitely a 100% feels that way but the other side of me knows that some asteroid hit some other asteroid at a certain angle like eight billion years ago and that's why I don't have hair in my head. Right? It's like there's like
Speaker 2:there's just I just
Speaker 1:I just fully see it as if you knew all the rules of the universe, you could just predict it forward in a perfect way. So I don't really believe in free will at my core, so I think that kind of bleeds out to these places where it feels like.
Speaker 2:Yeah. Maybe I don't either. Maybe I'm just it's like trying to pretend I do because I feel like I want that control or something. Maybe I don't actually believe it. Because if I really think about what I'm saying is in contradiction to free will
Speaker 1:Yeah. And I'm just
Speaker 2:I'm just sort of like hanging on to something that I don't really believe.
Speaker 1:I think it definitely feels like we have free will and I think it's probably also not right to, like, make everyone blameless. But
Speaker 2:No. Right. It doesn't like get you out of your consequences. It doesn't change the circumstances that like actions lead to. It's still like all of those things apply.
Speaker 2:It's just
Speaker 1:I just love the empathy of empathy that it brings of like, this could have literally been me or this could have been anyone or like
Speaker 2:Yeah.
Speaker 1:A lot of things led to this besides like Uh-huh. There is like a rotten spirit at the core of this person.
Speaker 2:Yeah. No. I've been I've been up close on some like situations in the recent weeks that have led to those thoughts. This is it didn't come out of nowhere. But I it's I feel like I'm in contrast.
Speaker 2:The way I view it is in contrast to everyone else around me, which is like, no, this person's at fault. No, this person it's like their fault. And I don't know. I just view it like they are who they are and that's how they would have that would have played out a 100 times out of a 100 times.
Speaker 1:Yeah. Yeah. I feel the same way.
Speaker 2:Cool. Well, you're supposed to disagree with me. That would have been way more fun. Tell me why we have free will. So
Speaker 1:the other thing, obviously, we gotta talk about the superconductor stuff.
Speaker 2:What? What happened with superconductor stuff? I don't even know what
Speaker 1:you're talking about.
Speaker 2:Are serious? I've been way unplugged.
Speaker 1:I mention that? Insane that you're this unplugged.
Speaker 2:All I think about right now is ASTRO and SST. Like, that's for weeks. So what is superconductor? Tell me more.
Speaker 1:You're you're you're gonna you're gonna feel so bad when I describe what's happening.
Speaker 2:Oh, no.
Speaker 1:Because it's
Speaker 2:Did you learn this on
Speaker 1:Hacker It's literally everywhere. Like, there's not a place that I can escape it. It's literally just everywhere.
Speaker 2:Is it on Twitter?
Speaker 1:It's on Twitter. I made a joke about it yesterday. I was like, I don't
Speaker 2:Are you serious?
Speaker 1:I was like, I'm not ready to become an expert on superconductors.
Speaker 2:Okay. Tell me, what is what is what is a superducker? What are we gonna be
Speaker 1:talking Okay. So I again
Speaker 2:Is this the alien thing?
Speaker 1:I'm not No. It's not. It's different. Okay. It's different.
Speaker 1:How do I explain this? Okay. So there is there's a paper that came out. I guess, why are you gonna feel bad? Because if you if
Speaker 2:you spend this time obsessed with
Speaker 1:the web framework, you're gonna feel like you everything you do is completely meaningless when I and I tell you
Speaker 2:what it is. Did everything just change? You're hyping this up. I'm so excited.
Speaker 1:Okay. There's a paper that came out, and it's people are trying to replicate it in the next couple days because it's seemingly like simple to replicate, where people, these three it's kinda confusing. They're they're kinda like arguing over who gets credit. But, these people from Korea, these researchers from Korea that that they've made a superconductor that is room temperature, that's stable at room temperature.
Speaker 2:Could you tell me what a superconductor is? Or are you getting there?
Speaker 1:Again, I'm not like a super expert on this, so I'm not a super expert on superconductors. You're not a super expert on superconductors. But I I'd like seen videos of superconductors in the past where, you've probably seen them too. They're like these there's there's there's material that's used in certain places. So the word itself that gives you a hint, a superconductor means you can put energy in it and it transfers through with very very little lost.
Speaker 2:Okay. That makes sense. A conductive
Speaker 1:It's like 99.99% energy transmission. Right? So if you think about
Speaker 2:a So I just now You just now got me out of the branch that's not like a train conductor.
Speaker 1:No. No. No. For trains. So we're gonna
Speaker 2:we're gonna
Speaker 1:get back there. Okay. There's so if we think about a power line, like, a power plant can only Mhmm. Supply power to a certain distance because the heat loss from the power lines eventually, like, becomes too much. And we lose a ton of energy even with the current distance.
Speaker 1:Like, some percentage of energy is just lost in the transmission. So a superconductor can theoretically transmit energy at near infinite distances because the loss is is so so minimal.
Speaker 2:Okay.
Speaker 1:But the issue is every superconductor we found so far needed to be at some, like, crazy low temperature to operate. So, like, you need like liquid I've always seen videos of like liquid nitrogen and then putting the superconductor and like showing. Mhmm. There's other weird effect. I've got there's a name for it where it lets you like it's kind of like a magnetic effect where it hovers over over other, over like some other materials.
Speaker 2:So a superconductor hovers over other materials?
Speaker 1:Yeah. Again, I I don't know this stuff that well. I don't know what the actual, mechanism is but it has to do with something like the word superconductor implies like electrons flowing through it and something about that. There's like some effect that it creates where it can like hover over stuff. And you I feel like you must have seen this video, like imagine like a bunch of like gassy liquid nitrogen and something like kinda hovering over a surface.
Speaker 1:Have you ever seen
Speaker 2:No. You should look at my YouTube history. It is not that.
Speaker 1:There's a great video where someone built a hoverboard, that like is made from the stuff and they're constantly filling with liquid nitrogen to keep it cold enough
Speaker 2:to It's like a real hoverboard?
Speaker 1:Yeah. And they have Tony Hawk try to use it and it's so difficult because an actual hoverboard would be like it's so so so hard to ride. Yeah. So anyway, really low level material and obviously the constraint has been, it needs to be super cold, kind of expensive to produce. Someone has built so people are saying that we've now have a superconductor that's works at room temperature and ambient pressure.
Speaker 1:So you can just have it in any environment. And it's very cheap to make, theoretically. So this has implications on literally everything. So I
Speaker 2:was gonna say, this sounds like a big breakthrough.
Speaker 1:See it's like a it's a first small tiny step and like a lot needs to be done to get it to like progress it to a place where there's a lot of other variables that are Yeah. In play. But if you just let if you ignore that, it's kind of fantasize. Theoretically, you could build a crazy solar power plant across the desert in The Middle East and run power lines for the rest of the world.
Speaker 2:Oh, wow.
Speaker 1:You can, that has implications. There there's like some Does it Yeah. So
Speaker 2:I feel like this
Speaker 1:In fusion research, this solves like one, again, not it doesn't solve the whole problem, but it solves like one problem in fusion. Again, with the magnetic effect, the hovering effect, now you can build, like trains that are not Oh, wow. Like kind how the maglev trains work but again Yeah. Being more efficient. It has implications in like so like, people keep coming up with other areas where it could it could impact.
Speaker 1:So we might be on the cusp of
Speaker 2:AI super intelligence.
Speaker 1:No. It has implications in in, in semiconductors as well. GPUs and Yeah, exactly. So again, really far fetched, right, so that we're all fantasizing here. Because one, the paper might not be real.
Speaker 1:Two, there might be limitations that where it never leaves the lab, kinda like how graphite is. We might but I know if you let yourself fantasize, we might be on like a new era of like humanity.
Speaker 2:What do you think yeah. What do you think the percentage likelihood that like this pans out is?
Speaker 1:I have no idea. Like, I
Speaker 2:just I just You're an expert. No. You're or you're not a super expert, but you're expert opinion,
Speaker 1:do you think?
Speaker 2:Could you just explain conductors? If you can't explain superconductors, maybe just conductors. Let's start. You're not a super expert. Give me give me like, is it a coin flip?
Speaker 2:Is it like a 1% chance? Like a moonshot? What are
Speaker 1:we talking about? I don't know. I feel like I I I have no idea.
Speaker 2:Your intuition is good. I want to know what your gut is. Is this gonna happen?
Speaker 1:Something about it makes me feel like it's a yes.
Speaker 2:Yeah. I do feel like it checks this box for me that's like I do feel like there's gonna be these weird little dominoes that happen that just like completely alter everything. Yeah. So it checks out for that because I I immediately jumped to all these other fields that are, like, making breakthroughs. You think of AI, you think of fusion energy.
Speaker 2:I mean, both those things are way ahead of where people thought they'd be at this point in time. And you think about, like, something like this just being a catalyst. It's it's really exciting.
Speaker 1:What's really crazy about this is no new technology needed to first exist for this thing to be produced. I saw someone saying that you could have made this in the lab in, like, 1890.
Speaker 2:Do do you know anything about like the material or whatever is
Speaker 1:It's going on like some kind of lead thing. Lead.
Speaker 2:Yeah. It was lead all along.
Speaker 1:Yeah. We we Interesting. We ran away from lead and now we're running back.
Speaker 2:Presumably, it's not just lead. There's like they're doing something. They're adding something to lead.
Speaker 1:Yeah. There's some so basically, the they actually discovered this material initially in 1999, these these scientists. Uh-huh. And they were kind of and what's weird is they're not like they're not like these like really renowned scientists. They were kind of like mid tier scientists that were kind of playing with this on the side it's like a side project.
Speaker 1:And they finally got funding a couple years ago, two of them, to like explore it deeper and they sent in like a big shot scientist to babysit them. So there's three of them now. And they've they've they've been working on like refine the process to actually produce it and then now they finally figured it out. And they've been super paranoid that it was gonna leak and someone was gonna steal it because they were like so convinced that this is gonna work and it is, like, gonna change the Nobel Prize winning statues in every city type of situation, like Einstein level fame. Wow.
Speaker 1:So everything they've done indicates that they genuinely believe that's what's at stake here. Yeah. So I think the question of is it like fake is maybe I don't believe that it's fake. Does it actually work is the question. But it's so easy to produce at a bunch of labs.
Speaker 1:I'm I was actually wearing a shirt from this company. It's like a like a space company. They have all the equipment. Yeah. So they, they're engineers in the past couple and they've been tweeting about this for the past couple days.
Speaker 1:They've been trying to replicate it. And they should have everything in place by tomorrow or Saturday.
Speaker 2:Oh, wow. So we're gonna know soon if this is, like, something that could be replicated and if there's real stuff here.
Speaker 1:Yeah. Again, taking it to industrial, see that there's obviously a lot there but this Yeah. This doesn't break any physics laws. It's not like someone made infinite energy and that breaks all it's like, it follows everything.
Speaker 2:Yeah. Or like faster than speed of light or something. We've seen that, like oh, the neutrinos or whatever, yeah.
Speaker 1:Yeah, that's people brought up that paper as well where people were like, we know this is wrong but we don't know why. But they did, they published it anyway. So yeah, like this this could be a major thing. The other thing this brings up for me is, there's like this thing about us that I hate, which is actually, it's funny. Humans hate themselves a lot, right?
Speaker 1:Like, think
Speaker 2:if you like just look at I hate myself. No. Not like, looking
Speaker 1:at individual level, like, we like hate human, like the concept of humanity, hate them or like humans
Speaker 2:like Like scum, we ruin the earth,
Speaker 1:etcetera. And if you look at like any fantasy story, it's always like, there's a magical kingdom and
Speaker 2:then the humans arrived and they like, used
Speaker 1:they like destroyed it. And it's it's always like, we have this low level of hate of ourselves and I think that shows up in places where it shouldn't. So if we think about, like, we're all worried about climate change, everyone is convinced that the only way to solve it is to punish ourselves because that's kind of been, like, it just shows up everywhere. Like, in certain religions, like, the idea is you're bad and you need to, like, repent for who you are and then you get saved, you sacrifice sacrifice stuff. Stuff.
Speaker 1:Yeah. Yeah. If you limit yourself, you limit the bad parts, then, like, then there's, like, a solution for you. And I think people have this feeling with with climate change where they're, we need to stop having children and we need to stop, like, doing x y z. Need stop, stop, stop, stop, stop.
Speaker 1:And if we repent for the badness that we are and the greed that we have Yep. We will stop climate change. It's completely impractical because that's nice to say that from, like, a very developed country. Yeah. Saying all
Speaker 2:that No, means because countries Yeah. They're developing. It's Yeah.
Speaker 1:Billions of people, you're like denying them. But it's like it's like a kind of a messed up position but
Speaker 2:Oh, yeah.
Speaker 1:People believe there's like degrowth thing. Like the only way out of this is to like shrink ourselves because like, we were greedy and we grew too much. But my position has been, this happens all the time, like in history so many so many times. There's a famous one in India where everyone was all the scientists were predicting how there's a huge famine, millions of people are gonna die, it's gonna be disaster. We've hit the ceiling on like, the ecosystem can support for this population.
Speaker 1:Yeah. There's there's like a name for people that believe that think like this. It's it's like a very like natural place to think. You're like, okay, you can only support so many people. Then a scientist from somewhere in the Midwest actually, figured out a strain of, I think it was rice that had like crazy yield.
Speaker 1:It was like a multiple yield. And he literally saved like millions, billions of people because the yield was something that no one could predict. And we've just gone through this so many times where like one or two people will innovate something that was so outside the realm of any calculation. We just enter a new era of what's possible. So that's what I've been hoping for and and leaning on and this kind of I think I have a lot of bias because I want this to exist to like support Yeah.
Speaker 1:My way of thinking. So that's why I feel like it it can be the thing. But it's an example of the type of thing that can happen.
Speaker 2:Yeah. I'm definitely guilty of the other side. Like, I'm guilty of the, like, we need to stop doing things. Like, just that mentality of, like, retreating a bit. I'm definitely guilty of that.
Speaker 2:So it's nice to hear your perspective. It it does give me this optimism that, like, humans will figure it out. We will solve problems. We keep pushing. We will figure
Speaker 1:it out. It's a very natural place to go. Remember the it's called Malthusian trap, I think it's called. Malthusian or it's like Malthusianist or something like that. There you are.
Speaker 2:Pareto principle. Yeah. Yeah. Keep going. Again, swear,
Speaker 1:I don't think I actually know that many of these but
Speaker 2:somehow, the ones that I
Speaker 1:know always pop up. Look, guess I guess I've got I'm in control of the conversation to some degree. So yeah.
Speaker 2:So we're all gonna have hoverboards now. Wow.
Speaker 1:I think this video shows that hoverboards are kind of useless. No Like even Tony Hawk can just barely ride it. And if he can if he can't ride it, I'm not gonna be able
Speaker 2:ride it. But back to the future. I mean, they they look so cool and useful.
Speaker 1:Yeah. I think I think friction is just important. It's underrated. Yeah. But for train.
Speaker 1:Yeah.
Speaker 2:You can't really Maybe we'll get a Trains.
Speaker 1:Maybe we'll finally get high speed rail and
Speaker 2:I love this story. Yeah. No kidding. I love this story so much because it's like you realize all the ways we can make advancements. I don't even think about materials.
Speaker 2:Like, I don't think about material sciences and how much, like, they could just invent new combinations of things. There's probably, like, infinite combinations of things they could figure out that have important impact on society. It's like I don't know. I get I I'm in this little bubble with technology where I think about, like, all the advances we're making in technology, but there's, like, so many ways that humans can push the boundaries and learn new things. And you just start to imagine, like, aliens zipping around the universe with these materials that we are just now discovering.
Speaker 2:Like, we're little babies and we're figuring stuff out. I don't know. I love it.
Speaker 1:It's all interconnected too. Right? Like, this discovery might push the cost of compute a lot further down which then enables all sorts of things that all sorts of stuff that we just ignore because it's just not possible.
Speaker 2:Yeah. Which then, you know, we were able to do better research
Speaker 1:Exactly.
Speaker 2:Unlock stuff in fusion which leads to, like, more abundant energy and then it's just like this crazy
Speaker 1:Feedback loop. Yeah.
Speaker 2:Yeah. Yeah. Pareto principle or something.
Speaker 1:I mean, if this leads to fusion, then a fusion is super that makes energy super cheap, then we can suck carbon out of the air for really cheap. We can, like, desalinate
Speaker 2:water really.
Speaker 1:Like, there's just so many so many things. So it feels so much like a video game. It's, like, insane how much like a video game it feels.
Speaker 2:Yeah. Like, the the the arc of it.
Speaker 1:The Yeah. Like, you, like, research one technology and unlocks another one. And then, like, eventually, you're, like, working on high level things and
Speaker 2:Yeah. It that is in games. Yeah. Sorry. I I know these things.
Speaker 2:I said Pareto principle is a joke and you didn't laugh, and I want everyone to know I didn't think that was actually the Pareto principle. I don't know what the Pareto principle is. I don't remember. But when I said it, it was just a joke. Okay?
Speaker 2:I just had to
Speaker 1:say it. It joke. One of the few principles Adam's Adam remembers.
Speaker 2:I just remember that word for some reason. Yeah. It reminds me of Geppetto. I I think of Pinocchio.
Speaker 1:Well, Pareto was an Italian economist.
Speaker 2:Oh, is Geppetto? Okay.
Speaker 1:Maybe Schippetto is Italian. I know that's
Speaker 2:Same family, basically.
Speaker 1:Yeah. Two brothers. Oh, man. Yeah. So anyway, this has a huge impact on React server components too.
Speaker 1:No, doesn't. Wait. You
Speaker 2:you caught me a little distracted and I'm like, oh yeah. RSCs. Uh-huh. So bringing this back to the
Speaker 1:beginning where you're like, all I've been
Speaker 2:doing is web frameworks. So do are
Speaker 1:you even motivated to work on code splitting now?
Speaker 2:Oh, no. I super am. But, like, it does put everything in a new perspective. Now I gotta read the stories, and I gotta know I wanna see the paper. I don't understand it.
Speaker 2:I love when, like, a new big paper comes out and it's, like, monumental and to, like, skim it like you know what you're doing. Like, uh-oh, okay. Well, those those mass symbols look cool. Yeah. Sure.
Speaker 1:Yeah. Well, what's interesting is I think this is actually one of those things that is, I would say, most technical things like this are so outside the realm of understanding, like, it's just not even worth it. I think this what's amazing about it is it's so simple. It's like, it actually is very simple.
Speaker 2:It's understandable. Like, how how quickly do you think there's a team a two pizza team at Amazon doing something to try and use these materials in, like, the semiconductor lab or something? I feel like that's a thing. They're probably jumping on it so Their
Speaker 1:order of operations is first, put out a memo that says we are now a blank company. So now they're a superconductor company.
Speaker 2:Oh, okay.
Speaker 1:Then their CEO does an interview where they're like, we were actually ahead of everyone in this and it's it's
Speaker 2:I didn't know this was their thing.
Speaker 1:And then and then like then they like kinda try to build something and catch up. I'm I'm I'm just referring, I don't if you saw the AI comments that, the CEO of AWS was making.
Speaker 2:Oh, no. Didn't see it.
Speaker 1:Yeah. He basically gave an interview where he was just like, we're ahead in all this stuff and we're gonna do this all better and like Yeah. Like just kind of like didn't say stating it as though like they've always been ahead and they're like not even worried. I want it to be true
Speaker 2:so Interesting.
Speaker 1:I want it to be true so bad because I do want access to all the LLM stuff. I want it to be a lot cheaper. And from like a technical point of view, it's actually important that exists in AWS because I can't just be shipping like, a lot of companies like, the whole thing with all this AI stuff is you can hyper personalize stuff, like Yeah. But you can't just start sending hyper personal stuff to like the OpenAI API. If you're doing like anything medical, you can't just start shipping.
Speaker 1:Who knows what they so having stuff all in your own infrastructure is is important. And getting it cheaper is important. So I want it to be true but I'm like really skeptical that they have made they're like, the the image they're projecting externally anywhere lines up to where they are
Speaker 2:in They have been really quiet on, like, the LLM space. Like, they do a lot of AI stuff, obviously, or, like, a lot of machine learning stuff, I guess, with, like, SageMaker and all that family of stuff. But, like, the the LLM stuff, they're saying now that they're sort of like they're up there with all these people that are pushing the boundaries. They just have been quiet.
Speaker 1:He basically was saying that you can't count Microsoft. They outsource the work to OpenAI. So Azure is not relevant. They didn't like do it. They don't have any this expertise internally.
Speaker 1:We have the expertise The thing is, they can make one claim always for no matter what happens, which is they are operating at a bigger scale than anyone. And anyone, they have more customers than anyone so Yeah. Yes, they will obviously enter into any space at like a crazy advantageous position. But he he kind of went further than that being like he made it sound like, no, like this is like we've always been doing this.
Speaker 2:We've been doing this.
Speaker 1:Yeah. And like we were training our own LLMs. I would not be surprised if they didn't start doing that till after like the general public was aware of this stuff. I would be kind of surprised if they
Speaker 2:didn't Well, they're very customer obsessed. So like that makes sense that the customers come to them and they're like, I wanna do LLM stuff on my AWS account. Yeah.
Speaker 1:And they're like, we've always been doing LLM stuff.
Speaker 2:Yeah. It's coming. Well, have you seen the there I think we've talked about him before. There's an AWS hero who works at Hugging Face, and he publishes roughly an article every day. And he's got an article on how to train like, how to fine tune Lama two on SageMaker, which is really interesting to me.
Speaker 2:Like, I'd love to play with that and have, an OpenAI alternative. Maybe it's like 90% of OpenAI. I don't know what GPT four versus Lama two looks like. But Yeah. To have that on your own AWS account running on a serverless SageMaker, you know, inference endpoint, that's pretty cool.
Speaker 2:It's like a whole article walking through
Speaker 1:it. Yeah. It's a very important workflow to figure out us to actually make it into products because that's kind of the the limiter right now. That's why all the AI products that are getting funded are just like a UI wrapper around ChatGPT because you just Right. Think the other stuff is is too hard.
Speaker 2:But the scale of it, the economics of it do not make sense. Like Yeah. I can say as a company that has a lot of users, like, working for a company that has a lot of users, we can't just, like, outsource everything to OpenAI.
Speaker 1:Not much.
Speaker 2:Most of our users don't pay us. Yeah. And there's, like, ad revenue and stuff, but, like, that it would have to be, like, training your own LLMs on much more economic infrastructure. So that yeah. That's exciting to me.
Speaker 2:I I think AWS will get there. I think, like, I do look at them as the only cloud that matters.
Speaker 1:Mhmm.
Speaker 2:I mean, I guess, like, I know you think about Cloudflare and all these. But, like, in terms of the big three, like Azure, Google
Speaker 1:Yeah.
Speaker 2:I don't know. I just don't care
Speaker 1:about those things. They they will get there, and I want it to happen sooner than later because then it also gives us as a like, SST as a company an opportunity to, like Oh. Bring AI to people in a way they can actually, use it at scale and not just for, like, little toys. So that yeah. Yeah.
Speaker 1:I'm excited from that angle. As a user, I'm also excited. Like Boomi, we have this really great, feature that when a doctor goes to go see a patient, they get a nice summary of everything about the patient. So they don't have to, like, read through the charts. They know like, here's important things I need to know.
Speaker 1:And even we we like like Liz has come up with a prompt for OpenAI that generates a pretty good good thing. Yeah. And that's a perfect
Speaker 2:a perfect use case.
Speaker 1:Yeah. Because it doesn't have to be perfect because if it misses a detail, it's like not the end of the world, there's not like a crazy liability situation there. Like we can kind of send data to OpenAI as long as we don't say like the person's name or or who it is and they'll personally identify information but, again, like there's a there's a cost question. We like need to make sure that economically that all makes sense. I think it will for us because we're B2B, it's like everyone pays us and we can afford the but like, yeah, just like that complexity there, like I need it in AWS and I'd like to see that.
Speaker 2:It's coming. I'm excited. And now I'm excited that a new material entered the world and it's gonna change everything. And I don't care if it's still early, I'm buying in. I'm buying in on the hype.
Speaker 2:Well, there
Speaker 1:there's all those prediction market have you seen this prediction mark this console prediction market? Do you know what that is?
Speaker 2:Is this like futures? Like, is this like a banking thing?
Speaker 1:Yeah. Basically. It's basically like I think it came out of the crypto world. I don't know if it existed outside of that. Basically, someone can create a bet.
Speaker 1:It's just betting. It's a bet for anything and people can place their bet on either side and that creates lines on like what people think.
Speaker 2:Oh, so there's like a line on whether this is gonna plan out?
Speaker 1:Yeah, plan out. Exactly. So I think it's pretty low. I think it's like under 20% of people believe.
Speaker 2:That it will?
Speaker 1:Yeah. There's like only 20 only 20% of bets on will side.
Speaker 2:Yeah. So most people are betting on
Speaker 1:It's gonna lower and lower as like time
Speaker 2:So So that's everyone's mostly a pessimist. We're some of the few optimists out here. Is that what you're saying?
Speaker 1:Yeah. I mean, I just want it to be true so bad.
Speaker 2:I feel like those bettors are very efficient though. Feel like lines are very efficient and they're good.
Speaker 1:Those prediction marketing are just good things. Wisdom of the crowds or whatever. Yeah.
Speaker 2:Yeah. It's like if you have a jar of jelly beans. Have you ever heard that?
Speaker 1:No. What is that? No. Sure you've
Speaker 2:heard that. You've read 10 books. If you have, a jar of jelly beans and you ask like one person to guess how many jelly beans are in there, like, they're never gonna get anywhere close. But if you ask like a 100 people and you average all of their responses, pretty close.
Speaker 1:Oh, wow.
Speaker 2:It's interesting. Yeah. This is like there's science behind it. I always say that. I'm always like, there's a paper or something.
Speaker 2:I gotta just say these things lightly. Okay?
Speaker 1:Some of my best friends are scientists.
Speaker 2:Oh, really? No. Scientists?
Speaker 1:I'm just kidding. Or okay.
Speaker 2:I was like I
Speaker 1:was making I was like, that's what you're like. You're
Speaker 2:like, no,
Speaker 1:it it's true. Yeah.
Speaker 2:Yeah. I've got friends that
Speaker 1:are yeah.
Speaker 2:That's what I should say. Yeah.
Speaker 1:Okay. Alright.
Speaker 2:Well, I got a lot of work to do on Astro, AstroSight and AstroSEC. So I'm gonna hop off here while the world changing underneath me.
Speaker 1:Nice. That's work for that's work for me. You're doing work for me. Oh, yes. Sorry.
Speaker 1:Nice.
Speaker 2:I'm gonna give you some work.
Speaker 1:That's a good work. Adam's working for me today. It's called open source.
Speaker 2:Cool. This was fun. This was a
Speaker 1:good one.
Speaker 2:I don't know. I don't I don't always feel at the end of it like, that was a
Speaker 1:good one.
Speaker 2:This was a
Speaker 1:good one. Yeah. Exciting day. Like that? In general.
Speaker 2:Yeah. Exciting day. The world is changing. Okay.
Speaker 1:Cool. DataX.
