2024-JS-Party-Transcripts / Digging through Jerod Santo’s tool box_transcript.txt
willtheorangeguy's picture
add all 2024 transcripts
f52f1b6 verified
[0.00 --> 14.58] this is jsparty your weekly celebration of javascript and the web find us on the web
[14.58 --> 22.56] at jsparty.fm there you'll find our popular and recommended episodes clips aka jsparty
[22.56 --> 28.94] the good parts and more big thanks to our partners at fly.io over three million apps
[28.94 --> 35.62] have launched on fly including ours you can too in five minutes or less learn how at fly.io
[35.62 --> 43.82] okay hey it's party time y'all what's up party people i want to say about a new sponsor of ours
[43.82 --> 52.40] jam.dev yes jam.dev is one click bug reports that devs love it's just too easy get jam for free today
[52.40 --> 60.42] at jam.dev and today i'm here with danny grant the ceo and co-founder of jam so danny how do you
[60.42 --> 67.14] describe jam if you've ever reported a bug you've probably had this happen to you you see the bug
[67.14 --> 73.48] you write all the information into a ticket engineer opens the ticket writes works fine on my end closes
[73.48 --> 79.58] the ticket that's because those of us like me who create tickets never put the information that
[79.58 --> 87.06] engineers actually need because we don't know and the words that we use in english to describe an
[87.06 --> 93.54] issue are never specific enough for an engineer like if i write that login didn't work didn't work
[93.54 --> 99.14] could mean so many different things so jam eliminates all of this miscommunication it's a
[99.14 --> 105.08] tool that lets someone like me a product manager or a qa person or someone in support one click to grab
[105.08 --> 110.34] what's on the screen plus everything in dev tools console logs network requests the timing waterfall
[110.34 --> 116.24] session metadata everything and package it into one link in the ticket so an engineer never has to ask
[116.24 --> 122.50] a follow-up question so i've reported many bugs before as a pm as an owner as a whatever and that
[122.50 --> 130.08] sounds like it saves a ton of time this saves afternoons of debugging you no longer have to jump on
[130.08 --> 136.38] a call and share screen to debug you no longer have to show a pm how to find the console and look
[136.38 --> 141.12] for logs engineers say it saves them at least an hour per issue and it's mostly just that back and
[141.12 --> 146.32] forth they no longer have to do but what i hear from product managers who use jam is they used to
[146.32 --> 150.84] after reporting a ticket get a bunch of follow-up questions from engineers and now they create a ticket
[150.84 --> 156.86] and they never hear about it again okay friends go to jam.dev and learn more about what jam is doing for
[156.86 --> 165.04] teams to make bug reporting and all that fun stuff super easy super fast get jam for free today jam.dev
[165.04 --> 167.42] again jam.dev
[186.86 --> 198.44] hello jsparty people those lovely familiar bmc beats mean that this is another episode of jsparty
[198.44 --> 203.50] your party about javascript in the web i'm kball i'm your host this week i'm joined by my friend
[203.50 --> 209.82] jared jared santo how are you doing i'm doing well how are you doing i am excited for this actually so
[209.82 --> 215.86] jared case y'all don't know is behind like every jsparty episode you know he's one of the original
[215.86 --> 222.82] changelog folks he uh even when he's not hosting he's got his ear on what's going on and so today
[222.82 --> 229.06] i'm excited to actually turn the camera around take a deep dive into jared sometime last year we did a
[229.06 --> 234.28] dive into another one of our panelists toolbox we went into nick niecy's toolbox and we covered a
[234.28 --> 238.74] bunch of fun stuff you know nick is a tool geek yes i think we could have spent that whole episode
[238.74 --> 245.00] looking at his vim.files or other things like that but today we're going to dive into jared's
[245.00 --> 251.26] toolbox and i'm kind of a little excited to see how different it is given that jared is not just
[251.26 --> 255.74] a developer but he's also a business owner a long-time podcaster all these other things so
[255.74 --> 263.64] right jared let's go i'm excited man i am your guinea pig your lab rat i am a dead frog covered in
[263.64 --> 270.90] formaldehyde ready to be dissected by uk ball let's go dive deep okay well so let's start with coding
[270.90 --> 276.20] okay because that's you know what everybody here has in common that's where it all starts so you know
[276.20 --> 283.04] what are what tools do you use okay so long journey i mean text editor is the primary tool that i think
[283.04 --> 291.30] of when it comes to coding old school vimmer i've told the story before but i was forced by my college
[291.30 --> 299.50] teacher to my programming 101 teacher and 102 to use vim he said you're going to ssh into a unix box
[299.50 --> 305.68] and you're going to code in vim he said something like pico or nano are great editors if you are
[305.68 --> 311.40] writing a email to your grandma but if you're going to write code you're going to use vim
[311.40 --> 317.96] and so i learned vim not because i saw some fancy video on youtube but because i just didn't have
[317.96 --> 323.28] any other choices came out i had to learn vim if i wanted to pass my class i mean same to be honest i
[323.28 --> 329.46] learned vim you know it was a fortran class i think taken as part of like my either e or physics
[329.46 --> 333.74] or something like that yeah so i don't know if i would have made it over that hump because it does
[333.74 --> 338.28] have that steep initial learning curve had i had a choice but i did have a choice i made it over the
[338.28 --> 346.42] hump and my college teacher was very proficient in vim and he would live code in in class and we would
[346.42 --> 352.00] watch him and be like oh wow you can really move fast if you get good at this thing so that was also
[352.00 --> 358.16] motivating so i learned vim used it for many years still use it off and on all the time mostly in
[358.16 --> 365.58] the terminal mostly ssh into a machine that's my go-to in terminal editor however i do like a nice
[365.58 --> 372.96] graphical user interface specifically for text editing and so i picked up text mate back in the day
[372.96 --> 381.28] eventually moved over to sublime text and sublime text 2 never hopped on the vs code train although i
[381.28 --> 386.34] have installed and i have used it i use sublime text for a very long time and nowadays i have
[386.34 --> 393.40] recently within the last six months switched to zed as i find it the only other sublime text alike that
[393.40 --> 400.94] is fast enough and light enough that it feels sublime-esque sublime was always just faster than all
[400.94 --> 407.14] the other ones and vs code started fast and then has like slowly like gained in my opinion some weight
[407.14 --> 415.18] and so i've been using zed and really i still keep sublime text launched and just for like one-offs
[415.18 --> 420.06] they still can't beat it and like i'm just going to open up a buffer and like do some stuff and do
[420.06 --> 425.04] some text manipulation and then just get rid of it it's still faster but for like project work
[425.04 --> 430.66] i'm chilling in zed so let's dive into zed a little bit what does it look like for somebody who hasn't
[430.66 --> 436.66] used it zed is very much an editor in the ilk of vs code or sublime text i mean it's going to look like
[436.66 --> 443.66] that project you know the file directory on the left you know buffers on the right here i am stuck in
[443.66 --> 449.84] the middle with you i don't know how you describe it it's a text editor you know it's uh it's built by
[449.84 --> 458.84] the major developer that worked on atom atom inside of github nathan sobo and it's built from the ground
[458.84 --> 465.54] up to be really fast and so that's why i appreciate it it also has fancy things built right in
[465.54 --> 473.80] but very much i think feature feature for feature it's going to feel the most like vs code but in my
[473.80 --> 480.54] opinion a little bit faster a little bit better and so i've been i've been using that uh open source now
[480.54 --> 486.42] so that's rad nice what's the plug-in story for zed well now we're going to get into jared's toolbox
[486.42 --> 493.74] and i am not really a plug-in guy so like there are plugins i have a few i install some stuff it's
[493.74 --> 501.22] fledgling i would say it's nowhere near what vs code has so if you are a plug-in person nowhere near what
[501.22 --> 506.76] vim has or neo vim just because it's a pretty new editor and a pretty new community and they've been
[506.76 --> 513.34] building like the extensibility story after they built a few of their other stories up and so it's
[513.34 --> 519.16] like bare bones in there and a lot of first party stuff so they're trying to batteries include that
[519.16 --> 525.80] sucker so a lot of it i just don't care about like it's built in but if you are heavy into plugins and
[525.80 --> 531.54] extensibility like snippet support just landed recently which was holding me off for a long time
[531.54 --> 534.82] because i do have some snippets that i use commonly and i just couldn't even use them in zed
[534.82 --> 541.32] that's there now so like it's getting there but yeah plugins are probably the current thing that's
[541.32 --> 547.14] limping along in that ecosystem just because it's a pretty new editor and i see on their blog their
[547.14 --> 555.04] latest blog post is introducing zed ai are you using the llm based coding tools inside of zed yes so
[555.04 --> 562.30] they have they first built copilot right in when copilot kind of got big and then they realized
[562.30 --> 567.62] there's more to life than copilot and so they made it to where you can switch in different models
[567.62 --> 573.86] different tools and then they've just recently launched like a first party zed ai so you're
[573.86 --> 578.88] plugging into their deal it's a i think it's going to be a paid arm it is a business so they're trying
[578.88 --> 582.96] to find ways of making money i think teams is one of those ways and i think that ai is another one of
[582.96 --> 590.36] those ways that they'll make money off of this editor i just plug it into llama 3.2 and so i do use
[590.36 --> 597.00] llama inside of zed but you can use whatever model you like chat gpt mistral blah blah blah blah blah
[597.00 --> 602.38] all right well so staying on the software train for a little bit any other tools outside of editors
[602.38 --> 607.90] that you find yourself going to uh the terminal i mean i'm a big terminal guy so we could live in
[607.90 --> 613.12] the terminal for a while if you want to are you do you use tmux do you have a are you one of those
[613.12 --> 618.14] who like oh i need my custom terminal setup like what does it look like not so much anymore i was
[618.14 --> 624.76] definitely a tweaker and a configurator as a as a youth when i had the free time and the desire to like
[624.76 --> 628.32] really make my system look cool i used to have one of those terminals that would
[628.32 --> 633.58] come down from the top did you ever have this one where it was like it was a hot key away they're
[633.58 --> 638.62] probably still out there and it was like a heads-up display for your terminal so you just like hit a
[638.62 --> 642.52] hot key and it would slide down over the top of your other stuff and you'd be like coding and then
[642.52 --> 646.80] you'd hit it again and it just slide back up again and when i worked in an office that was cool because
[646.80 --> 650.82] like people would see that and they'd be like wow this guy's hardcore i don't care about such
[650.82 --> 657.40] things anymore so i use like built-in terminal.app i don't have an iterm i'm not using wes term or
[657.40 --> 664.78] kitty or any of the fancy stuff that nick nisi thinks is rad they are rad i'm sure but terminal.app
[664.78 --> 673.10] is just good enough for me i'm a longtime tmux user and i use tmux mostly via a configurator tool
[673.10 --> 680.74] called smug smug which if you have been around long enough you're a tmux guy right i am yeah so do
[680.74 --> 686.60] you use a configurator i know they're like a tmux desk sessions thing no i i have like a i have an
[686.60 --> 692.98] old tmux config that i have that my my finger is like embedded in my brain and fingers that i have
[692.98 --> 697.96] been passed down from machine to machine over the probably decades at this point yeah most of my
[697.96 --> 702.62] configs are kind of set in stone because i'm just like set in my ways you know so i don't really need
[702.62 --> 708.96] to like i'll add a new a new function into my bash rc or whatever it's called csh rc at this point like
[708.96 --> 713.64] once a year i'll add a new function and be like oh that's a nice quality of life improvement but
[713.64 --> 716.94] i've got so much built up over the years that i don't think about it as much anymore
[716.94 --> 725.04] however smug is a tmux configurator so there's also tmux sessions which is a plugin for tmux that
[725.04 --> 730.70] does a similar thing and there's an old ruby gem called tmuxinator which is where i got started and
[730.70 --> 739.28] basically like define your tmux sessions in a yaml file or in a list of yamls and you can start
[739.28 --> 749.34] up tmux with all your settings into a project bam and then you can suspend it restart it etc etc and so
[749.34 --> 756.30] smug is now a a basically a rewrite of tmuxinator in go which means you know universal it's actually
[756.30 --> 760.70] fast yeah well not just that but just it's not gonna like you don't have to do the whole gem
[760.70 --> 765.12] file thing you know like there's no gem install you're not like which version of ruby is this oh
[765.12 --> 770.86] i gotta get a different version of ruby universal binaries for the win i think it's just like brew
[770.86 --> 776.30] install smug and i use smug to control tmux and then you're inside tmux obviously but my tmux is
[776.30 --> 782.38] pretty basic you know just like a couple open not tabs what are they called panels yeah a couple open
[782.38 --> 788.68] panels you know switch between them just basics basics but i do use it all day every day there
[788.68 --> 794.60] is something interesting there in terms of like i feel like being a developer you have to be learning
[794.60 --> 801.28] all the time but like i have a bandwidth for how much i can learn in any particular time and stuff
[801.28 --> 807.24] like my tools right it only changes very slowly because it's it's good enough i agree and i think it's
[807.24 --> 812.44] the kind of thing like i think the toolbox analogy is really fit because when you're first getting
[812.44 --> 816.84] started you got like this empty box and you're like okay i need a bunch of tools and so you go
[816.84 --> 821.88] out and you find them and you test certain ones and you find the one that you like and then you have
[821.88 --> 826.20] to learn how to use that tool and you configure it and you tweak it and you customize it and so it's
[826.20 --> 829.96] like now i have this tool you're not just going to throw that out like you're going to hold on to
[829.96 --> 833.52] that sucker and then you're going to go find another tool and eventually your toolbox kind of full
[833.52 --> 838.44] and at this point in my career and probably yours as well it's like i might change a tool here or
[838.44 --> 844.68] there or tweak one like every quarter or once a year or you know i get excited when i have a new tool
[844.68 --> 848.66] i'm like hey i grabbed one you know because i just i got a full toolbox i don't really need
[848.66 --> 856.80] to be acquiring anymore absolutely i will say the the sort of rise of the ai assisted coding stuff and
[856.80 --> 863.04] like how now with something like claude uh or sonnet and things like that like right it really is good
[863.04 --> 868.14] enough that if you don't start figuring out how to put that in your toolbox you are getting left
[868.14 --> 873.24] behind so that's been a forcing function for me me too and i'm probably slightly behind the curve
[873.24 --> 879.82] even from you on that i am using them but i'm not i'm sure i'm not fully leveraging because i haven't
[879.82 --> 884.62] tested all of the different ones and made sure i'm using the best one i haven't tried cody i haven't
[884.62 --> 889.96] tried claude i have tried cursor i mean we don't want to maybe get into ai later but i've tried a bunch of
[889.96 --> 895.84] these things but i'm also kind of like a vanilla llm user at this point kind of also letting some
[895.84 --> 901.42] of it shake out totally and these funded companies build the tools better to where it's like okay this
[901.42 --> 906.24] one's amazing go get it you know all right so we've talked about terminals we live in the terminal a
[906.24 --> 912.62] little bit anything else that are go-tos do you you know have maybe let's get into dev frameworks and
[912.62 --> 918.22] coding languages like do you have a go-to for quick scripts versus product versus what have you
[918.22 --> 924.94] yeah so changelog.com is all elixir so it's using the phoenix framework i've been writing that i've
[924.94 --> 931.66] been maintaining that code base and advancing it since 2016 which is probably the longest single
[931.66 --> 939.06] code base of my career as a contractor i've you know worked on projects and then moved on or i think
[939.06 --> 944.74] i've maintained something for a few years i did have long-term customers so maybe like three or four or
[944.74 --> 949.66] five years it'd be like the longest i worked on a code a single code base but i've been on this one
[949.66 --> 956.28] since 2016 so like that's coming up on eight years i guess and it's all elixir in the back end and
[956.28 --> 964.66] html oriented i'm very much an html oriented web developer with javascript sprinkles and that's the
[964.66 --> 969.84] way i like it and that's the way i've written it and that's works just fine for me there's no
[969.84 --> 978.68] spa framework in the front end of changelog.com and that's languages now that's for i guess product
[978.68 --> 983.68] that'll be a product right for one-offs and scripts i still usually just start with bash and then go
[983.68 --> 990.28] immediately to ruby as soon as i'm outside of bash i just can't get more expressive you can do a lot in
[990.28 --> 994.92] elixir and i have a i have like two elixir scripts i wrote and i'm like there's just more ceremony here
[994.92 --> 1002.10] and ruby like for text manipulation and looping and command line stuff and like shelling out and
[1002.10 --> 1007.44] getting a result and then i know i've spent time with sed and auk and i can do all those things but
[1007.44 --> 1012.92] like ruby for me is just way faster to get that stuff done totally i still like even if i have to
[1012.92 --> 1017.52] do like some math that's not immediately obvious to my head i just like go to the terminal hit irb and
[1017.52 --> 1023.88] there i go seriously so fast it's the closest you get to like your thoughts becoming code like your
[1023.88 --> 1029.76] pseudocode is basically like add a dot here and some curly braces and now it's actual code you know
[1029.76 --> 1035.62] and for me that's just probably never going to be replaced also because i got you know so many years
[1035.62 --> 1043.02] right in that language before i went to elixir that it's just so easy to get stuff done so yeah i got a
[1043.02 --> 1048.68] ton of ruby scripts just you know all over my hard drive um in the terminal i guess one other thing i
[1048.68 --> 1054.44] didn't mention a very cool tool that i added in my toolbox last year is a2n we've heard of a2n
[1054.44 --> 1060.70] i have heard of it i have not added it to my toolbox yet so talk through it yeah this is a very
[1060.70 --> 1065.86] cool one it's basically like i like a tool that just upgrades your life and doesn't require anything
[1065.86 --> 1071.48] else like there's no learning curve there's no adoption really this is a shell history upgrade
[1071.48 --> 1077.40] basically built by a gal named ellie huxtable and we've had her on the show a couple of times
[1077.40 --> 1083.36] she's she's whip smart and great great user experience person and basically like you know
[1083.36 --> 1089.82] the up arrow or the control r searching your shell history if you've done that which we all have a
[1089.82 --> 1095.90] hundred times like this just basically improves that in every in every possible way and so you install
[1095.90 --> 1101.86] it it runs in the background and it takes over certain keystrokes in your terminal in order to
[1101.86 --> 1107.36] have fuzzy search on your shell history a lot like fzf would be if you can figure that yourself
[1107.36 --> 1113.10] but it's very pretty as well and then it also offers advanced things i don't care about like
[1113.10 --> 1118.40] syncing your history across machines and all this kind of cool stuff i'm a one machine person so i don't
[1118.40 --> 1123.20] really have that problem but if you had a desktop and a laptop and you wanted shared shell history
[1123.20 --> 1128.08] costs everything like she has a service that encrypts that and synchronizes it around and
[1128.08 --> 1133.76] it's an awesome cool it also provides stats which is fun like you can look at your most used commands
[1133.76 --> 1140.74] and stuff like that and so for someone who absolutely depends on control r yes command r just working
[1140.74 --> 1148.16] like is it drop in replacement you won't okay there's one little thing that it changes which is really
[1148.16 --> 1153.84] just like an orientation thing that took me a minute which is when you control r command or
[1153.84 --> 1158.90] whichever one it is which for those who haven't done this before this provides like backwards search
[1158.90 --> 1166.62] of your recent command so i can hit control r and type l and it'll be like ls l something else you know
[1166.62 --> 1172.36] and you can like pick that one and just hit enter and it'll execute and so this is very handy when you do
[1172.36 --> 1178.08] that in classic terminal it'll just like do it right there where you are and when you do that with
[1178.08 --> 1185.22] a to n it will bring up kind of a reverse chronological list not reverse chronological
[1185.22 --> 1192.26] reverse last in first out list like your most recent used going upward that match the current
[1192.26 --> 1196.42] search and because of that it moves your search down to the bottom of the screen and so it literally
[1196.42 --> 1201.78] just moves it from where it was to the bottom and that just took me maybe a couple days to like
[1201.78 --> 1206.68] it bugged me for a day or two because i'm used to it just being i'm staring at it and now it moves it
[1206.68 --> 1212.14] down to the bottom other than that which is a tiny thing it's a drop in replacement and it's better
[1212.14 --> 1218.98] in every conceivable way so i just install it and run it and have it look back so try that one out and
[1218.98 --> 1225.26] maybe if you do have any uh issues that it changes because those of us with long standing habits it
[1225.26 --> 1232.40] doesn't take much to be like you know this ruined my life in one minor nitpicky way and that's
[1232.40 --> 1237.20] sometimes enough friction to be like not worth it for for us um so i'd be interested to hear your
[1237.20 --> 1244.14] results but man i i installed it and have not looked back all right well i just installed it i'll let you
[1244.14 --> 1251.38] know cool it doesn't appear to have broken my command r nope so that's a big one that i that i installed
[1251.38 --> 1260.28] probably within the last year other than that i'm pretty vanilla terminal tools besides tmox smug a2n
[1260.28 --> 1266.94] let's talk maybe about some of the other non-software things then because you have this other side of
[1266.94 --> 1274.68] that you're a small business owner you're a podcaster for longer than almost anyone yeah what are your go-tos
[1274.68 --> 1281.12] let's start with with podcasting maybe what are your go-tos there sure so in software world as you
[1281.12 --> 1286.58] know cable we use riverside to record and that's a web app that we pay for it's a software as a
[1286.58 --> 1294.26] service web app that puts i think almost every newfangled web technology into play in order to have a
[1294.26 --> 1298.82] really nice experience for us and that handles a lot of the problems that we used to have to work
[1298.82 --> 1304.18] around it used to be a lot harder to podcast we had skype and we had multiple enders and we had
[1304.18 --> 1309.88] you know this that and the other thing call recorders blah blah blah riverside has really simplified
[1309.88 --> 1317.02] the tools that we need and we've been using that for a couple of years now pretty pretty successfully
[1317.02 --> 1323.24] there are all competitors to that which are also good so lots of cool tools in the podcasting space
[1323.24 --> 1330.28] and then we take we take the recorded stuff and we do all our editing in adobe audition now there's a
[1330.28 --> 1334.40] lot of tooling around that now that didn't exist before as well around editing specifically
[1334.40 --> 1340.60] whether it's descript and the ability to read the transcript delete words out of the transcript and
[1340.60 --> 1348.18] it edits the audio which is just a really cool idea but for us we've been doing it this way for so long
[1348.18 --> 1354.44] that we just prefer kind of the power and control that we have in audition that you give up when you
[1354.44 --> 1359.48] use some of these online tools for editing so everything goes into audition all of our files are synced
[1359.48 --> 1364.94] via dropbox so we don't think about dropbox very often but it's just a core piece of our business
[1364.94 --> 1371.14] 100 long time users of that and then everything else in terms of publishing is all just self-built
[1371.14 --> 1378.74] so i built a open source web app like the one i mentioned that does everything from like we have a
[1378.74 --> 1387.22] mp3 file to completely published syndicated promoted blah blah blah is all just coded up
[1387.22 --> 1394.16] so custom tool a custom tool gotta love those custom tools oh yeah well the nice thing about them
[1394.16 --> 1401.92] is that they're custom so it's both a gift and a curse right so like we can do whatever our hearts
[1401.92 --> 1406.56] imagine and we've done some really cool stuff like our transcripts get synchronized over to github
[1406.56 --> 1411.90] and they're open source on github and so you can actually help improve them there and then if you
[1411.90 --> 1416.52] improve them there they get synchronized back and you know they get sucked back into our database
[1416.52 --> 1423.38] stuff like that we do a lot with chapters and with mp3 metadata that you couldn't do elsewhere
[1423.38 --> 1429.64] so we've been able to really customize it to be exactly the way you want and then the curse is
[1429.64 --> 1434.12] if we want something new we got to build it like everything you have to build it all you know
[1434.12 --> 1440.62] yeah it's broken i gotta fix it if we want it we gotta build it and so it's that onstanding
[1440.62 --> 1447.02] you know gift and curse of custom software it's really paid dividends though we were on wordpress
[1447.02 --> 1453.24] back in the day and look at us now i mean we could have been a wp engine customer we would be so mad
[1453.24 --> 1459.48] right now oh man well and you clearly have not fallen into the common developer challenge of
[1459.48 --> 1464.86] rewriting the platform more often than you publish no we couldn't do that because we published five or
[1464.86 --> 1471.36] six times a week i'm not that fast of a coder but we definitely thought about doing some rewrites
[1471.36 --> 1477.58] and of course when you when you build something eight years ago you know it has its warts the
[1477.58 --> 1483.62] technologies that you pick are no longer best in breed i actually think i picked a pretty good tech
[1483.62 --> 1489.34] for this but i'm actually very pleasantly surprised that i don't have any sort of itch i think there was
[1489.34 --> 1494.32] once when like should we go jam stack with this because it's so static content like a lot of it's
[1494.32 --> 1499.86] static content and so pre-built html makes a lot of sense and we're not doing that we're doing
[1499.86 --> 1505.08] caching we do some stuff like our mp3s get pushed to r2 and then cdn from there so there are and our
[1505.08 --> 1510.50] feeds as well which are just massive xml files at this point because we want to put all of our
[1510.50 --> 1516.72] episodes in there so we have like 12 megabyte xml files that we're serving up which does get slow
[1516.72 --> 1521.64] even with a fast programming language if you're if you're dynamically producing that each time so we push
[1521.64 --> 1526.10] those off to r2 and then we we serve them from the from the cdn from there so we've done a few
[1526.10 --> 1532.40] things that are jam stack ish but i definitely have thought once or twice maybe we should just
[1532.40 --> 1538.52] switch to an entire jam stack approach and it's never been worth the lift because there's so much
[1538.52 --> 1542.86] surface area to the app at this point like you think it's a simple app and it is and then you go
[1542.86 --> 1547.60] look at all the different stuff it does and you're like it's just a lot of stuff here that i had to
[1547.60 --> 1554.56] rewrite it's simple in the surface area it exposes to people but there's a lot of complexity
[1554.56 --> 1560.92] under the covers totally at the risk of diverging too far are there features or functionality that
[1560.92 --> 1567.14] you want to add to the app that you're looking for in the future well we definitely want to
[1567.14 --> 1572.56] provide like i think there's definitely some stuff we could do with our transcripts
[1572.56 --> 1580.80] in our episodes that are language model focused that would provide value similar to just like a
[1580.80 --> 1587.06] really upgraded search functionality where you can say questions like have they ever talked about
[1587.06 --> 1592.02] jam stack on jsparty and just get an answer you know versus like i'm gonna go search we did have
[1592.02 --> 1597.16] somebody who built a thing like that but it was more like talk to an llm version of jared and kball
[1597.16 --> 1601.62] and with our personalities but the end product like you play with it for five minutes and then it's done
[1601.62 --> 1607.30] you're like okay that was that is what it is but i think having a librarian so to speak because we
[1607.30 --> 1611.72] have thousands of episodes now and we get questions like have you guys ever done this show and i have
[1611.72 --> 1615.88] to go find it through the search functionality or not and then be like i thought we did but i can't
[1615.88 --> 1621.00] find it i think we could definitely build something there that's a big one that would be cool but just
[1621.00 --> 1625.94] nice to have and then one thing that we really want to do we have taken steps towards with our custom
[1625.94 --> 1631.30] feeds but haven't gone totally it's like bring our membership program completely on to the site
[1631.30 --> 1638.00] and off of the supercast which is another tool we use for our memberships that would be cool
[1638.00 --> 1645.02] that would be cool yeah complete custom put up filters maybe llm based filters is this about this
[1645.02 --> 1650.96] or that right then you get real expensive so they got to be paying right yeah exactly i did build custom
[1650.96 --> 1656.02] feeds this year which was something that our subscribers have asked for for a very long time
[1656.02 --> 1662.30] one small wrinkle in our membership is for instance if you're just a js party listener
[1662.30 --> 1667.94] and you want to support js party you sign up for changelog plus plus it feels good you support us
[1667.94 --> 1673.32] but because of the way supercast works they can only take one of our feeds and turn that into the
[1673.32 --> 1678.42] private feeds for everybody and so we have like a master feed of our plus plus content that we send
[1678.42 --> 1684.22] to supercast and so all of a sudden now you have to get all of our episodes instead of just js party
[1684.22 --> 1688.16] which is you know people understand like sorry that's just the way it works please just delete
[1688.16 --> 1693.16] the ones you don't like or if you can find a podcast app that provides filters inside the app
[1693.16 --> 1697.98] which there are real nerdy podcast apps where you're like i want to subscribe to this feed but only if
[1697.98 --> 1703.36] this string matches or whatever please do that instead and so that was a bummer because it's we you
[1703.36 --> 1706.62] know we say it's better but when you sign up and you're like this is actually worse than what i was
[1706.62 --> 1712.96] doing before besides the bonus content and the ad free so now that i built custom feeds it solved that
[1712.96 --> 1718.14] problem you can go in and create a custom js party only feed and subscribe to that and i thought i
[1718.14 --> 1723.80] had to bring everything first party to get that done and then i realized why just build the custom
[1723.80 --> 1729.04] feeds feature and let all i needed on our app was to know if you're a plus plus member and because
[1729.04 --> 1735.46] supercast uses stripe it's our stripe account i can just hit the stripe api figure it all out and so i'm
[1735.46 --> 1740.54] kind of sidestepping which has been really nice because it lets me solve that problem for folks but
[1740.54 --> 1743.86] then i'm like now i don't have less motivation to get off supercast because that was one of the
[1743.86 --> 1748.18] major reasons i wanted off and now it's mostly about money because they take a little bit of
[1748.18 --> 1755.06] money which is fine but we'd obviously save some by not having to use them and just that autonomy and
[1755.06 --> 1762.52] like complete control of the experience which are less less motivating than you know custom feeds so
[1762.52 --> 1769.26] that's part of the app this year and people are loving it they're loving it yeah anyways we're yeah
[1769.26 --> 1775.26] we're upstream now take us back take us back to the main river i mean i think it's fun to think about
[1775.26 --> 1781.88] we talk about building you know selecting your tool chest but as a business you're building your own
[1781.88 --> 1787.06] tool chest as you go right your business is not this software the software is the tools for your
[1787.06 --> 1792.60] business yeah 100 and you have to decide like just like anything else do we build or buy you know
[1792.60 --> 1799.66] and there's a lot of off-the-shelf tools there are way more now than there were in 2016 and 2015 when
[1799.66 --> 1809.16] we made this choice where it's like would i build a custom platform today probably not it would depend
[1809.16 --> 1814.30] on what exactly we're trying to build but we do some partner broadcasts so like we produce grafonda's
[1814.30 --> 1820.76] big tent and for them we're like just go sign up for transistor it's a great service uh fireside was
[1820.76 --> 1826.26] another good one that just got acquired by john nunamaker who's a friend of ours and so that one's
[1826.26 --> 1831.66] going to be getting better they're like there's options and most people do not need what we built
[1831.66 --> 1838.10] back then but now that we have it you know use it to make our business better well there's no shortage
[1838.10 --> 1843.82] of helpful ai tools out there but using these ai tools means you got to switch back and forth
[1843.82 --> 1850.12] back and forth between yet one more tool so instead of simplifying your workflow it just gets more
[1850.12 --> 1856.04] complicated but that's not how it works when you're using notion notion is the perfect place
[1856.04 --> 1861.82] to organize lots of stuff tasks tracking your habits writing beautiful docs collaborating with
[1861.82 --> 1868.02] your team knowledge bases and the more content you add to notion the more this cool thing called
[1868.02 --> 1875.58] notion ai can personalize all of the responses for you unlike generic chatbots notion ai already has the
[1875.58 --> 1883.86] context of your work plus it has multiple knowledge sources it uses ai knowledge from gpt4 and clod and
[1883.86 --> 1889.98] that helps you chat about any topic and here's the kicker now in beta notion ai can search across
[1889.98 --> 1897.00] slack discussions google docs sheets slides and even more tools like github and jira those are coming soon
[1897.00 --> 1903.58] and unlike specialized tools or legacy suites that have you bouncing between different applications
[1903.58 --> 1910.74] notion is seamlessly integrated infinitely flexible and beautifully easy to use so you are empowered to
[1910.74 --> 1917.46] do your most meaningful work inside notion from small teams to massive fortune 500 companies these
[1917.46 --> 1926.22] teams both small and large use notion to send less email cancel more meetings save time searching for
[1926.22 --> 1932.48] their work and they reduce spending on tools which helps everyone stay on the same page you can try notion
[1932.48 --> 1939.78] for free today when you go to notion.com slash jsparty that's all lowercase letters notion.com
[1939.78 --> 1947.30] slash jsparty and hey notion is powerful it's easy to use notion.ai is amazing and when you use our link
[1947.30 --> 1953.22] you are supporting jsparty and i know you love that once again notion.com slash jsparty
[1953.22 --> 1962.50] let's talk a little bit about the business because i would bet there's a fair number of people here who
[1962.50 --> 1966.06] want to run their own business in some form or another maybe it's just a freelance business right
[1966.06 --> 1972.26] getting out from under people's thumbs what are the tools you use to run the business what did you use
[1972.26 --> 1977.00] back when you were contracting like what what would you lean on if somebody's in that or what's what's in
[1977.00 --> 1986.26] that side of your tool chest right okay so obviously you have things like payroll and invoicing and
[1986.26 --> 1991.16] then you have collaboration and communications and i can go a little bit through checklists on that so
[1991.16 --> 1996.44] we use fresh books adam for those who don't know adam sokoviak my business partner and co-host of the
[1996.44 --> 2003.70] changelog he signed us up for fresh books like probably a decade ago and i've always used it i used
[2003.70 --> 2009.42] harvest when i was a contractor i really liked harvest had lots of flexibility and was also
[2009.42 --> 2013.18] simple they had a really i mean i think they're still doing their thing i was talking the past tense
[2013.18 --> 2016.76] because i don't use it anymore but it sounds like they're dead or something like i'm sure you can go
[2016.76 --> 2021.86] out there to get harvest.com and check it out today i really liked harvest for invoicing but
[2021.86 --> 2028.16] uh fresh books is what adam was using and totally serviceable good service i don't know there's my
[2028.16 --> 2035.76] review it works well for invoicing we use gusto.com for payroll and they're great because they provide
[2035.76 --> 2043.68] all the kind of all the things that small businesses need but don't want to have to hand roll hr stuff
[2043.68 --> 2052.20] vacation stuff blah blah blah built into gusto taxes etc dropbox is a big one as i already mentioned
[2052.20 --> 2059.30] we use that for all of our file sharing and stuff and then everything else is like slack uh zulip we've
[2059.30 --> 2065.18] added for our community now we're kind of transitioning over to zulip and i can't think of anything else what
[2065.18 --> 2069.22] else is there for our business k-ball i mean you really don't depending on what you're doing you
[2069.22 --> 2077.52] really don't need that much like i i run my whole side business off of fresh books also and like google
[2077.52 --> 2081.68] suite essentially right yeah i guess we do use google suite as well for for email
[2081.68 --> 2087.20] and like and docs it's pretty much all you need for for something like consulting or i'm doing
[2087.20 --> 2093.94] coaching right right like that like you need a way to communicate with people and you need a way to
[2093.94 --> 2098.70] build people yeah totally and whatever service you're delivering so a way to deliver your service
[2098.70 --> 2107.36] right it's kind of i think on the sales side or the customer side adam has tried multiple crm tools
[2107.36 --> 2113.16] over the years i'm not sure if we've ever landed on a crm that we've been like this is the one for us
[2113.16 --> 2118.34] but we aren't a typical sales team either so i'm not sure what you would say about that i just don't
[2118.34 --> 2122.52] use it so i know we've i've been signed up for lots of different services over the years like we're gonna
[2122.52 --> 2127.98] try this one now okay yeah i mean it depends right and as you highlight as your business skills depending
[2127.98 --> 2133.78] on what you're doing you need different things but like to get started it's really not much yeah i mean
[2133.78 --> 2137.84] all you gotta do is ask yourself like what am i doing here and then how do i do each thing it's
[2137.84 --> 2142.96] like well we are selling ads right we're selling these sponsorship campaigns so we have to be able
[2142.96 --> 2148.12] to sell one produce one and then invoice somebody at the end of the day and so you figure that out
[2148.12 --> 2151.16] and it's like well we're hiring somebody okay we're gonna be able to pay them and you figure that out
[2151.16 --> 2159.26] so you just kind of add these things as the needs arise and uh no there's so i mean to i guess
[2159.26 --> 2163.58] silicon valley and the greater tech industries credit like there's a lot of good services you
[2163.58 --> 2167.68] can build your business on nowadays yeah to where it's not that hard to get started it's really not
[2167.68 --> 2173.54] like you have a plethora of choices there's competition prices aren't too bad you can start
[2173.54 --> 2180.36] a small business relatively easy and cheap uh using internet-based tools yeah it's shockingly easy i feel
[2180.36 --> 2184.06] like the core problem is figuring out what what is something you can sell that people want to buy
[2184.06 --> 2190.38] like that's still the really hard part isn't it i mean yeah i mean that's the core of things but all
[2190.38 --> 2195.06] these operational headaches that used to do capital you need all this other stuff for like most businesses
[2195.06 --> 2202.06] not a big deal there's options yep so that's podcasting tools business tools what other kind
[2202.06 --> 2207.28] of tools you want to talk about what other kind of tools you got man all right so um well how about
[2207.28 --> 2212.42] as a podcast listener you mentioned like geeking out on more extensive tooling like i'm still on apple
[2212.42 --> 2217.10] podcasts what do you use oh we gotta get you off there man all right actually it's gotten a lot
[2217.10 --> 2222.90] better but tell me the tools what should i be doing so i i'm a big fan of indie podcast apps
[2222.90 --> 2229.00] and i think that you can find one that you can fall in love with and i like to have be able to have
[2229.00 --> 2233.10] a relationship with the person who's building the tool i just feel like that's one of the reasons
[2233.10 --> 2238.04] why zulip is so cool for us it's like we're like we know the zulip people now and if there's a
[2238.04 --> 2241.16] thing that we want to talk about with them like we can just talk about it whereas if you don't like
[2241.16 --> 2246.06] something about your apple podcasts you know go take a long walk off a short bridge you might as
[2246.06 --> 2250.72] well because you're not going to talk to apple about it you know you're sol so you know i can
[2250.72 --> 2255.06] tell you which podcasting app i use but there's a whole bunch of small ones that are really great
[2255.06 --> 2261.90] overcast is the one that i use pocket casts is really cool and open source owned by automatic so
[2261.90 --> 2268.78] you know so so open source in quotes sure i don't know i mean like what do you i mean after what
[2268.78 --> 2273.38] after what they've been doing what do you call that anymore right like yeah i don't know good
[2273.38 --> 2278.18] question i mean it's mit and everything like that but you know matt's proven himself to be uh
[2278.18 --> 2283.08] i don't know kind of unreliable or on what's the word for someone who you can't who's uncertain
[2283.08 --> 2287.32] you know you don't know what he's gonna do next i mean wordpress is theoretically gpl
[2287.32 --> 2294.04] like it is gpl right so now it's just like trademark and uh services i don't know it's
[2294.04 --> 2299.70] getting murky for sure but as a pocket cast user it's just cool that it's open source you can watch
[2299.70 --> 2304.32] them build it you're not necessarily i wouldn't go any further than that myself but it's just cool
[2304.32 --> 2310.38] that it's out there and what's interesting was that it wasn't open source matt mullenweg buys it
[2310.38 --> 2315.70] via automatic any open source is a thing so like the guy used to have a lot of street cred with me
[2315.70 --> 2319.84] that's why i'm very confused at this point because i thought that was very cool maybe he went on one
[2319.84 --> 2325.80] of those uh ayahuasca trips that has everybody freaking out yeah yeah he was enlightened and he
[2325.80 --> 2331.62] saw the light and he wanted he wanted to change his ways i don't know so that's a good one there's
[2331.62 --> 2339.84] podcast addict there's castro there's all these ones and they just have specifically the thing i like
[2339.84 --> 2345.60] about them is the chapter support is like the ability to use the advanced features and all the new
[2345.60 --> 2352.14] burgeoning podcasting namespace stuff so like podcasting community is trying to innovate and
[2352.14 --> 2359.56] change and make podcasting better but because apple and spotify kind of have a stranglehold on those
[2359.56 --> 2366.08] audiences it's hard and so the more people get using the indie apps who are actually innovating and
[2366.08 --> 2373.28] trying new things then the better off the whole ecosystem is of like open podcasting and so i'm always
[2373.28 --> 2378.98] in favor of the open side of podcasting even though we exist everywhere but much rather have
[2378.98 --> 2383.60] you on apple podcast than spotify because spotify is like the worst way to listen to podcasts i just
[2383.60 --> 2388.56] can't understand it i've just been on apple podcast by default right i've been there for so long
[2388.56 --> 2393.12] yeah why not but all right i'm inspired i'm gonna pull my phone out and install overcast
[2393.12 --> 2398.66] okay cool if you're listening to this on spotify there's better worlds out there i'll tell you
[2398.66 --> 2405.18] there's podcast apps that support chapters in such a way that as a podcaster i can name a chapter
[2405.18 --> 2411.26] i can add a link so that when you're looking at that chapter you can click on the name of the chapter
[2411.26 --> 2418.42] and follow the link to the thing that we're talking about so if we're discussing an article that we just
[2418.42 --> 2423.10] read i can put the link to the article in there and you can click on it and read it while we talk
[2423.10 --> 2429.34] about it and i can attach imagery so while that chapter is active it becomes the it takes over
[2429.34 --> 2434.22] your cover art for the podcast that you're listening to and so i can reference a diagram
[2434.22 --> 2442.16] or a meme and put that meme in your podcast app while we talk about it spotify won't let you do that
[2442.16 --> 2448.12] so if you want better better memes you want the better means man it's all about the memes
[2448.12 --> 2453.18] anyways that's my sales pitch for indies plus you support independent developers which feels good you
[2453.18 --> 2459.50] know totally bringing this back around so we've talked about coding languages and dev frameworks
[2459.50 --> 2464.00] podcasting running the business other good stuff i realized when we talked about coding though we
[2464.00 --> 2470.64] barely touched on javascript and we are on js party so when you're writing javascript what are your
[2470.64 --> 2477.58] go-to libraries frameworks etc oh cable i'm gonna kill your street cred right here you're gonna get
[2477.58 --> 2483.74] me called out as an imposter on js party um as i said earlier i'm a javascript sprinkles person but
[2483.74 --> 2489.16] i'm also a right tool for the right job person so i've definitely used react and i've actually went
[2489.16 --> 2494.76] out and chose to pull react into a project and so it's not like i'm just anti i do not like single
[2494.76 --> 2500.20] page apps i don't like to use them i don't like to write them saying this is one of the
[2500.20 --> 2507.30] people who wrote one of the early popular single page apps to exist when i rewrote the groveshark
[2507.30 --> 2517.64] flash app which had millions of users in html and javascript back in like 2012 2011 no it was pre
[2517.64 --> 2525.76] yeah 2010 so i like server rendered html and javascript what do i use i use all the built-in
[2525.76 --> 2532.52] stuff mostly so i use the dom apis query selector all i'll write small functions that wrap that
[2532.52 --> 2538.38] our app right now changelog.com which like i said is eight years old uses a jquery alike
[2538.38 --> 2546.02] called umbrella js that was like a super light version of jquery's api and event handlers and
[2546.02 --> 2551.68] callbacks and all the stuff that people think is terrible practice today i don't have a favorite
[2551.68 --> 2558.90] front-end framework because i just don't use them as there there's an unpopular opinion i haven't
[2558.90 --> 2565.82] used any i mean i've used react i've done toy apps but i don't really count those and so i just use
[2565.82 --> 2572.72] javascript man i don't use typescript i like node.js i've used dino and i like that node.js i think is a
[2572.72 --> 2578.78] great platform and i've used it successfully multiple times really like it and if javascript
[2578.78 --> 2585.62] was a little bit more ergonomic for me i would probably use it even more it's gotten better
[2585.62 --> 2592.28] but it used you know i also have like a long-standing i remember in the old days so it's hard for me to
[2592.28 --> 2597.04] have like that change of emotion around it you know like i'm i'm on js party because i love web
[2597.04 --> 2601.68] development i love the web platform i'm not a javascript lover and i'm a typescript hater y'all
[2601.68 --> 2608.06] know that so there you go you called me out there's where i stand i don't really use front-end frameworks
[2608.06 --> 2615.02] well i was i was going to look and try to see you know how much javascript is there on changelog.com
[2615.02 --> 2621.56] not very much and the core app js if i load the home page it's minified but i get a grand total of
[2621.56 --> 2630.38] 64 kilobytes that's not bad right that's not bad yeah that's not much i don't want much i want my
[2630.38 --> 2635.04] web pages to load as fast as possible to as many people as possible and constrain the devices
[2635.04 --> 2642.26] and if i can get away with it i would use zero not because i don't like it because i think that
[2642.26 --> 2647.50] that's ultimately a more rock solid and fast experience for most people so i sprinkle it on
[2647.50 --> 2653.14] in the form of i guess 64 kilobytes i think maybe more for the player i'm trying to figure out that
[2653.14 --> 2658.34] looks like that's loaded separately the player i think should be bundled in though is it wow yeah
[2658.34 --> 2664.06] it just looks like it's not and we used to use turbo links to make a single page app like experience
[2664.06 --> 2671.14] but i actually have also removed that just because people aren't really browsing our website as we
[2671.14 --> 2675.88] wanted them to back when we first built it when we had like a news feed we thought there'd be commentary
[2675.88 --> 2680.80] blah blah blah and you'd like listen to a show and like checking out different stuff and so we wanted
[2680.80 --> 2687.92] that player to stay um and so we built turbo links based spa like experience where the urls change
[2687.92 --> 2693.64] but you're not reloading the whole page and we rocked that for probably six years and i took it out just
[2693.64 --> 2699.44] last year because it's just not the way people used our website and it's just additional bloat and had a
[2699.44 --> 2705.28] few things where it would introduce little bugs here and there that were manageable but if you can just
[2705.28 --> 2710.30] not have that class of problems then why not so i took that out as well the player is the majority
[2710.30 --> 2716.76] of the javascript and then you have some stuff like some overlays there's not much there's really
[2716.76 --> 2721.42] not i mean i'm looking at this i'm looking at my network tab right now and like my extensions are
[2721.42 --> 2728.38] injecting more javascript than your website has that makes me feel good k-ball i like that yeah
[2728.38 --> 2735.44] all right so go to tools no javascript not no javascript but just just enough just a pinch just a pinch
[2735.44 --> 2739.72] no i don't have a go to front-end framework if i was going to build something today
[2739.72 --> 2744.70] i have used svelte recently i take that back i wouldn't call my go-to framework but i use svelte
[2744.70 --> 2749.92] and i thought this thing is cool so i that one got me i would probably grab svelte kit or
[2749.92 --> 2755.58] preact maybe just but it depends on what i'm building of course but the chances of me doing
[2755.58 --> 2762.00] a single page app unless it's like i'm building a gmail competitor or something very low very low
[2762.00 --> 2764.04] more likely a spotify competitor
[2764.04 --> 2772.58] someone should replace spotify just for podcasting just for podcasting yeah i got no problem with the
[2772.58 --> 2777.20] music player and it's gotten better on podcasting they're so they're supporting transcripts now and
[2777.20 --> 2782.86] they do support chapters halfway which is better than it used to be used to be no way so it's not
[2782.86 --> 2787.00] like they aren't trying but when they do they do it their own way like you're gonna write the spotify
[2787.00 --> 2792.86] version they're not going to adopt an open standard it's always like some engineer had to show off
[2792.86 --> 2798.82] inside spotify and build their own spec you know that kind of stuff makes me mad too anyways now i'm
[2798.82 --> 2803.74] just rambling and ranting now we're rambling we're we're i think we've wrapped the the gamut yeah of
[2803.74 --> 2808.34] tools is there anything you use on a daily or weekly basis that we have not talked about yet
[2808.34 --> 2819.18] well we briefly mentioned ai stuff and i have recently switched my standard usage off chat gpt so
[2819.18 --> 2826.42] i was just a chat gpt for standard usage for the first two years when they come out november two
[2826.42 --> 2833.96] years ago so not quite two years call it 18 months and then llama three just got good enough and so i'm
[2833.96 --> 2841.48] out there beating the drum of like why not use the open ish version versus the purely server side
[2841.48 --> 2847.92] thing and so i've cut back on my chat gpt i still have it on my phone i use it mostly for like create
[2847.92 --> 2853.94] me an image of this thing because it's just so so easy so it's so fun too yes my wife absolutely loves
[2853.94 --> 2858.58] that feature in fact she's on the free plan i'm on the paid plan for chat gpt and so she gets like
[2858.58 --> 2863.20] three a day two or three a day and so she'll prompt like two or three and then she'll be like
[2863.20 --> 2868.44] can you paste this prompt into your phone and send me the picture because i want a fifth sixth
[2868.44 --> 2872.76] and seventh attempt at this image the maddening part about those image generators is they just
[2872.76 --> 2877.14] can't spell right you notice this yep because they're not actually spelling words you cannot
[2877.14 --> 2881.58] put text in there no i mean you can but they're going to spell it wrong and it's going to be weird
[2881.58 --> 2886.66] and you can tell it like nah you spelled that wrong and basically it's like i don't know how to spell
[2886.66 --> 2893.18] i'm just drawing pixels you know yep which is hilarious but needs to be fixed so i use it for that
[2893.18 --> 2900.94] but i've installed olama on my macbook and i'm using a desktop app called enchanted which is
[2900.94 --> 2909.50] basically a chat gpt-esque uh gui for interacting with various lms and you can configure it which one
[2909.50 --> 2916.34] to use via server uri similar to the way you can with zed or vs code or you know vim and so i have it
[2916.34 --> 2925.04] using llama 3.2 and i've been pretty happy with that setup for just you know answer my questions
[2925.04 --> 2930.64] and generate some text and whatever i'll ask it coding questions i'm still not like i haven't
[2930.64 --> 2935.92] figured out using the coding tools inside of the code editor quite as much i'm still because i started
[2935.92 --> 2940.34] off just like i'll go ask jasgpt and i'll come back i did that for a while and that's where i kind
[2940.34 --> 2944.16] of feel like i'm vanilla and i'm probably behind the curve zed has some stuff where you can like
[2944.16 --> 2948.50] highlight a thing and then like send that in his context and stuff but i just haven't gotten that
[2948.50 --> 2953.12] far i'm sure you could probably help me with some of this yeah i mean that's one of the places i found
[2953.12 --> 2958.76] cursor to be really ahead of the curve in a couple of ways one is their sort of ability to let you
[2958.76 --> 2963.20] specify context and you can like add files and add them to your context and things like that
[2963.20 --> 2968.40] but the other thing they have that i think is really nice is they have their own custom diffing model
[2968.40 --> 2975.74] so you get something back from whether it's you know llama or i love sonnet uh as a tool here or
[2975.74 --> 2982.50] you know gpt4 or whatever you know kind of llm model you're using but then applying that to your
[2982.50 --> 2987.60] file correctly is actually not always trivial like they're not always giving you good diffs and so
[2987.60 --> 2992.54] what cursor has is they have their own proprietary model that is like take this thing that comes back
[2992.54 --> 2999.20] from the llm and turn it into an actual pliable diff and that seems to i think make a big difference
[2999.20 --> 3003.86] now i don't know zed may be doing something similar like there's probably it's pretty clear that that's
[3003.86 --> 3008.54] a need and so i would imagine that anyone who's building a business around this is going to be
[3008.54 --> 3014.60] building those tools that is one of the places where the open source variations on this really fall
[3014.60 --> 3019.42] short they are just not nearly as good at applying the changes that come back yeah they're pretty good at
[3019.42 --> 3026.20] just being a chat bot you know and doing what chat gpt basically does but yeah turning it into an
[3026.20 --> 3034.84] overall wrapped product is always been where open source tends to lose i did download cursor it just
[3034.84 --> 3039.70] bugs me that like i understand why they're like we need to just be our own editor like to me as a
[3039.70 --> 3046.36] product person i totally 100 get that and i think if they continue to do what they're doing they'll get
[3046.36 --> 3052.16] it to where it's good enough and it is a vs code fork so it's not like it's completely foreign to
[3052.16 --> 3057.40] anybody but it's just a crappier editor i mean it is and so like i don't want to switch all my things
[3057.40 --> 3062.94] in order to go get that experience however i did have a memory leak in a node app that i built and i
[3062.94 --> 3067.22] didn't put any work into this memory leak i just knew it was a chromium thing like puppeteer
[3067.22 --> 3073.92] eventually chromium is just like just leaking memory and it would crash my app server right and it just
[3073.92 --> 3077.70] started crashing it was crashed like once a week at first and then it started crashing like once every
[3077.70 --> 3082.80] couple of days and running as a server on fly and i would just get sick of the crash reports crashing
[3082.80 --> 3087.44] is no big deal honestly because it just reboots the thing and it comes back up again and anyways i was
[3087.44 --> 3091.94] like i don't really want to deal with this memory issue problem you know i'm not freeing something
[3091.94 --> 3097.48] here or there i'm calling const when i should call letter i don't know what i'm doing i'm realizing
[3097.48 --> 3102.90] that's probably not the problem but i just like gave my file it's like a node it's like server.js
[3102.90 --> 3109.00] right i just took my node app and i just put it in a cursor and i just said i got this memory problem
[3109.00 --> 3114.00] and it's like i'm gonna rewrite this for you and then i was like yeah that looks like it might fix
[3114.00 --> 3119.78] the problem paste it in haven't had an app crash since so super successful with that project but i just
[3119.78 --> 3124.50] didn't like it's not yeah well and i i have a similar problem right so i use the vim bindings
[3124.50 --> 3132.04] because of course and it's slow it's just slow and the undo like the implementation or the
[3132.04 --> 3137.78] interaction of the undo with the ai completes is broken it's totally worked so if i'm editing
[3137.78 --> 3142.02] something where i'm just like i want to go in i know the change i want to make i'm going to make it
[3142.02 --> 3148.74] i will still open up new vim in a terminal that said for larger scale transformations like worth it
[3148.74 --> 3154.00] it's totally worth it and you can do like multi-file transformations you can do a single file
[3154.00 --> 3158.28] we're like refactor this to do this i've done similar things where i'm like this is broken it
[3158.28 --> 3163.70] does this fix it and it'll just yeah do it and and you look at the code you're like actually this is
[3163.70 --> 3167.70] pretty much had i known that i would have wrote this you know yes and it doesn't even know me
[3167.70 --> 3171.48] it's just writing code and i'm like yeah i'll just go ahead and accept that and the thing is
[3171.48 --> 3176.46] you can't turn your brain off with it right like because it will still do things wrong or it'll
[3176.46 --> 3180.56] misinterpret or whatever but it's like as if you had somebody you you know a really junior
[3180.56 --> 3184.46] developer you could delegate to you're like go write this code for me no that's not right no
[3184.46 --> 3190.94] that's not right change this do it this way okay great go but it's so fast yeah they're getting
[3190.94 --> 3196.18] there it's gonna get there it's just not knowing exactly like what it looks like in the meantime
[3196.18 --> 3200.84] and what to what's worth your while and what's not yeah and so i'm just kind of like still
[3200.84 --> 3206.06] wading into the deep end i'm not like deep into it they've also got the scaling issues right so like
[3206.06 --> 3210.88] cursor has had their hacker news moment or their blow up online moment or whatever and like
[3210.88 --> 3216.78] because they do things like route stuff through their own router so that they you know are proxying
[3216.78 --> 3220.80] for you and they have their their diffing models like they're running their own models and stuff like
[3220.80 --> 3228.14] traffic goes through their servers and they're a small team and so when like they get hit with these
[3228.14 --> 3234.50] massive surges of traffic like they don't always handle it well sometimes it it's like moving through
[3234.50 --> 3241.36] molasses well that's just a time and money problem you know money and time will fix that one eventually
[3241.36 --> 3248.74] they will solve that but yes it is it is interesting once you once i've like made my brain shift to okay
[3248.74 --> 3253.46] i can do this work and it's it's interesting going into like i went into a legacy project the other day
[3253.46 --> 3259.42] i was like i don't remember how any of this stuff is working and it was so convenient to be able to
[3259.42 --> 3266.04] just load it up in cursor ask the chat like what is this doing okay change it to do this thing and have
[3266.04 --> 3273.08] it just work i'm a believer now i'm a true believer this stuff is you know it's imperfect it's broken it's
[3273.08 --> 3278.62] not intelligent per intelligence right it's probably the biggest breakthrough in terms of coding
[3278.62 --> 3284.22] productivity i've seen in my lifestyle lifespan i think that's fair i was trying to think back is there
[3284.22 --> 3290.40] other than other major breakthroughs but i mean i guess there's the long slow rise of open source
[3290.40 --> 3295.94] stuff and api availability right like so it's not new that we're making things easier for software
[3295.94 --> 3302.68] developers because like you want to start a new service now you can integrate with every other
[3302.68 --> 3308.84] service out there very quickly just throw up open source you know it's shockingly fast to get a very
[3308.84 --> 3314.98] powerful application log in with your google account send email do this do that do that like
[3314.98 --> 3321.54] you could set that all up in a few days and things that would have taken weeks and weeks of custom code
[3321.54 --> 3327.38] back in the day so right there has been an accumulation of improvements over time right but like in terms of
[3327.38 --> 3332.78] a single step function change it's shocking yeah i was thinking like this like the first programming
[3332.78 --> 3337.70] language that was above assembly or something but even those were probably smaller incremental changes
[3337.70 --> 3344.02] that eventually became big in terms of productivity compared to this which seems to be although there
[3344.02 --> 3350.14] are steps along the way but at the end of this the productivity boost is going to be just astronomical
[3350.14 --> 3357.14] across the world for sure for sure well i hope it leads to honestly more businesses that are not
[3357.14 --> 3362.64] software businesses able to do what you did of like hey you know we're building our own tools for this
[3362.64 --> 3366.64] because the tools out there aren't there yet or they're not good or they don't handle our niche use case
[3366.64 --> 3371.92] right like that's i think the really cool thing about these tools is they enable a swath of software
[3371.92 --> 3376.70] development that probably wouldn't have happened before because it wasn't economical yeah well said
[3376.70 --> 3383.64] i hope so i think that will probably happen and if it's anything like past innovations it will produce
[3383.64 --> 3389.66] not less work but more work just at a different layer of the stack yeah and it'll take time to get
[3389.66 --> 3396.38] there right it takes time to adjust but i think we will anyway that's a good tool closing right
[3396.38 --> 3401.70] oh yeah you bring your tools in you had your toolbox but this new tool it's worth uh pulling
[3401.70 --> 3407.40] into that toolbox 100 at the end of the day you know if you're trying to build something the toolbox
[3407.40 --> 3412.04] is just a means to the end you know like i would happily throw it out if i could get the end result
[3412.04 --> 3417.26] without it wouldn't you k-ball mostly i like a few of my tools maybe there's a little bit of joy in
[3417.26 --> 3424.18] there sure maybe i would hesitantly throw things out but i think overall i'd be willing to part with
[3424.18 --> 3429.68] pretty much all of my tools in order to get the end result faster cheaper and you know without cutting
[3429.68 --> 3436.58] my finger absolutely all right well let's call that a day thank you jared thanks man thanks for uh
[3436.58 --> 3443.46] dissecting my tools and call me out on the front end man everyone's gonna think i'm a imposter now thanks
[3443.46 --> 3452.76] if the shoe fits i was gonna say no lies detected like no that's just true it's how i feel you know
[3452.76 --> 3454.84] it's all good so it is what it is
[3454.84 --> 3476.08] all right that is jsparty for this week thanks for hanging with us hopefully you find at least one
[3476.08 --> 3481.64] of my tools useful in your work if not go back and listen to the episode called digging through
[3481.64 --> 3488.28] nick neesey's toolbox lots of gems in there and if you haven't yet check out changelog news it's the
[3488.28 --> 3494.94] industry's only weekly newsletter that's also a podcast one reader calls it so good he considers it
[3494.94 --> 3502.72] a competitive advantage read and listen to the latest issue at changelog.com slash news big thanks
[3502.72 --> 3509.50] once again to our partners at fly.io to our beat freak the mysterious breakmaster cylinder and to our
[3509.50 --> 3516.66] longtime sponsors at sentry we love sentry you might too use code changelog save 100 bucks on the team
[3516.66 --> 3523.76] plan next up on the pod tanner lindsley and his tan stack stay tuned right here we'll have that episode
[3523.76 --> 3525.34] ready for you next week
[3525.34 --> 3535.34] you