One last party

Interactive Transcript

spk_0 This is JS Party, your weekly celebration of JavaScript in the web.

spk_0 If you like this show, you will love the Change Log.

spk_0 It's news on Mondays, deep technical interviews on Wednesdays, and on Friday's an awesome

spk_0 talk show.

spk_0 That's a lot like JS Party.

spk_0 Now that I'm thinking about it, find us by searching for the Change Log wherever you

spk_0 listen to podcasts.

spk_0 Big thanks to our partners at Fly to IO.

spk_0 Launch your app in five minutes or less.

spk_0 All around the world.

spk_0 Learn how at fly.io.

spk_0 Okay, hey, it is part time, y'all.

spk_0 What's up friends?

spk_0 And party people, Adam here.

spk_0 I'm sitting with Danny Grant, co-founder and CEO of Jam.

spk_0 Dev, one of our new sponsors.

spk_0 Yes, Jam.dev is one click bug reports that devs love.

spk_0 It's just too easy.

spk_0 Get Jam for free today at jam.dev.

spk_0 So Danny, how do you explain what jam is and how it helps teams be more effective?

spk_0 Jam is the fastest way to capture a bug in a way that developers can debug it faster.

spk_0 It's a browser plugin that hooks into DevTools.

spk_0 So when anyone on your team spots a bug, they can one click capture what happened on the

spk_0 stream plus everything in DevTools.

spk_0 Console logs, network requests, session information.

spk_0 And it grabs it into a link so that when you open up the ticket, you have every single

spk_0 thing you need to debug.

spk_0 You don't have to ask a single follow-up question.

spk_0 This is for teams who want to spend their time building new features, not fixing old ones.

spk_0 I think that the most impact a software engineer can have is on building the future for their

spk_0 customers.

spk_0 It's on making things easier.

spk_0 It's building what's next.

spk_0 And so we want to make sure that you're not spending your whole afternoon just trying

spk_0 to repro a bug.

spk_0 Instead, you have everything you need to just get started.

spk_0 OK, friends, go to jam.dev and learn more about what jam is doing for teams to make

spk_0 bug reporting and all that fun stuff.

spk_0 Super easy, super fast.

spk_0 Get jam for free today.

spk_0 Jam.dev.

spk_0 Again, jam.dev.

spk_0 Oh, yes.

spk_0 You know the sound of those breakmasters, cylinder beats, means it's time.

spk_0 Once again, for J.S. Party, I'm Jared, your internet friend.

spk_0 And I'm joined by my friends, Nick Nesey.

spk_0 Now I have the time of my life.

spk_0 Oh, wait, wait.

spk_0 Please keep going.

spk_0 Please never stop.

spk_0 I've failed like this before.

spk_0 And K-Ball.

spk_0 And so we come to the end of the road.

spk_0 Hello.

spk_0 And Amy.

spk_0 I'm going to spare everybody.

spk_0 Hello.

spk_0 Thought you guys might hit me up with the triple song.

spk_0 All right.

spk_0 Fair enough.

spk_0 And we are here for one last J.S. Party.

spk_0 You've probably heard by now, since we delayed our finale quite a bit, that yes, J.S.

spk_0 Party is coming to an end, along with all other change log pods that aren't named the

spk_0 change log.

spk_0 But we've had a blast.

spk_0 We hope you've had a blast.

spk_0 And we are here today to have one last party.

spk_0 Traditionally, we do a New Year's party.

spk_0 In fact, for the last three or four, maybe five years, we've had our first episode of

spk_0 every year has been called New Year's party in which we had made some predictions.

spk_0 We had made some resolutions.

spk_0 And then we hold our feet to the fire the following year and see who was right, who was wrong

spk_0 and who failed miserably.

spk_0 Usually all of us are wrong and failing.

spk_0 But we thought we'd do that even though it's almost February, it might be February, but

spk_0 time we hit publish on this sucker.

spk_0 And so everyone's kind of over it with predictions and stuff.

spk_0 But you know what?

spk_0 It's our last show.

spk_0 We're going to do what we want to do around these parts.

spk_0 And one thing I really wanted to do, although Chris, bone, skull, hiller, is supposed to be

spk_0 here.

spk_0 He might hop in mid-show.

spk_0 We'll see what happens with the schedule.

spk_0 But he had this epic prediction, which was seemingly a surefire winner was that he predicted

spk_0 Jimmy Carter would die in 2024 because he'd been on hospice for a very long time.

spk_0 And dog, wow, I'm gonna say dog, but yeah, I guess so.

spk_0 I'm not happy the guy died.

spk_0 But he almost made it.

spk_0 He almost made it all the way to the end.

spk_0 It was like a day or two.

spk_0 What day did Jimmy Carter die?

spk_0 Was it like December 28th or something like this?

spk_0 He made it all the way to the end of the year.

spk_0 Outlasting most people who live, making it what 100 years.

spk_0 And Chris was almost wrong.

spk_0 Chris was almost wrong, but the last hour, President Jimmy Carter passed away.

spk_0 December 29th.

spk_0 So two days shy.

spk_0 Chris thought he was going to lose and he slipped something into the water.

spk_0 Yeah.

spk_0 He must have insider access there, making sure that it happened.

spk_0 But that was, I thought funny because we've, when you made the prediction, we're like,

spk_0 yeah, well, that's not a hard one because, you know, he's expected to die any day now.

spk_0 And but man, we continue to outlast all the predictions until the very end of the year.

spk_0 So I think he might be the only one that was right.

spk_0 Do you remember yours?

spk_0 I was just looking going back.

spk_0 I don't remember what it was this year.

spk_0 I have it here.

spk_0 Okay.

spk_0 You predicted that towards the second half of next year, now this was back a year ago.

spk_0 Now so you were talking about 24 that we're going to get back to something that feels a

spk_0 little bit more stable in the tech industry, specifically with regard to layoffs, overall

spk_0 vibes, etc.

spk_0 How did you rate yourself on that one?

spk_0 I would rate myself a loss on that.

spk_0 Yeah.

spk_0 I think it's taken a lot longer than I expected to get back to normal.

spk_0 I don't think it is back to normal.

spk_0 I think we are, there are a lot of things that sort of moved in a more positive direction.

spk_0 I don't think it's felt as negative as it did before, but less negative is not the

spk_0 same thing as a prediction of fact, a normal and positive vibes.

spk_0 I'm going to give myself the big L there.

spk_0 Fair enough.

spk_0 Nick, you had a prediction as well.

spk_0 Do you remember what it was?

spk_0 I sure don't.

spk_0 I could make up whatever I want right now.

spk_0 You predicted that TypeScript would lose market share to an other company called JavaScript.

spk_0 That's false.

spk_0 No, you actually predicted that next, JS, the facade, a next facade to the Versel advertisement

spk_0 that it is would finally shine through and it's going to really affect react market share.

spk_0 People start to realize that next was merely a Versel advertisement.

spk_0 Now I'm paraphrasing you and characterizing what you said, something like that.

spk_0 It would be so much so a drop off of usage because people would be upset about that, I guess,

spk_0 that they're going to stop using react.

spk_0 And I was right.

spk_0 OK, didn't we put a hard prediction on Astro taking over relative to next?

spk_0 Yeah, we're going to look at state of JS survey results.

spk_0 Maybe if you could find that quickly, we could do that.

spk_0 I didn't go that far.

spk_0 And we're going to see if Astro had stepped up its usage base over previous years versus

spk_0 next.

spk_0 I remember that, yes.

spk_0 And Amy, you weren't here for last year's New Year's party because you're a new party

spk_0 animal.

spk_0 You joined at the most inopportune time.

spk_0 So I think I'm the only winner here.

spk_0 Yeah, pretty much the only way to win is not to play.

spk_0 That's how I usually win as I just have everybody else make predictions and then I just

spk_0 don't do them myself.

spk_0 So I don't get made fun of.

spk_0 But yeah, you win by not participating.

spk_0 So congratulations to you.

spk_0 I also don't think I did participate.

spk_0 But at least I couldn't find anything in the chapters or quick scan of the transcript

spk_0 of things that I predicted.

spk_0 I remember the previous years predicting that the AI plateau would occur.

spk_0 And gosh, I just feel like sometimes I think that's right.

spk_0 And then other times I think it's not right.

spk_0 I did get the state of JS survey results.

spk_0 So one, Astro continued to rise.

spk_0 Okay.

spk_0 It is going up.

spk_0 It has in 2024 reached 23% response rate for usage, which puts it still at less than half

spk_0 of next JS, which did stop its meteoric rise and drop off by two percentage points.

spk_0 So maybe we could say that that Nick got, you know, at least a B plus on that of like,

spk_0 okay, next stopped rising, right, started falling, astro kept rising.

spk_0 But there's a long way to cross over.

spk_0 I wish it would have predicted that remix would be a thing in 2025 or 2024, because I don't

spk_0 even know if that's right.

spk_0 It's like they're taking a nap.

spk_0 Well, say what happened with remes?

spk_0 So remes got bought by shop of right.

spk_0 Yep.

spk_0 And they're using it as Shopify.

spk_0 But then is it going silent or is there no activity or just they decided to they quote

spk_0 take a nap.

spk_0 So they also do react router and they found that they were basically putting the same bit

spk_0 of code in both and react router has so much historic pull within the react community

spk_0 that they decided to go all in on react router and let remix quote take a nap.

spk_0 But I think the idea is that they're going to pour a lot more resources into react router

spk_0 and then turn remix more into like the framework pieces.

spk_0 Whatever that means, though, that's like framework adjacent stuff or whatever,

spk_0 but they're going to let react router shine a little bit more than it was with remix.

spk_0 I do feel like when we talked with Kent C. Dodds at the conference a couple of months ago

spk_0 that shipped on JS party, he was excited about some of the work they were doing with react server

spk_0 components.

spk_0 Maybe that was coming out under the auspices of react router rather than under remix.

spk_0 I'm not sure, but you know, kind of having a second opinion and not just having next

spk_0 be the only people doing anything with react server components.

spk_0 And so there's something going on in that space.

spk_0 Yeah, but this is the main problem because depending on who you ask, like you can't figure

spk_0 out what's actually happening because I do remember that conversation.

spk_0 And I think that like the RSC stuff is more specifically react 19, but maybe taking some

spk_0 turn with that within react router.

spk_0 But then like the remix stuff in react router is that just framework mode in there because

spk_0 they have library and framework mode now.

spk_0 And then I heard that remix would be like their RSC framework that they build now that

spk_0 react 19 is out.

spk_0 But who knows if that will ever ship or what it would be.

spk_0 So I think there's a there's a message that's lost somewhere in here.

spk_0 Well, I think the main message is that RSC and SSR and all that stuff in order to get

spk_0 that technology to work is hard because I would have put my bet on RSC and all that stuff

spk_0 taking off within this past year.

spk_0 But there's really no other framework right now that offers that as a production ready

spk_0 option.

spk_0 I think expose probably the one that's getting the closest.

spk_0 But that's probably also why next didn't lose as much market shares because that's really

spk_0 the only framework that you can use RSC in.

spk_0 I mean, it also took the react team a year to like release the final version of or not

spk_0 final version, but officially let 19 be what it's supposed to be.

spk_0 I mean, it's been in beta all year.

spk_0 To the extent that next released dependent on a non production version of it.

spk_0 Right.

spk_0 Is it all worth it?

spk_0 Is all this worth it?

spk_0 No.

spk_0 No, I don't do JavaScript anymore.

spk_0 It's not worth it.

spk_0 This is why we have to end the show because it's just not worth it anymore for y'all.

spk_0 We kid a little bit.

spk_0 But I will say RSC is pretty amazing.

spk_0 I've been so obviously I work on the Redwood JS core team and because I worked on the Redwood

spk_0 JS core team, I try and make sure that I understand all of the frameworks and everything that

spk_0 they're doing that way.

spk_0 I can better educate people on Redwood knowing, well, next does this remix.

spk_0 Does this or this is when you want to reach for reach for Astro.

spk_0 And so I've been building a project on next with their app router and the RSC stuff is really

spk_0 nice.

spk_0 I'm a huge fan of GraphQL.

spk_0 That's really what Redwood stuff is built on, but you have to do a lot.

spk_0 You have to write mutations and queries and create a use Apollo hook to write a function

spk_0 to send it off.

spk_0 And the fact that you can just write your server query right there within your component

spk_0 is really nice.

spk_0 It's a lot less code.

spk_0 So I understand the draw that people have for RSCs and saying this is the future.

spk_0 Have you found yourself ever getting confused about what's going on there?

spk_0 I don't find myself getting confused as much as I do just walking in these mental circles.

spk_0 So I'll start a new project and it'll be like, oh, be really nice to write this in Redwood

spk_0 because then I get GraphQL and Storybook and testing all this stuff built in.

spk_0 But then I can move faster if I have next with RSC and all these other pieces.

spk_0 But then the fact that Astro ships with less JavaScript and will compile everything

spk_0 down, that's really nice.

spk_0 The fact that has marked down support built in.

spk_0 But then like this server loader action pattern that remix has is really smooth and it's

spk_0 very clear what's on the server, what's on the client.

spk_0 So then I end up walking in circles, which is not helpful to anyone or my projects.

spk_0 It's like sometimes you just have to pick a technology and stick with it.

spk_0 And it's like everybody has their pros and cons.

spk_0 You just get comfortable with where those edges are and then you just push through the hard

spk_0 parts instead of constantly just walking in circles.

spk_0 I feel like you've just described quite eloquently why so many developers never ship any blog

spk_0 posts, but they ship their blog 10 times.

spk_0 Hold out.

spk_0 Ah, how many different versions of your blog has there been Nick?

spk_0 Oh, at least six.

spk_0 Yeah.

spk_0 And how many have used PHP to write the comments directly into the best one?

spk_0 Oh, dear.

spk_0 You know, there's a there's a lore or a myth in golf.

spk_0 I think it came from some pithy book that somebody wrote that the very first swing that

spk_0 you take is like your perfect swing and you're the rest of your life.

spk_0 You're like trying to find your way back.

spk_0 I know that gets tried all because I at well, I did have a good swing one time.

spk_0 Anyways, I don't know if it's true at all.

spk_0 It's probably not because you don't have a golf yet.

spk_0 How could you swing good?

spk_0 It's stupid, but I'm starting to think maybe that's true with Nick Nese's blog.

spk_0 You know, like the first one you wrote that very first one where the PHP re re-wrote itself

spk_0 and the comments and everything like that might have been the apex of Nick Nese's blog.

spk_0 Might have been.

spk_0 I went to the way back machine to see if they had any remnants of it and not that version.

spk_0 That's a it's a crying shame, Nick.

spk_0 It's a crying shame.

spk_0 Isn't this what we're all reaching for now and we're talking about building AIs that can build new AIs?

spk_0 We're all trying to get the PHP that re-writes its own file.

spk_0 We are.

spk_0 Jared, I want to ask you about a prediction.

spk_0 Okay.

spk_0 I know you didn't have, you said you at least don't remember one and I said,

spk_0 but I feel like you have been having a bit of a renaissance lately as I've been listening to you

spk_0 on other programs.

spk_0 Thank you.

spk_0 Yeah.

spk_0 Say more.

spk_0 You're more into VIM, right?

spk_0 Neavam?

spk_0 True.

spk_0 You're taking a look at it.

spk_0 At least you've switched off of terminal.app, which is like.

spk_0 This is true.

spk_0 Yeah.

spk_0 I mean, evolving.

spk_0 You know, this is the evolution of a man.

spk_0 Keep going.

spk_0 I want to hear about that because I'm still there.

spk_0 I got shamed into it.

spk_0 Okay, well, I got shamed out of it.

spk_0 Shamed in, shamed out.

spk_0 Do you have more to say to Nick?

spk_0 You want me to tell him about terminal.app.

spk_0 I do.

spk_0 And I'm like the only piece left of that trifecta would be to move to TypeScript.

spk_0 But is TypeScript moving to you?

spk_0 Oh, that's, if we were continuing this show, that would be the prediction I would make is

spk_0 that Jared will be writing full TypeScript without knowing it by the end of the year.

spk_0 Without knowing it.

spk_0 That might be the case because, you know, I have found, as I used it every once in a

spk_0 while, it's telling me stuff that I wouldn't know otherwise.

spk_0 You know, and I'm like, is this, is your TypeScript, you know, peak in time?

spk_0 I'm going to turn out under the covers there because the way I write my code, you shouldn't

spk_0 know this about my code.

spk_0 It doesn't have any sort of information like this available to you.

spk_0 I think that there's been two big, big things that just came out like towards the end of

spk_0 the year that is making that prediction for you.

spk_0 And the first is that node, well, in 23, so in 20 coming soon to the stable version,

spk_0 will let you just run TypeScript files.

spk_0 Right.

spk_0 It won't do any type checking.

spk_0 It'll just erase the types and then run the code, which is awesome because it

spk_0 removes one step.

spk_0 But then almost think, can I code I want?

spk_0 Yeah.

spk_0 Erase the types and then run the code.

spk_0 I mean, you're close.

spk_0 You're very close.

spk_0 Close.

spk_0 The other piece is like, I think it was just last week that the TypeScript team announced

spk_0 a new flag, Erasable syntax only.

spk_0 So it will disallow you from using things that have a runtime component like enums and

spk_0 enums come to mind.

spk_0 I'm thinking of, there's another piece that I can't think of right now.

spk_0 But that is all, I think, in service of the typed, what is the name of it?

spk_0 Type says comments.

spk_0 Type annotations.

spk_0 Yeah.

spk_0 The TC39 proposal of, right.

spk_0 I think it's like the balls are rolling.

spk_0 We're not going to be there at the end of 2025, but I think we'll be closer.

spk_0 What are we going to get first?

spk_0 That or temporal?

spk_0 That for sure.

spk_0 Temporal feels like it falls in the same bucket as like fusion power, right?

spk_0 It's always just a certain amount of time in the future.

spk_0 Yeah.

spk_0 And full self-driving autonomous vehicles are always 18 months away.

spk_0 Hey, that's the thing when Mr Zuckerberg said that they're going to have mid-level engineers

spk_0 that are all AI this year.

spk_0 That is his equivalent to full self-driving.

spk_0 That's my prediction.

spk_0 Meaning he's wrong.

spk_0 Meaning, yeah, it's always three months away.

spk_0 Right.

spk_0 Like a full self-driving.

spk_0 I mean, if they want a mid-engineer, I'm for hire.

spk_0 I'm available.

spk_0 They lost a pretty good engineer, didn't they, Nick?

spk_0 I haven't seen him.

spk_0 Yeah, OK.

spk_0 Yeah, Zuckerberg hasn't seen him recently either.

spk_0 Kable, you were going to say something.

spk_0 Well, few different things that I think might be interesting.

spk_0 One is kind of in the spirit of looking back.

spk_0 And since this is the last JS party episode ever, I was looking back to the first episodes

spk_0 that at least Nick and I were involved in, which were back in, what was this?

spk_0 March or April of 2027.

spk_0 2027.

spk_0 No.

spk_0 No, 2028.

spk_0 Sorry.

spk_0 April 28, 2018.

spk_0 There you go.

spk_0 I can talk, honestly.

spk_0 I can tell you.

spk_0 That's like this is some sort of future prediction.

spk_0 Back to the future.

spk_0 Which is actually funny, because if you think about back to the future, if they were going back

spk_0 in time, how much time did they go back?

spk_0 30 years, right?

spk_0 First time in the little 50s.

spk_0 Yeah.

spk_0 So if they go back and if that were happening today, they'd be going back to 1995.

spk_0 No.

spk_0 That was a glorious year.

spk_0 Let's do it.

spk_0 Why did you ruin my day?

spk_0 Windows 95 came out.

spk_0 It was amazing.

spk_0 Remember the friends advertisement for Windows 95?

spk_0 Really beaten down Brendan Eich's door.

spk_0 Hey, don't make that language down.

spk_0 Take more than 10 days.

spk_0 It's worth it.

spk_0 Yeah.

spk_0 I mean, there's a bunch of stuff we could do around how we're closer to different times

spk_0 than we once were.

spk_0 But yeah, like college students today, graduating college students, none of them were born

spk_0 when 9-11 happened.

spk_0 Yes.

spk_0 Right.

spk_0 There's a whole bunch of that stuff.

spk_0 But we're talking to developers.

spk_0 They're all young anyways.

spk_0 Oh, yeah.

spk_0 I mean, we're old codgers.

spk_0 We're old codgers.

spk_0 But this is kind of what I want to do.

spk_0 So looking back at those first episodes of JS Party that Nick and I were on back a mammoth

spk_0 is 20, 28 teams.

spk_0 So that's seven years ago.

spk_0 What has changed in that time?

spk_0 What were we talking about at the time?

spk_0 Any guesses?

spk_0 We were talking about React and TypeScript.

spk_0 That's a lot I was talking about.

spk_0 I was talking about.

spk_0 That just summarizes Nick's life for the last decade.

spk_0 Had React come out yet?

spk_0 React was out, right?

spk_0 I was.

spk_0 Okay.

spk_0 I was like, I was building WordPress blogs in 2060.

spk_0 React was out.

spk_0 But, oh, and actually, next was out as well.

spk_0 The first next version was in 2016.

spk_0 Looks like.

spk_0 So we had those things have stayed the same.

spk_0 One thing that hasn't, we were talking about Bootstrap.

spk_0 And wondering if CSS Grid was going to change how everyone was using Bootstrap.

spk_0 So CSS Grid did change things, but not a bootstrap kind of went on away on its own path.

spk_0 The tailwind revolution.

spk_0 When did that happen?

spk_0 That's more recent, right?

spk_0 It just hit version four.

spk_0 So I'm going to say it's probably like four years old, five years old, but it didn't

spk_0 pick up steam until maybe two years ago.

spk_0 I'm just making stuff up now.

spk_0 I think that's a big one.

spk_0 I think we were already talking about temporal.

spk_0 So as you predict when that's going to land.

spk_0 And I was also noticing we were talking about USB access, Web USB, and some of these

spk_0 other like taking JavaScript out from being just web related.

spk_0 That's a trend that I think has continued to flourish of JavaScript everywhere.

spk_0 Yeah, talking about Johnny five or Dweenows.

spk_0 I remember PWA's was a thing back then.

spk_0 Progressive web apps like what are progressive web apps?

spk_0 That's kind of died, right?

spk_0 Yeah.

spk_0 To a certain extent.

spk_0 Web versus mobile, right?

spk_0 Like the iPhone app store versus web apps.

spk_0 And can we can the web do more as a platform for mobile?

spk_0 When did React Native get big?

spk_0 Because that seems to be continuing down.

spk_0 And you have Expo is really making waves.

spk_0 You have like, what was it?

spk_0 Blue Sky was the first Expo app that was top of the App Store for a while.

spk_0 There's a bunch of other Expo apps happening.

spk_0 So React Native's initial launch was 2015, but you asked when it got big.

spk_0 And I remember making a splash right away.

spk_0 Obviously I had Facebook behind it.

spk_0 So and solved the problem that a lot of people had.

spk_0 And React was already getting steam.

spk_0 So I'm not sure if it came out of the gate steaming or if it built steam over time,

spk_0 but 2015.

spk_0 So definitely was out and available back when we rebooted the pot.

spk_0 But Amy was just writing PHP.

spk_0 I mean, did you with your page?

spk_0 I was writing J query.

spk_0 OK.

spk_0 Hey, I wrote some J query the other day.

spk_0 So I was just thinking when was that Google like language for mobile that was supposed to

spk_0 make it easier for new sites to load?

spk_0 No, it wasn't flutter.

spk_0 But AMP.

spk_0 Yes.

spk_0 That was terrible.

spk_0 Yeah, when did that start?

spk_0 That's because they would like re-host your pages on their own.

spk_0 Yes.

spk_0 And I was thinking about 2016.

spk_0 I was also working on a project where we had to like, recode the entire site and AMP.

spk_0 And it was miserable.

spk_0 I remember seeing that mean like, no, thanks.

spk_0 I did not want them to re-host my site.

spk_0 Thank you very much.

spk_0 But they had a big care behind it because it was going to get you better search results.

spk_0 And that's just like pure money, right?

spk_0 Yep.

spk_0 And load everything faster.

spk_0 Yep.

spk_0 The moral of the story was they eventually got rid of that version of the site.

spk_0 So that died along with the language.

spk_0 It's crazy how a company of that stature and resources and skills and everything could

spk_0 still create a turn like that.

spk_0 You know, like not saying like the technology was bad or even maybe the idea wasn't bad,

spk_0 but like they still couldn't get the market to actually accept it.

spk_0 You know, like it's just.

spk_0 One share part of it was the developer experience, right?

spk_0 Which I think also says something about the code and the projects that we work on.

spk_0 So it has to feel good to work on and work in.

spk_0 I think that's a good point.

spk_0 I think back when we first rebooted the show, developer experience was not a common

spk_0 term in our industry.

spk_0 And now it's like there's developer experience conferences.

spk_0 There's companies called DX.

spk_0 I just won at least.

spk_0 And there's a heavy focus on that as like a core aspect of, you know, developer facing

spk_0 products and services, right?

spk_0 What's up, nerds?

spk_0 I'm here to curb Mackie, co founder and CEO of Fly.

spk_0 You know, we love fly.

spk_0 So Kurt, I want to talk to you about the magic of the cloud.

spk_0 Yeah, thoughts on this, right?

spk_0 Right.

spk_0 I think it's valuable to understand the magic mind of cloud because you can build better

spk_0 features for users.

spk_0 Basically, if you'd understand that, you can do a lot of stuff, particularly now that

spk_0 people are doing LLM stuff, but you can do a lot of stuff if you get that and can be

spk_0 creative with it.

spk_0 So when you say clouds aren't magic because you're building a public cloud for developers

spk_0 and you go on to explain exactly how it works, what does that mean to you?

spk_0 And some ways it means these all came from somewhere like there was a simpler time before

spk_0 clouds where we get a server at Rackshack and we just say, or tell net into it even and

spk_0 put files somewhere and run the web servers ourselves to serve them up to users.

spk_0 Clouds are not magic on top of that.

spk_0 There's just more complicated ways of doing those same things in a way that meets the needs

spk_0 of a lot of people instead of just one.

spk_0 One of the things I think that people miss out on, and a lot of this is actually because

spk_0 AWS and GCP have created such big black box abstractions, like LAMDA is really black boxy.

spk_0 You can't like pick apart LAMDA and see how it works from the outside.

spk_0 You have to sort of just use what's there.

spk_0 But the reality is like LAMDA is not all that complicated.

spk_0 It's just a modern way to launch little VMs and serve some requests from them and let

spk_0 them like kind of pause and resume and free up like physical compute time.

spk_0 The interesting thing about understanding how clouds work is it lets you build kind

spk_0 of features for your users you'd ever would expect it.

spk_0 An archononical version of this for us is that like when we looked at how we wanted

spk_0 to isolate user code, we decided to just expose this machine's concept, which is a much

spk_0 lower level abstraction of LAMDA that you could use to build LAMDA on top of.

spk_0 And what machines are is just these VMs that are designed to start really fast or designed

spk_0 to stop and then restart really fast or designed to suspend sort of like your laptop does

spk_0 when it closes and resume really fast when you tell them to.

spk_0 And what we found is that giving people those parameters actually there's like new apps

spk_0 being built that couldn't be built before specifically because we went solo level and

spk_0 made such a minimal abstraction on top of generally like Linux kernel features.

spk_0 A lot of our platform is actually just exposing a nice UX around Linux kernel features,

spk_0 which I think is kind of interesting.

spk_0 But like you still need to understand what they're doing to get the most use out of them.

spk_0 Very cool. Okay.

spk_0 So experience the magic of fly and get told the secrets of fly because that's what they

spk_0 want you to do.

spk_0 They want to share all the secrets behind the magic of the fly cloud, the cloud for productive

spk_0 developers, the cloud for developers who ship learn more and get started for free at fly.io

spk_0 again fly.io.

spk_0 My new official title is developer experience engineer.

spk_0 Oh wow.

spk_0 What are you doing?

spk_0 JavaScript.

spk_0 Get her issues.

spk_0 Yeah, exactly.

spk_0 I think it's funny that you think that they couldn't produce a turd like that company

spk_0 that big.

spk_0 Let me introduce you to Gemini.

spk_0 Gauntlet thrown.

spk_0 I haven't tried Gemini.

spk_0 I've largely stayed away from Google's AI stuff ever since they had like the black George

spk_0 Washington thing.

spk_0 I was just like these guys aren't ready for prime time here.

spk_0 If you remember the the one name of it was Bard.

spk_0 Oh, I do remember.

spk_0 Yeah.

spk_0 Well, I was thinking back to this too.

spk_0 And I did the 2025 thing when you're reminiscing as I asked Claude what was big in 2018.

spk_0 And it says that HTTPS essentially became essential.

spk_0 Let's encrypt.

spk_0 Oh, yeah.

spk_0 Let's encrypt was world changing.

spk_0 Yeah.

spk_0 While you while we talk Nick asked Claude what happened in 2028 in the world of JavaScript

spk_0 and web development.

spk_0 I mean, part of why this is interesting is looking back like that seems like a whole different

spk_0 world.

spk_0 So if we're going to like project forward, where do we see this going?

spk_0 I predict if we look forward seven years, 99% of software code is generated by LLM is

spk_0 rather than written by humans.

spk_0 It's funny that you say 99% because I just covered a story in news where a

spk_0 pull request was opened against LLM.cpp, which is a low level C and C++ program.

spk_0 That was doubling performance of some SIMD instructions lower level than I even understood

spk_0 that the code was obviously the people working on understood.

spk_0 And they opened up a very large pull request that doubled the performance of this particular

spk_0 subsystem.

spk_0 And it was 99% written by deepseek.

spk_0 And the opener of the request that all I had to do was prompt it, provide a couple tests,

spk_0 and just guide it along the way.

spk_0 And here's your code.

spk_0 And I was like, that's pretty impressive.

spk_0 And that's 2025 style.

spk_0 So you're thinking 2032 seven years down the road.

spk_0 Yeah, I could think I could be the case.

spk_0 What Claude say Nick?

spk_0 It's only given me propaganda about the 2028 Olympics.

spk_0 I thought you're going to say the election.

spk_0 I was like, well, either way, it'd be propaganda.

spk_0 Yeah, that's funny.

spk_0 I don't know what's going to happen, but probably this country is going to win in the downhill.

spk_0 It does say Gartner predicts significant advancement in generative AI with 40% of major

spk_0 companies expected to implement digital twin profiles for employee skills and behaviors.

spk_0 Digital twin profiles.

spk_0 That's like straight up.

spk_0 That sounds like a bad idea.

spk_0 I mean, when is that overstopped us?

spk_0 I'm a huge advocate of these tools for a lot of use cases,

spk_0 but I will say that most of the ideas being put out there by leaders and

spk_0 gen AI of what they think the future should look like with these tools are terrible ideas.

spk_0 Like these are really useful things.

spk_0 We should be figuring out how to use them to make our lives better.

spk_0 I'm not sure that like Sam Altman and some of these other folks who are putting

spk_0 that out realize just how dystopian they sound when they are projecting what they think that world

spk_0 should look like.

spk_0 It's getting harder and harder to get attention on the internet.

spk_0 You know, so sometimes you have to say the most outlander's most dystopian or utopian in your mind

spk_0 stuff in order to get a headline or a viral post or something.

spk_0 So I take a lot of that with a grain of salt.

spk_0 I don't think they even believe half the crap they're putting out.

spk_0 You know, though, I'm thinking back to our time together in New York, Gable,

spk_0 I just kept getting bashed over the head mostly by you, but by pretty much everyone else,

spk_0 cursor, cursor, cursor, I still haven't touched it.

spk_0 But and you're still gamefully employed.

spk_0 Some what?

spk_0 Yeah.

spk_0 But I, my, it did like enlighten me to at least look for this workflow.

spk_0 My main problem with cursor is I don't want to use VS code.

spk_0 Same.

spk_0 Or therefore.

spk_0 Also my main problem with it.

spk_0 And yet I still ended up going over to it because I could see when I started getting used to

spk_0 using it such improvements in my productivity.

spk_0 Yeah.

spk_0 And I'm getting the same improvements now, but in the comfort of me of him and it's working

spk_0 quite well.

spk_0 I'm using a plugin called Avanti and I just had to give it an anthropic key and put $25 on

spk_0 that and it's, it's going to town.

spk_0 But it, uh, it has changed the way that I write code, which is I just talk about my code

spk_0 a lot with a robot rather than writing it.

spk_0 Like probably 60% of the time is just talking about it.

spk_0 And then I write it, but it's like, I have somebody always there always listening.

spk_0 And a lot of times, yeah, it's complete crap that it's giving me back.

spk_0 So like that's where the, the skill set comes in of like distinguishing pretty quickly.

spk_0 Like, I don't think you know what you're talking about here and going the right way.

spk_0 But at the end of the day, it is so much better than staring at a blank terminal or blank

spk_0 file editor and starting from scratch on something, especially when it's like,

spk_0 I just know I need to do, you know, this quick thing, like a reduce.

spk_0 And I could, I know how to do it.

spk_0 I'm just being lazy.

spk_0 So I can just have the computer do it.

spk_0 And it's, it is changing the way that I write code and the way that I interact with everything.

spk_0 Is it good?

spk_0 I don't know.

spk_0 Right.

spk_0 It's at what Linnler you're using this though.

spk_0 Like if you were to characterize a reduced call versus write me a method versus create

spk_0 a series of classes that interact or a module system or like what level of your code

spk_0 abstraction?

spk_0 Are you generally operating that?

spk_0 Hmm.

spk_0 I think I'm at the top and the bottom, the high level and the low level.

spk_0 So meaning a lot of those reduced calls like the simple stuff, you know, I have that.

spk_0 I also have copilot that's still in there just trying its darned it's just doing stuff.

spk_0 Like, it's like this.

spk_0 Like this.

spk_0 It's pretty awful.

spk_0 But the, and then at a high level, like a lot of times I will describe like, Hey, I'm thinking

spk_0 about making this change to the code.

spk_0 And sometimes I'll just give it the code.

spk_0 Like in Avanti, I can just the file that I'm at, I can just say at file or at, you know,

spk_0 I can give it context on all of the files that I'm asking about.

spk_0 Right.

spk_0 And then ask, like, I think that this would be better if I provided this kind of call back here.

spk_0 What are the pros and cons of doing that?

spk_0 And I have like a high level architectural architecture discussion on that.

spk_0 But in the middle area where I'm not running those low level things and I'm not doing the

spk_0 architecture by myself, it's kind of like the, the general implementation is where I'm

spk_0 still like fully in charge and doing things without any AI.

spk_0 Amy, you're not in a long as are you having some more experiences or does it resonate

spk_0 with you or.

spk_0 Yeah.

spk_0 So a few weeks ago, Kable, you mentioned cursor and like that at command changed everything

spk_0 for me and sent me on a deep dive with cursor.

spk_0 Like, what all can it do?

spk_0 And it's pretty amazing.

spk_0 And some of it, I think still comes down to your context window and the information that

spk_0 you provide it.

spk_0 So that's interesting that you're talking about this high level and low level because part

spk_0 of that journey, like I realized there's such a thing as a dot cursor rules file.

spk_0 Amazing.

spk_0 So what this file is is you can put it in any project you're working on and cursor the

spk_0 website has a directory of all these different files that people have created.

spk_0 And so in it, you would, it says like, I'm on a next 15 app router project using tailwinds

spk_0 that uses ShadCN and RedX and all this other stuff.

spk_0 And so then when it makes suggestions, it's going to give you suggestions that are

spk_0 aligned with that framework.

spk_0 And you can also outline, this is my folder structure.

spk_0 So it knows this is where I want my components to live.

spk_0 This is where I want my pages to live.

spk_0 And so the feedback that it's giving you is a lot better.

spk_0 So when you're talking about a high level perspective, like that's where all of that content

spk_0 lives.

spk_0 So it's pretty amazing.

spk_0 You can also, which I didn't realize, set a file inside your cursor settings to say,

spk_0 this is generally how I like to code.

spk_0 Or this is the type of response that I want.

spk_0 And so I found a tweet that somebody on the cursor team had provided and said, this is

spk_0 what I use in my cursor settings file.

spk_0 So I copied and pasted that tweaked it a little bit.

spk_0 But the main difference was it was asking for more direct and terse explanations.

spk_0 And so when you're talking about just trying to get to the code as fast as possible, that

spk_0 really helped cut down on its feedback of, you know, cursor.

spk_0 Wow, that's a great question.

spk_0 Like it'll cut on that and just give you your answer.

spk_0 The other thing is the documentation.

spk_0 If you, if there's not the documentation for whatever technology you're using in their

spk_0 repo, you can say, here is a link to the docs and I'll go to that website and scrape

spk_0 all the documentation, which is also really interesting.

spk_0 So I have used it on a few projects to write all of the code if it's a side project.

spk_0 So for example, a couple of weekends ago, Raycast released, Raycast focus.

spk_0 And so I thought, oh, this would be interesting if we had like a refocus app.

spk_0 So a lot of times if I'm heads down on something and I know I have to step away to say, walk

spk_0 the dogs, I don't want to do that because I know I'm going to get distracted.

spk_0 I'm not going to sit back down and work on what I was working on.

spk_0 I'll lose that focus.

spk_0 And so I was like, I just want to remind her in the middle of my computer to say, this

spk_0 is what you were working on.

spk_0 This is what you need to go back to.

spk_0 I've never written a Raycast extension before.

spk_0 And so I did all that with AI in about an hour and it ended up writing an Apple script,

spk_0 which I'm not very proficient in Apple script.

spk_0 But the fact that it could just do all that by itself was pretty incredible.

spk_0 One of the things that we've experimented with that is very interesting in this domain

spk_0 is like one of the places where LLM based coding agents are really good is like translate

spk_0 this thing into this language, right?

spk_0 So translate across languages.

spk_0 So we started experimenting with we have a contractor who's proficient in Python and

spk_0 our main systems are using golang and Ruby and TypeScript.

spk_0 And we said, no, you're doing an experimental thing right at Python.

spk_0 We get the deliverable and we say cursor translate this Python into Ruby using these guidelines

spk_0 and it'll just go and suddenly the language of development for the original development

spk_0 is far less important because you can translate it in 30 minutes.

spk_0 Yeah, that's awesome.

spk_0 But I still don't quite know.

spk_0 Sarah Dressner said this and this sums it up so well.

spk_0 But she just made the point like I like coding.

spk_0 I like the job that I do.

spk_0 I liked my craft and it's not that I want to be resistant towards the future or whatever

spk_0 comes next.

spk_0 I think that there'll still be a job.

spk_0 It'll just look different.

spk_0 The question is, I don't know that I like my different job.

spk_0 And I think that's still to be determined.

spk_0 Yeah, I think there's a lot of questions around what does it look like

spk_0 when you're doing this.

spk_0 I was talking with somebody about who's on the older side about what it was like when

spk_0 compiled languages first came out where they were having to shift from literally writing

spk_0 programs using assembler instruction by instruction to these higher level abstractions.

spk_0 It's a very different way of thinking about it.

spk_0 But it's still programming and you made that shift.

spk_0 I've found for me writing code with LLMS, I actually have to be careful the right amount

spk_0 that I ask it to do not only because if you give it too much without the right things

spk_0 that can go off the rails, but also because just my brain won't keep up.

spk_0 And it's both less fun.

spk_0 And then if I have to come back and change or modify this, I'm less able to do it.

spk_0 Like so I will ask it in a chat oriented way to do something.

spk_0 But I found that for me, it typically has to be one train, one conceptual transformation

spk_0 at a time.

spk_0 So my brain is still mutating the code.

spk_0 And the fact I'm not doing the keystrokes to change it ends up being less important.

spk_0 Then like, I'm able to do it at a size that is my brain is still doing the same types

spk_0 of functioning that it was doing before, if that makes sense.

spk_0 If I ask it to implement the whole thing, that's neither fun nor often the right solution

spk_0 unless it's a throwaway or one-time script or something like that where I'm not going

spk_0 to need to maintain it over time.

spk_0 Well, I think that's the point with that Ray Castics in too, it was very small.

spk_0 And so when you talk about the cursor being able to hold all of that in context, it was

spk_0 able to do that because it wasn't a large application.

spk_0 It was also a side project.

spk_0 So I didn't have to worry about branding.

spk_0 I didn't have certain specifications that a lot of times if we're working for companies

spk_0 or there's branding and specs and all those things involved, sometimes it's harder to

spk_0 get the exact response that you're looking for when you need it to fit those specifications.

spk_0 I think that maintenance aspects is key because so much of what we do is maintenance.

spk_0 And it's one thing to create something, but then to maintain it, there's another level

spk_0 of understanding or care.

spk_0 I don't know what it is, but I find that when you detach yourself from the nitty-gritty,

spk_0 which more and more so you can do, obviously we're still checking and making sure, no,

spk_0 don't write like this, write it like that.

spk_0 I assume those kind of changes eventually go away to where we're doing that way less with

spk_0 our LLMs and the code they spit out the first time is actually like dang, better than I could do.

spk_0 But then we don't grok it ever.

spk_0 Like we grok the outputs or the test suite passes.

spk_0 And there's this interesting study that came out recently on metacognitive laziness.

spk_0 And this is the danger of using Gen.J.V.I with learning specifically because you actually

spk_0 tend not to learn very well from them, even though they have all the information.

spk_0 It's just almost too easy.

spk_0 It's like they just tell you the answer and you're like cool and then you don't learn it because

spk_0 there was no actual going and getting it.

spk_0 And I just wonder over time, I think the answer to that problem is like, well,

spk_0 they have to get better so that we don't have to actually do that.

spk_0 We just can stay at a higher level, but I don't think that that's, I think that is where

spk_0 you have plateaued at least for now.

spk_0 I disagree pretty vehemently with the idea that they have to get better so we don't have to.

spk_0 I think we have to figure out how to learn these things in a way that we're

spk_0 or use these things in a way that we're still learning.

spk_0 On the coding side, like the biggest thing I've seen, the difference,

spk_0 biggest difference I've seen between people who are able to use these things really like get

spk_0 orders of magnitude productivity boost versus like incremental is you don't turn your brain off.

spk_0 You don't just let it do all the things.

spk_0 You figure out how to like use it to more rapidly inform your brain.

spk_0 But I think because it's so easy to to to turn your brain off, we don't have that instinct.

spk_0 So like I think maybe that's a UI tool builder problem, like maybe applying these directly in

spk_0 the tools that we've previously done leads to this. Maybe I don't I think if we leave it up to

spk_0 the end user, we're probably doomed because most of us are lazy and we will be lazy by default.

spk_0 But I don't think the answer is for the tools to get better enough. We don't have to care.

spk_0 I think that we need to figure out how we can use these tools in a way that helps us think

spk_0 better rather than turning our brains off. You know how many developers literally copy paste

spk_0 their answers from Stack Overflow and just put it in their program. And then they're like,

spk_0 how does this work? Like I don't know what it worked. So I moved on. I mean, that's going to be the

spk_0 next thing. Only obviously Stack Overflow is dying a quick death. But it's going to be the same.

spk_0 There's so many people like anybody who all four of us were not or three of us and you were

spk_0 not in a long and agreement as you said that, K-Ball. That's like the 1% of developers that actually

spk_0 are going to mean because everyone else is like, yeah, I'm busy. I'm going to go to happy hour.

spk_0 So I'm just going to paste it in there. It worked. Ship it. Let's move on. I'm not going to learn it.

spk_0 Why would I have to learn it? It works like 98% of the time. There's always been that stigma around,

spk_0 like I don't want people to know that I pasted from Stack Overflow. So I will manually type this in.

spk_0 It's called shape. Yeah. Oh, manually type this in. I'll change a few characters just in case.

spk_0 We'll have that comment there. No. It's like that. But it's like, I mean, part of that was like

spk_0 an exercise. And like, if I'm typing this out, it is forcing me to go line by line and understand

spk_0 and kind of at least give me the opportunity to critique as I'm going along by line. It just

spk_0 slows me down enough to not do that. But there's always like that stigma of like, oh, I don't want people

spk_0 to know that I had to stack overflow to get this. Do you have, do you feel like there's that same stigma

spk_0 for AI tools? Like, like, hey, well, if you just say ask cursor to rewrite this, you know, whatever.

spk_0 And does it feel like you wrote it? Does it matter? It's a good question. I think at the end of the day,

spk_0 what matters is does it get the outcome that you're looking for? Where I think there's always been

spk_0 a difference between like the developers who are just copying, pasting versus the other is in the same

spk_0 domain that Jared, you're pointing at, which is maintenance is being able to evolve a system over time.

spk_0 Right. The best developers have a more or less complete and evolving mental model of what their

spk_0 software does, how it relates to the business problem that's solving the mental model of that

spk_0 problem itself and like the ways in which those things interact. And what that enables you to do

spk_0 is see possibilities. It enables you to see how you can evolve the system to address different

spk_0 problems, so just those problems better. If you don't understand any of that, you don't have any of

spk_0 that visibility. So that differential is already there. This probably exacerbates it because I have

spk_0 seen, you know, you look at like somebody who's doing this very visibly in public, I think is Simon

spk_0 Willisson. He is supporting like 300 different open source projects and doing all these other things.

spk_0 He's like 100x his productivity by using LLMs. He's doing it really well. And I think he's not

spk_0 unique. I've seen other developers who are able to take these tools and use them to do ungodly

spk_0 amounts of work. So this concept of a 10x developer that for so long has been like an argument thing.

spk_0 Like I think when you talk about how do you effectively use LLM based tools, you can get to 100x.

spk_0 And that's going to have an impact. That's going to mean that if you're in that copy and place

spk_0 and just accept and don't think about it mode, like you are going to be replaced by someone else

spk_0 who's able to use these things more effectively and keep evolving the system. And we're seeing

spk_0 company after company at the big scale saying we're not hiring engineers now because we're seeing

spk_0 such productivity boons from this. Like I think that's going to that is a challenge that we have

spk_0 to face. Now the opportunity here is now we can be in a world of abundance when it comes to software.

spk_0 Right. You can run a very effective software company on two, three, four engineers and do a heck

spk_0 of a lot if those engineers are really embracing these tools and understanding how to do all this

spk_0 amazing stuff. And that opens a ton of places and opportunities where the economics wouldn't have

spk_0 worked before because software development was too expensive. But if you're just blindly applying

spk_0 the tools like I think this is a period where like there's going to be a shakeout. You hear that

spk_0 y'all don't let that don't let that instinct that copy base instinct that take it away curse or

spk_0 instinct don't let that take over you have to feel the shame and rise above or be replaced.

spk_0 Well, if it writes some code for you, that's fine. And you have to answer the question,

spk_0 do I understand this code? And if the answer is no, you're not done. And maybe the way that you

spk_0 get to understand it is actually using it all. And one of the places I think these things are

spk_0 really cool for me. And it's like I can hop into a code base. I haven't touched in six months.

spk_0 I don't remember how things are. I can open up cursor and say, wait, how is this thing working

spk_0 again? And it will go and it will find the right things and it'll give me a bunch of good information

spk_0 and a jumping off point. And then I can say, okay, great. I need this to change in this way.

spk_0 Find me where else it needs to like it's incredible in terms of hopping in and like not having

spk_0 that massive code context switching cost of this is a 12 month old or six month old code base.

spk_0 I haven't touched in that long. But you do have to ask yourself that question. Like, do you want

spk_0 to understand what this is doing? And if the answer is no, don't push it. I mean, we've been telling

spk_0 junior devs that for a long time, right? Like, have you test? Did you test your code? Do you know how it

spk_0 works? Why did you push this in a PR? Have you guys seen Google hunting? Yes. A long time ago.

spk_0 Kable have probably hasn't come on. Kable. Well, there's this particular thing. So will hunting.

spk_0 That's Matt Damon's character. Kable, he's a savant. He's an intellectually brilliant person.

spk_0 Can do math like a whiz kid. And he gets put to work on these these advanced math equations

spk_0 that nobody else could solve and these theorems and stuff. And he's working with a guy who's like

spk_0 a accomplished mathematician. He's got like awards and he's like the smartest guy in the room all

spk_0 the time. And he's doing stuff that for will is very, very easy. But for everybody else in the

spk_0 world, it's very hard. And he's like a very smart mathematician. And so for him, it's hard to

spk_0 understand for will. He's like, and this is a whole scene where he gets mad. Will takes the

spk_0 proof and he's like throws it in the fire. He's like, this is so easy for me. He have an idea

spk_0 how easy this is. And he's like, practically crying. He's like, because he wants the he wants to

spk_0 understand. And I say all that because like, isn't that going to be us with our tools pretty soon

spk_0 where it's like, okay, you wrote the code now. Could you please tell me what you're doing here?

spk_0 Because I'd like to understand how how you did that. And these things going to be like

spk_0 you simple tin. What how many times I have to explain myself to you? Just go away. Be gone. I

spk_0 will take care of everything from here. You have no idea what I'm talking about. I mean, the other

spk_0 thing that's possible that we could explore is maybe they're the programming languages we're

spk_0 writing today are not the right programming languages to be writing in this world. Because

spk_0 going back to that other metaphor, how many of us understand the assembly that's generated by

spk_0 the code that we write? Very few. But we still have code that expresses the higher level concepts

spk_0 and abstractions that we're trying to write that we understand those abstractions. And we may not

spk_0 understand the way that's mapped into an implementation. But the compilers are good enough we can rely on

spk_0 them. They pretty much always work. It's pretty rare that you have to dive through the levels of,

spk_0 oh, shoot, this compiler is actually generating bad assembly. So my abstraction isn't being rendered

spk_0 properly. So maybe the answer is actually we need to develop that next level of abstraction,

spk_0 which is how we express the conceptual abstractions we're implementing to an LLM. And then the LLM

spk_0 writes JavaScript or golang or whatever. And there's a whole aside on like what are the right good

spk_0 language targets for LLM? I think yeah. Actually, golang is a great target because it's relatively simple.

spk_0 It's strongly typed. It's fast to compile. It's fast to test. And so you can do a lot of validation

spk_0 of the generated code automatically. And it's fairly explicit. So it can't get too clever. Like,

spk_0 I'm using an LLM to write a lot of Ruby. And like when it tries to do metaprogramming, you're like,

spk_0 whoa, whoa, whoa, whoa, what the heck are you doing? Yeah, smoke a dooby to keep up, you know, all of this.

spk_0 That's interesting. You say that because I actually asked Chris Brando from the fall through podcast

spk_0 on Change, login friends, fall through is the go time spin-off pod for those listening much like

spk_0 this functional we should get to that is going to be the the JS party spin-off pod. So if you're

spk_0 listening this far and you're like, I can't believe this is the last JS party ever, you know,

spk_0 the single tier is rolling down your cheek. Have no fear. The spirit of JS party lives on

spk_0 in the dysfunctional podcast starring K-Bonank Nesey and Amy like the whole crew pretty much

spk_0 except for yours truly who would love to be a guest whenever you guys aren't talking about

spk_0 TypeScript. And anyways, so that's that's coming. This functional is coming. So I say a team for that.

spk_0 We'll talk about it more in detail. But I asked Chris the question about is go a good output language

spk_0 for an LLM? And I will tell you, he said no. So maybe go listen to that K-Bon and see why he thinks

spk_0 that as a go expert himself. Interesting conversation. I do think there's like some missing layer in here

spk_0 where it's like not English or your native tongue. And it's not TypeScript of course. But it's

spk_0 like some sort of pseudo code in between that we could talk, which is easier than we're currently

spk_0 doing. But more formal than just like me typing out, hey computer do this. That I think might be

spk_0 the sweet spot. What do you all think about that? Jared, are you proposing AppleScript? Amy seems

spk_0 to like it. It's like that would be the one time where I were talking about somebody just

spk_0 taking whatever code. I didn't even read the AppleScript that it created because there's just a

spk_0 side project. That's a beautiful thing. If you don't have to read it, it's a successful move.

spk_0 If you did, it's very readable. Tell application, whatever application to say this thing. Yeah.

spk_0 As you say to K-Bon's point, my favorite thing with code recently when it makes a suggestion is

spk_0 when I say, why did you do that? And it will tell me. And I'm like, oh, and it feels like suddenly

spk_0 I have a teacher again that I haven't had in 20 years. I can just ask questions too.

spk_0 Yeah. It's so annoying. It makes it so what's the opposite of condescending?

spk_0 It's, that's, that's where those that cursor settings file is. You can say in there, do not

spk_0 be condescending to me. Don't you dare be condescending to me. It's wonderful.

spk_0 I'd rather be condescending to me because I'm like, you know, I ask you a question and it's like,

spk_0 oh, I think that this would be a great idea. And then I point out some very obvious flaw.

spk_0 And it's like, you know, you're absolutely right. We should do this.

spk_0 So in your settings file, it should be like stand up for yourself more. It's also like when I'm

spk_0 using conviction. When I'm doing this, it's not like, it's not like I'm an expert on this, right?

spk_0 I'm asking the LLM because I'm not an expert. And so I'm just asking like, hey,

spk_0 is this actually the way that like, I don't know, a next developer would do this? And it's like,

spk_0 oh, you're absolutely right. And I'm like, are you just telling me I'm right? Or do you actually,

spk_0 are you actually like scanning to see like, yes, next developers do prefer this pattern?

spk_0 No, it's just telling you. I know. And that's the problem.

spk_0 Exactly. They're still a little dysfunctional. These tools.

spk_0 Oh, good play.

spk_0 Good segue. Let's talk about this functional dot F M.

spk_0 Well, friends, I'm here with a friend of mine, Michael Greenwich co founder and CEO of WorkOS.

spk_0 We're big fans of WorkOS here. Michael, tell me about auth kit. What is this? How's it work?

spk_0 Why'd you make it? WorkOS has been building stuff in authentication for a long time since the

spk_0 very beginning. But we really focused initially on just enterprise off single sign on

spk_0 sample authentication. But a year or two into that, we heard from more people that they wanted

spk_0 all the auth stuff covered to factor off password auth, you know, with blocking passwords that

spk_0 have been reused. They wanted auth with, you know, other third party systems. And they wanted

spk_0 really work to handle all the business logic around time together identities, provisioning users,

spk_0 and even more advanced things like real based access control and permissions. So we started thinking

spk_0 about that more how we could offer it as an API. And then we realized we had this amazing experience

spk_0 with Radix with this API. Really, the component system for building front end experiences for

spk_0 developers. Radix is downloaded tens of millions of times every month for doing exactly this.

spk_0 So we glued those two things together and we built auth kit. So auth kit is the easiest way to add auth

spk_0 to any app, not just next JS, if you're building a Rails app or Django app or just straight up express

spk_0 app or something. It comes with a hosted login box. So you can customize that, you can style it,

spk_0 you can build your own login experience too. It's extremely modular. You can just use the back

spk_0 and APIs in a headless fashion. But out of the box, it gives you everything you need to be able to

spk_0 serve customers and it's tied into the work OS platform so you can really, really quickly add

spk_0 any enterprise features you need. So we have a lot of companies that start using it because they

spk_0 anticipate they're going to grow a market and want to serve enterprise and they don't want to have

spk_0 to re-architect their auth stack when they do that. So it's kind of a way to like future proof your

spk_0 auth system for your future growth. And we have people that have done that. People that started off

spk_0 and they're like, oh, I'm just kicking the tires and just doing this and then poof, their app gets

spk_0 a bunch of traction. It starts growing, it's awesome. And they go close Coinbase or Disney or United

spk_0 Airlines or you know, it's like a major customer. And instead of saying, oh no, sorry, we don't have

spk_0 any of these enterprise things and we're going to have to rebuild everything. Just go into the work OS

spk_0 dashboard and check a box and you're done. Aside from the fact that auth kit is just awesome.

spk_0 The real awesome thing is that it is free for up to one million users. Yes, one million monthly

spk_0 active users are included in this other gate. So use it from day one and when you need to scale

spk_0 to enterprise, you're already ready. Too easy. You can learn more at authkit.com or of course work OS.com.

spk_0 Big fans, check it out. One million users for free. Wow. Work OS.com or authkit.com.

spk_0 Let's talk about dysfunctional.fm. Why this name? I thought a figure you guys are going to

spk_0 broaden. It's not going to be JS talk the whole time. So but the word function is in there. So

spk_0 tell me more about the name dysfunction. We'll talk about what y'all are thinking.

spk_0 Real quick, I just want to say that we did move beyond a JS specific name because we're writing

spk_0 less JavaScript as a whole as a group. And so we wanted to be able to broaden our horizons. And

spk_0 basically at this point, I'm copying exactly what was on that last change. Look in friends because

spk_0 they said the exact same thing. So they beat you to it. They're definitely on the ball. If we

spk_0 had to compare spin offs right now and not necessarily any fault of your own, but fall through with

spk_0 them folks are like on top. They got like five episodes out and they're doing video. They've got a

spk_0 cast of characters as a mile long. But let's talk about your guys as broadcast. I was going to say,

spk_0 do we want to talk about the behind the scenes of the sort of yes, we are no we aren't. Yes, we

spk_0 are no we aren't JS party side of things that makes dysfunctional seem so dysfunctional right now.

spk_0 Or let's talk about the future of dysfunctional. And how awesome it's going to be when we get it out

spk_0 there, which first episodes are being scheduled to record. And yes, it's going beyond

spk_0 JavaScript and the web. So the core thinking behind the name dysfunctional was actually kind of tied

spk_0 into an experience. Nick and I both had and then we were talking with people at a conference and

spk_0 after we had just learned OJS party is going to go off the air and we're sort of doing our market

spk_0 research of like wait, what is missing out there? And like one of the things with most podcasts

spk_0 in this space and conferences and all of these is they're extremely aspirational. We go and we

spk_0 interview one of the core team members on next or on Redwood or on something else and it's like

spk_0 talking about the cutting edge and where things are. And then we go back to our day jobs and we're

spk_0 still centering a div or whatever the equivalent is in our local environment and language. And like

spk_0 feeling like there's a disconnect, like it's important to have that aspirational content.

spk_0 But there's not enough stuff out there talking about what is the day to day of working as a developer

spk_0 in a variety of different settings and companies. And what's more, how often that feels dysfunctional?

spk_0 Right, I think there's a reason that these like yes, I blew up prod on my first day,

spk_0 like stories are so popular because we've all had that experience. But we don't necessarily all

spk_0 want to broadcast that experience and it can make it then feel isolating if you're out there.

spk_0 And you're like, man, I'm hearing these people talking about all this amazing cutting edge content.

spk_0 And here I am like still working on PHP or whatever it might be. So that's I think the thinking

spk_0 behind dysfunctional. And it's something that I think we all identify with of like, you know,

spk_0 it's great to show up on the show like this and be like, I'm K ball from JS party. Like,

spk_0 I've had people be like, oh my gosh, that's amazing. And I'm like, dude, I'm just a guy.

spk_0 I still screw up and break things and all this stuff. Like I'm dysfunctional at heart as well.

spk_0 Anything to add or subtract, Amy or Nick. Did you catch the functional part?

spk_0 We'll talk about that too. Functional programming. It's a show about functional programming.

spk_0 I mean, being functional as well. Oh, functioning. Yeah. As you go about your program.

spk_0 Functioning adults mostly. Yeah. Will there be any a hoi hois?

spk_0 Uh, absolutely. Are you taking that with you? I mean, I feel like that's yours.

spk_0 That's that's a Nick Nisi. Like you can ask chat GPT and it will tell you that Nick Nisi says a

spk_0 hoi hoi does it really? Wow. That's amazing. I want to go ask you that. What do you,

spk_0 what do you say that? What does Nick Nisi say? Let's say who is Nick Nisi? What is Nick Nisi's tagline?

spk_0 Is that a tagline? Nick? What do you call that? A call sign? Yeah. It's like his nick Nisi's call sign.

spk_0 Rappers do that in their songs, right? DJ Khaled or, you know, like something like that.

spk_0 All the any rappers say hoi hoi in their songs, Nick. It's just yet coming soon.

spk_0 I was curious if it even knew who I was. Yeah. It's close, but it thinks that I partnered with

spk_0 the University of Northern Iowa. Oh, really? Nope. No. When they've been going for the other Amy

spk_0 Dutton. Maybe it doesn't see anything about fiddle playing though. All right. So here's

spk_0 a chat GPT. I couldn't find any information indicating that Nick Nisi has a call sign.

spk_0 He is a JavaScript, a substitute for a developer from Oman, Nebraska, and a panelist on the

spk_0 JS Party podcast. He's also active on social media platforms such as Instagram and X formally

spk_0 Twitter. What does he say at the start of JS party episodes? Sorry. Now I'm just prompting out

spk_0 out. Trying to get chat GPT into the range. I was blown away. I think it introduced me to the

spk_0 I realized that he and I are known by chat GPT. Amy, it sounds like it knows you, but not so well.

spk_0 So I don't know. It's just wanted to throw in the new everything else except for the University.

spk_0 Except for that one. That's a typical hallucination. Here's another one. At the beginning of JS Party

spk_0 episodes, Nick Nisi typically introduces himself with a symbol. Hello. Yeah. That's Claudette. It said

spk_0 that I say, Hey, everybody. In some episodes, he may also share a brief personal update or

spk_0 comment before diving into the main discussion. Bad GPT. Terrible. That's dysfunctional.

spk_0 It does. You guys should have a GPT as your first guest. Is there going to be guests on this

spk_0 new show? Yeah. Okay. I think we're going to have some episodes. You guys didn't have some

spk_0 to. I don't want to see what kind of a show it is. Sometimes shows are just us talking. Maybe

spk_0 you guys are going to do that every time. Yeah. I think that one thing that we want to try and do is

spk_0 build the rapport between us as well. I think some of the best episodes of JS Party are the ones

spk_0 where it's like this. We're just riffing with each other. We have bring our own personalities

spk_0 into it. There's a lot of value in talking to guests about their open source project or whatever.

spk_0 But a lot of times it's a lot harder to go deeper than surface level on any of those. I think

spk_0 that we want to have a healthy balance between that. I like it. I also think it's easier to talk about

spk_0 the things that are going wrong if you've got more of that rapport. We'll have some guests that we

spk_0 can bring in and be like, All right, tell me about something really dysfunctional in your day job.

spk_0 My guess is some people will say, Sure, let me go. And others will freeze. And they'll be like,

spk_0 what mentally skinny their India's exactly right? Right. Now that we've got Nick out of meta,

spk_0 we're all at places where we can talk about what's wrong. That's right. Freedom. Free Nick.

spk_0 Nick is free. Although probably most of your dysfunctional stories were all up in there. So I

spk_0 feel like there's a treasure trove if you could just tap it. Do you have a statute of limitations

spk_0 on that NDA, Nick? Are you free right now? Did I sign an NDA? I don't know. You have to tell us.

spk_0 I don't know. All right. For the most part, it wasn't. It was not any more dysfunctional than any

spk_0 other place. Fair. I did love reading your write up on LinkedIn where you were just kind of

spk_0 transitioning and you were just saying, like, this is like normal company, just like anybody else's

spk_0 where we're trying to write programming. We're working with other engineers that yes, they're smart,

spk_0 but we also write bugs. It was very, very human. I appreciate it. Did you talk about all your

spk_0 bugs during your interview at Work OS? Wait until they hired you and then you're like, I also write

spk_0 bugs. It was a fun interview process there. It really was. I did get a chance to shine in a horrible

spk_0 way, like miserably, bumble through something during an interview. Yeah. Databases are, like,

spk_0 I can use them, but I was mostly using a GUI tool to reset the database and things like that.

spk_0 Sure. Which I thought would be bad, but it's what I do day to day. Cool with it.

spk_0 Yeah. Did you come in the door and say it's me, Nick Nesey from JS Party, oh, oh, hey, hey.

spk_0 Wait for the simple hello instead. I thought that I, like, work OS has been a sponsor of ours.

spk_0 Like we've known those folks for a long time. We have friends over there. So your reputation may

spk_0 have preceded you. Maybe, but I didn't lean on that. And I realized that I never do, which is

spk_0 probably a good thing, but like I never lean into the network that I've been growing too much.

spk_0 And I should maybe change that. But anyway, yeah, I really enjoyed like one thing that I thought was

spk_0 really cool that they had me do that I haven't been through it any other company is they had me go

spk_0 through some of their public documentation and write a friction lock on like here's.

spk_0 Oh, that's cool. Like go through it and tell me everything that you hit friction on while you're

spk_0 going through it. And I came up with like six pages of things, which is pretty good.

spk_0 Especially for a developer experience role. That's great. Yeah. Totally.

spk_0 Very cool. K-ball. Let's know in your world. What's new in my world? I'm writing software with LLMs.

spk_0 I'm writing software that is itself using LLMs. So I'm as you can tell I'm way deep down in

spk_0 that rabbit hole. Other than that, I've been doing some interviews for software engineering daily.

spk_0 Which is fun. You know, getting the interview bug out and then on dysfunctional we can talk about

spk_0 all the things that I break. Nice. And Amy, is there anything you're excited about or

spk_0 interested in looking forward to? Yeah. So I committed to building 12 projects this year. I've

spk_0 built 12.com where I'm just an application a month. And so I'm feeling the heat of the fact that

spk_0 it's the last weekend, January. But you know, K-ball is talking about that to next engineer.

spk_0 I've been trying to figure out, okay, how can I move faster? What are pieces that I can bring with me?

spk_0 And the fact that I am limited to just a month to crank something out is like both a really good

spk_0 thing and a really hard thing. So that's what I'm excited about and been working on.

spk_0 Very cool. I love that. Well, as I said at the top or maybe in the middle, the spirit of JSP

spk_0 already will certainly live on both in the dysfunctional pod and also on Change Login Friends where

spk_0 we'll be pulling lots of different things into including everyone's favorite front and

spk_0 feud will become a Change Login Friends game show. And in fact, your dysfunctional trio here will be

spk_0 on an upcoming episode of that. So definitely subscribe to the Change Log if you want more like this.

spk_0 And of course, to dysfunctional, if you want more like that, anything else we want to say before we

spk_0 call to show. Oh boy, oh boy. These last seven years for K-ball and I have been amazing. Like,

spk_0 it's been so much fun being on the show with you and just like all of the people that we've met and

spk_0 talked to, it's been a ton of fun. And I'm really excited for where the Change Log goes. In the future,

spk_0 I've been listening to all of the episodes and it's been a lot of fun. Continue doing the great work.

spk_0 Thank you. And it's been a blast for me as well. I think that of all the shows that we are

spk_0 retiring. This one is the hardest for sure because it's very close to my heart. K-ball, anything to say?

spk_0 I feel like Nick said everything. So I don't know. Yeah, this has been, it's wow. I think I've

spk_0 mentioned this before, but I've been on JS Party longer than I've ever been at a single job.

spk_0 So y'all, and then particularly looking at Nick and Jared who've been on at the whole time,

spk_0 love you Amy, but you're too new. Y'all have been a part of my professional life for more than

spk_0 any single coworker in a job setting. So that's crazy. I mean, it's wild to think about. And

spk_0 y'all, the listeners who are listening out there, y'all have gotten to hear me more than

spk_0 any of my coworkers have. So you know me. I'm so sorry. Yes, maybe I should be apologizing, but I

spk_0 guess the thing that I would say there then is out to listeners whether or not you come along with

spk_0 us on the dysfunctional ride as we do that. Like don't be strangers. We're just people.

spk_0 If you miss hearing from us, reach out to us. You can, I think we're all findable on LinkedIn and

spk_0 socials and various other things. You can send us messages and we will probably reply. So yeah,

spk_0 I mean, don't be strangers. There you go. Do not be strangers. Definitely reach out. We

spk_0 appreciate you listening to us all these years. 350 episodes. That is an accomplishment. That's

spk_0 quite a few. Most podcasts don't make it to 100 episodes. So that's pretty cool. And we wouldn't

spk_0 have been able to do it, especially in a continued sustained fashion. If it wasn't for you all listening

spk_0 every week, because as you know, we have bills to pay and we pay those bills by having advertisers

spk_0 and those advertisers don't want to advertise unless we have listeners. And so to make the things

spk_0 go around, you are a huge part of that equation. We appreciate each and every one of you for listening

spk_0 all these years. It's been a blast. I've learned a lot about TypeScript. I've learned a lot about

spk_0 podcasting and have grown a lot as podcasters through this. So definitely stay tuned for more from

spk_0 dysfunctional change log and reach out like K-Ball said. Okay. Jared, we'll have an open

spk_0 invitation for you to join anytime you want to talk about TypeScript. I'm ready. I'm available.

spk_0 I have more free time now. I don't have a diaspartiates pod to record. So pick any topic besides that one

spk_0 and send the invite over. And we'll be ready to rock. All right. For Nick, Amy and K-Ball, I'm Jared.

spk_0 This is JAS Party and we'll talk to you on some other show.

spk_0 Okay. So we have more people to thank and I'm sure I'll forget someone but I do want to shout out

spk_0 every JAS Party panelists who helped make the pod awesome through the years starting at the start

spk_0 with the original hosting trio. Thank you to Michael Rogers, Alex, Sexton, and Rachel White.

spk_0 Then came the episode 20 rebooters, Susan Tinn, Chris, Bone Skull Hiller, Faras,

spk_0 Abuqadijay, and Safia Abdullah. We had so many awesome co-hosts along the way shout out to Divya,

spk_0 Emma Boschian, Ali Spiddle, Amelia Wattenberger, Jessica Sachs, and of course, a male who's saying

spk_0 JAS Party wouldn't have been the same without you all. Thanks also to the people behind people on

spk_0 Mike. Adams to Koviaq, Jason Backens, Brian Luzano, Adam Clark, Jake Stutzman, and to all of our

spk_0 guests over the years. Thank you for spending your time with us. And one more thank you to our

spk_0 longtime partners at fly.io for supporting JAS Party all these years. To break master cylinder for

spk_0 cranking out the best beats in the biz. You can still get your fix of the JAS Party theme song,

spk_0 this outro song, and various remixes, you know, by listening to Change Log beats on your favorite

spk_0 music streaming platform. So that's cool. Alright, that's it, the JAS Party is over. You don't have

spk_0 to be annoying.

spk_0 Oh

spk_0 Whoa whoa

spk_0 It must have been called very much

spk_0 It just really must have been called there in the shadow

spk_0 Never have some light on your face

spk_0 To never have sunlight on your face

spk_0 Content to let me shine that's your way

spk_0 They're always walking a step behind

spk_0 So I was the one with all the glory

spk_0 So you're really the one with all the glory

spk_0 You're the one with all the strength

spk_0 The beautiful face without a name for so long

spk_0 A beautiful smile to hide the pain

spk_0 Did you ever know that you're my hero

spk_0 It's everything I would like it to be

spk_0 I can fly higher than an eagle

spk_0 Thank you for putting the tailwind beneath our wings

spk_0 Adam

One last party

Interactive Transcript

Topics Covered

Related Episodes

Out on a Limb | Nov 2nd | Hillhurst United Church

How Do You Respond to Jesus? John 5:16–47 :Daily Devotional, Daily Bible Study

Episode 240: Otto Aerospace CEO Paul Touw talks teardrop, laminar flow, and Phantom 3500

Burnout, Breakthroughs, and the Power of Getting Back Outdoors | Gun Talk Hunt

Share Episode