Databased: Why Most Content Doesn't Work ft. Web Dev Cody

[Music] hi everybody welcome back to another episode of databased we have a very special episode today so webdev Cody is joining us how you doing Cody doing good thanks for having me on with you yeah I'm looking forward to talking to you too so I have um not only in sort of following your content and chatting with you a lot over the last uh 18 months just gotten to feel like I got to know a lot more about um what you do and appreciate the work you do but part of what what I've been spending more time doing is figuring out how to do our own version of like making content and and and talking about technical products even doing some like education and stuff like that so I definitely have a much healthier appreciation at this point for sort of like how sort of deep that that job is and that skill set is and stuff like that so yeah I'm pretty excited about this conversation today the funny thing is like as much times you spend on camera online right my my default assumption is like well you must be so comfortable right now but you said you told me before we started you're like this is the first podcast I've ever done yeah I mean I have over a thousand videos on my channel and even this like a one-on-one I'm a little anxious you know I don't know what it is but you just mix it up a little bit so the internet is a is a weird place and like the kind of like uh what what having a video where I think you're up to you're above 200,000 subscribers now right you're is that right yeah I'm close to two 250 yeah 250,000 subscribers so with that kind of thing I have to tell this story really quick there was like a we had CH chatted with each other online for like a year and then and I watched all your videos so I had seen you a lot and i' still remember we had this one meeting where we jumped on and I'm like okay so Cody I'm thinking that maybe it'd be fun if we talked about and then you kind of were like yeah by the way hi nice to meet you I stopped and I realized it was this like surreal internet thing where like I actually felt like we knew each other because I had seen you talk all the time and you were like this is kind of the first time like I've ever like talk to you like on camera so you have this like celebrity experience thing right like someone feels like they know you and in some some respects maybe they kind of do but like you're like I've I've never met you before so yeah I don't know how much uh it feels strange to you that you have a quarter of a million people that are subscribed to your channel and so many people are kind of watching what you do it's kind of weird to get used to I mean like I just feel like a normal person just making like tutorials but recently I was walking through the store and some random person stopped me he's like hey do you like do YouTube videos and I'm like Yep this is it this is my moment this is it yeah but yeah going to your um what you mentioned before I mean I I remember watching through one of your like three-hour tutorials where you built a story book generator with conx and Ai and like even that alone you kind of feel like you get to know someone just by watching them teach for three hours especially teaching if it's just them doing like educational or entertainment content you don't really get to know them but teaching it's like you really get a Vibe for like who they are I think well the thing yeah I mean we'll talk a lot more about all the stuff but like I I look at the videos I do and all I see is all the things I want to do better like especially those old videos cuz like I think that you know the people that have really sort of understand kind of like and have put the practice in to be really natural on camera you know there's a there's a version of being really succinct that kind of like almost feels very natural like you get into flow relatively easily and I look at my old videos there's so many ums and AZ and then I'm showing things that aren't the important part of the project and stuff like that and I think there's a way that like somehow like within let's say videos that you do you sort of like manage to kind of focus on the like Salient parts of the project that are probably the ones that like matter to the person and there's not as much like transition time clicking around files trying to be like where was that line of code that so slowly over time like trying to balance like the the difference between like polish and like organic right like I do want to be myself but also too you want to like respect the viewer and that like they they don't necessarily want to sit there for 60 seconds and watch you search around for like what line of the file you know that but I was kind of curious like just because you you have so many videos as you mentioned like how did you kind of get started this like what was the what was your path to becoming a technical Creator yeah I think originally I did it just cuz I was like passionate about coding and like when I started learning how to code this was back in like 2009 so there's no Discord there's no slack there's no communities you can join it's just like forums you can find online and you don't feel like you actually like jump into a zoom and talk to somebody and so you kind of feel like back then you're learning alone and now with YouTube and like all these other social media platforms you have an opportunity to inspire others and teach others so that's like why why I started it I just like teaching I don't know it's just fun and then like you get a comment saying like bro you just completely explain that better than like 20 other videos I watched like that just keeps you going because you know you're helping people and I I will say like the reasonings for keeping on with the channel has kind of like ebbed and flowed over the the months like sometimes I want to do it because I'm chasing views and then I like course correct myself I'm like no I need to stay on helping people so it's been like a a Juggle sometimes a struggle but that's the original reason I got into it yeah I mean you mentioned kind of the EB and flow and like I think that like one of the things I've noticed about you is that like both on like let's say on Twitter X or whatever we're calling it now but also like your your videos and like you tend to like follow your nose a lot right so like if you're having a moment where you're like a little fed up with webtech like a little while ago you went and did some goling for a while and then you tried PHP and so so I kind of Wonder like yeah how you think about balancing that like my audience knows me for like react nextjs and like but you know I'm kind of tired of that I need you know what's I'm passionate about right now is just trying this new thing so like how do you think about balancing those two things and like do you worry about like your audience coming with you or is it not really what motivates you I mean that's definitely a major factor of why I keep doing nextjs stuff is like I've built a whole Niche around nextjs in a react and it would be dumb for me like I kind of treat my channel as a business so it' be dumb for me to switch over to something like laravel and just keep pushing out laravel content because I've seen a bunch of other channels where they try to switch to a different stack or framework and their videos just bomb so like I'm going to sit here after work in like record a video and research and edit I want to get some views and so that's kind of like why I stick with nextjs but also I do think that react and xjs is like the largest ecosystem at least on social media it makes it seem like it's the largest ecosystem it gets the most views but going with the follow your nose comment at some point I do get fed up I mean I'm doing web development fulltime 8 hours a day and then after work I'm like okay let's make another web development video where I teach a nextjs tutorial for the 12th time that I've done it from scratch and it's like I got sometimes I gota like you know get that uh that Hiatus and just like get a a clear fresh view on coding um and recently I've been doing Game Dev I'll probably burn out after like a a month or two of that but it's been fun like I got back into object Orient programming and like trying to get back to like the gang of four design patterns like you don't seem like you do that as much with web development because it's like build a page build a component connect to a rest API and it's very like straightforward versus something that's more like a native desktop type of thing I feel like you have to think about your code AS components and I'm sure you saw that with your rust engine that you guys built like it's a completely different type of programming than just making rest apis and a UI to integrate with that yeah so the getting a four this is back to visitor pattern and all these kinds of like uh classic objectoriented software architecture stuff so yeah what what what uh Tech is the game I I've seen you working on this game but what what actual um stack is that what language is it and stuff like that right now it's just typescript I'm doing typescript and then the back is just expressjs for like the websocket server I'm assuming at some point I'll need achievements and storing like authentication I was thinking about bring in convicts for that I'm not going to lie just because it's so easy and then um I was thinking about there's like another service called haora which basically you can deploy your game server and they will scale up rooms automatically for you I I did that on a previous sponsor video which after using it I'm like this is this is pretty cool for a game yeah just typescript I'm trying to St stay Court to typescript on that one but it's still kind of more object oriented patterns within the typescript yeah the the funny thing is you mentioning like writing the game and you you all are rapid fire events firing over the socket and stuff like this so before that you were working on a game on convex so we we ran into 18 months ago we had a game called AI town that like kind of had had its moments where in fact it kind of keeps rearing its head every now and then but it was kind of wedging something that wasn't convex wasn't really designed for on top of convex because you thought well I wonder if we could kind of make like a a game that has it it it's this isn't like 30fps or anything like that but it's relatively latency sensitive kind of thing and um it was not easy to make it work to be honest and so it one thing I would say when we looked back on it we were like one thing was really good about it is like we learned more about Ai and about simulation stuff like that and it also too by running something on convex that convex really wasn't designed for we almost like hardened convex for like what it was designed for it's like it's like if you can make it not completely break for something that really tests it a lot more than like it but but at the end of the day then we'd have people that wanted to make games show up and say oh I want to make some high performance like latency sensive thing and be like yeah it's conx isn't really great for that so there's these other services that are like tailored Taylor Made for that that can do that stuff uh much easier I will say someone on my Discord has been making a game I think he's still using conx but it's a turnbas game I think comic shines the turnbas game turnbas games are good yep so but something where it's like oh there's going to be a collision between these two people and something has to happen at that moment where like even if it's not again like UDP packet 18 milliseconds FPS type latencies but it's still like 100 milliseconds or 150 you can do it but like having a database in middle sometimes like is that really quite the right abstraction so but yeah for turn-based games it works it works really well that one's pretty popular and yeah you mention just like if you're going to come home and like you know kind of get out the editor again after a full-time job I kind of wonder how you how do you balance all that right because like you have a full-time job and you have this thing that probably started as like a hobby and has almost like grown into like a second job now right so like how do you think about not going crazy with that much of a workload I mean if you go on MX that's like my psychosis that comes out you know that's when I go crazy I was going to say you do seem like you're skirting The Edge every now and then I'm trying to like readdress how I'm dealing with X unfortunately like certain social media platforms have meta to it right there's a certain algorithm you have to hit YouTube's different from Tik Tok Tik tok's different from X X unfortunately is just very like if you post something negative you're going to get a lot of views and I've been trying to like okay let's just stop being negative be positive and see what happens so I'm doing that for this year but as terms of like where do I get the time yeah I do the eight hour work day and a lot of the things that I'm learning at work I mean like if you spend eight hours at work learning something you should be able to make a video about it and so that's what I do I sit down and I just start recording um I've gotten really good at my editing so like I will literally sit down on my computer with no idea maybe I'll spend a minute thinking about like what I want to make and then I'll start recording I'll start talking and then afterwards I'll go downstairs hang out with my family and just quickly edit it and post it and I've gotten like I've tuned that down really well so I can crank out videos when I get really motivated but the research stuff like you still have to research like when someone asked me to do a sponsor video I got to read through their docs understand how their stuff works I got to Tinker with it to make sure I'm actually confident with teaching it that I just do sitting in bed like I'll be watching murder shows with my wife or my kids will be watching or playing Roblox and I'll just read the docs use cursor sometimes to just you know write a sentence and then I'll go back and watch TV for a minute so I I found ways to like take as many shortcuts as I can like you've seen Myan Channel with the thumbnails I don't do thumbnails at one point I did do the custom thumbnails and when I stopped I noticed a boost in My Views so like this is like two years ago and I saw that boost I'm like you know what what's the point of putting this extra 10 minutes in making thumbnails when I could just let YouTube pick them for me and it it worked good so I think that's kind of how I juggle it all I just take shortcuts I just do the bare minimum so I can stay motivated because if I have to do all this stuff for every single video like I probably would have burned out years ago I mean that's one thing we've learned is like the and we don't have our workflow down as well as you do but we're getting a little bit better but like certainly the the time off camera can easily be multiples of the time on camera if you're not sort of like very careful about how you structure your work and stuff like that it sounds like so you know when it comes to editing and kind of managing your Channel all that you're kind of still doing all that stuff yourself so you just like do your editing and like and so yeah I would imagine that means that it's very important to you to have a very efficient process so you're not spending you know you know two two hours recording and then six hours into the night like cutting everything up into a million pieces and gluing it back together except for those long videos like you've seen those six- hour videos I've done on my channel like some of those conicts those I mean you have to watch through the whole thing and spend six hours watching it sometimes I watch through at two times speed and so I'm used to hearing my voice like chipmunk mode just going straight through the video and I'll cut out the places where my guy screw this up yeah also if you're doing like a tutorial never stop the recording just take a second and pause and then just re explain it like delete code if you have to and rewrite it and re explain it in a better way that I do that a lot and I think that's kind of helped my content be easier to digest for beginners is because I will say a whole sentence and code something up and I'll think this was a bad way to explain it and I'll delete it all and just like re explain it sometimes but you don't see that because I edit everything one thing I'm trying to do is like for this year I've been watching the Prime in and Theo and it's like I'm kind of inspired with how they can sit there on a live stream and talk for hour straight yeah because my content is very different I edit everything because I make mess ups I say um uh and then I I see a bug and I have to fix that for 10 minutes so I'm trying to Pivot and do more live streaming so I can get better at just speaking I I think if you can speak without having like the hiccups and the ums and whatever you just come across as more confident and believable yeah I'm Blown Away by like when I'll watch Theo on Twitch he'll just be sitting there like H chat blah blah blah and like all right here we go and he'll just stare at the camera he like squares up the camera just like some people say react 19 is and I'm just how do you I think he does have a teleprompter though so I'm not sure like if he maybe before he starts getting into a video he already has something typed up I haven't really talked much about his his workflow I just know he has like a studio and a teleprompter and he has a team editing his stuff and researching his stuff I believe um but the primagen I mean he just he's just funny to watch on stream and he like he has it down packed where he can keep you entertained and engaged and then he actually teaches something and then goes off on tangent so you're still engaged so it's just that type of stuff that you have to learn in terms of like your social media platform you want to engage with YouTube is something different than Tik Tok but uh the thing you have to learn though I I mean tell me if you're read with this right but like is is like what's the version of you that's going to work right because like if you kind of try to do someone else like it doesn't I think it doesn't really work that well right so it's like you know I I just feel like the thing that we've discovered and other other people I talked to are I think they really good creators will talk about is like in some respects you can kind of only do you if you're going to do it credibly and do it at volume and all that kind of stuff and so you're just trying to kind of let like find the kind of version of yourself that sort of like projects confidently and I don't know if like that kind of makes sense to you and like how like how you think of like the version of Cody that like is on like is this just you is it you know I would have my own like adjectives right like you can kind of have a dry sense of humor sometimes right and you're very you're very like authentic right so like if you're having a bad day you kind of say like oh this is driving me crazy today or whatever so how much do you kind of think about the way you present yourself or do you just actually try not to think about it just be yourself I think I just try to be myself like I'm pretty sarcastic I'm pretty uh dry humor I would say I do try to get that into videos if I can like if I'm coding and about to say something and I can throw in some funny thing like that'll just keep people in engaged but I think just having it come naturally which again you're not going to know how to do that unless you practice speaking and like making an effort of like let's just throw in two jokes in this video or something like that especially at the beginning you want to have a hook in your videos and sometimes it's okay to plan that out and like okay let's something funny I could say or like let's make fun of JavaScript because why not let's make fun of laravel just because it's just like the meta that's going on at the time not that I actually like think that Lille or anything's bad it's just you know it's just thinking about something you could say to keep people engaged but yeah I think to answer your question I think most of my content is pretty authentic other than I have to edit stuff when I'm like doing more of like a educational teaching video I guess my suspicion too and something like I've tried to work at being better at is just the Comfort to be yourself it's not easy actually I think that like I mean it might be easy for you now you've had a lot of practice at it but like I think it's actually pretty difficult actually right to like stare on camera and somehow find a way to like I don't know very vulnerably be like the same way you are with your friends or in person or whatever like you know it's it's taken a little work at least for me to like try to be as comfortable with you know um when I'm making content but like I think when I do it's more effective like the more I can find a way to just like be willing to be myself it seems to work better so it sounds like you agree with that in general yeah I mean and when I watch your videos I mean it's really good educational content but if you feel like you're not truly being yourself I think just funny ways to do that I mean you you're literally like a CEO of a company right so you have to speak a lot to a lot of people I assume you've been to conferences and I've seen you speak out a conference which I've never done so that has to be anxiety producing I would say when you're on camera you feel like people are judging you and watching you and you're like oh I have to be careful what I say and at some point you just have to be like you know what screw this I'm just GNA say what I want to say if if I'm wrong they're going to correct me in the comments and I will do better next time well obviously you don't want to say something that's going to get you canceled you know what I'm saying like you have to have you got to be able to gauge what you're saying and not just be uh off the rails the funny thing is yeah I get what you mean but like for me that's what that practice thing like you you kind of coming in saying oh I haven't done a podcast before I'm a little nervous about this or whatever like talking to people at this point feels pretty comfortable talking to a crowd for me is even more comfortable talking to a camera is almost the scariest thing you know it's so because the lens is it's almost like it's infinitely big right like in practice our channel is not nearly as popular as your channel so it's actually smaller than the crowd I might talk to but the camera feels much more intimidating to me for some reason you know so I'm still working on kind of getting comfortable with that but one thing you'll find out with YouTube is that videos are ephemeral like you'll put it out there and within two days if you didn't get the views you're looking for you're not going to like there's a you might hit the YouTube algorithm and win the lottery and like your your video just get you know shot up to the Moon with views but all my videos I publish like with that after like two to three days that's it if you didn't get the 10,000 20,000 views you're looking for you're not going to get it and so from my persp perspective is like put it out there and just move on to the next video after you've seen the performance of that one you mentioned before one of the sort of Preparatory like tasks you have to do sometimes if you're like working with a new sponsor or something so we obviously first met each other through sponsor content and stuff like that I'm kind of curious these days how you think about like the relationship between like creators and sponsors right because I think especially in the technology Arena it it's a it's a uh complicated relationship sometimes where like you know it's very important to maintain your kind of authenticity with your audience and and have your true voice like credibly on camera but also too it could be a little complicated it feels like oh is this person saying they like this product just because they're being paid or whatever so like how do you think about like choosing sponsors these days and like balancing that like like making sure you're not diluting your voice by like promoting things that you're not excited about or whatever I try to limit how many sponsor videos I do and I only do like the long form videos so you have to have a product where I can make at least two hour 2our three hour video where I can truly give my viewers a understanding of what your product's doing I I don't ever do the like you know the 30 60c integrated Clips I just don't think those are useful and as me as a viewer of YouTube I literally skip those like when I'm watching Theo's Channel and I see the sponsor thing come up I click the right arrow and I skip it because no one wants to watch it at least I don't want to watch it but a full dedicated video If a sponsor comes to me and says hey I need you to actually like demo my product and build something cool that's when I'll be like yeah I'll work with you all and I'll do that because that's actually enjoyable to do it just takes a lot more work obviously you have to do the research and the planning and the editing and the recording and all that stuff but part of why you like the longer engagements is if you are going to do that kind of content it's it's better to create sort of a a genuine representation of what the product is which could take longer than just like kind of a equip like a standard ad or something where but that but that also means as you said you're just not doing as much the drama with Theo and like his sponsors and like his comments bro people on YouTube are just so I don't know why they hate sponsors so much it's like I don't know if they pick on him just because he's Theo but I I've never gotten that much hate for doing like sponsorships for example with conx I've actually gotten a lot of positive comments they're like bro thank you so much for hooking me on comic this is actually like one of the coolest service I've used and I haven't gotten that from a lot of the other sponsor videos I've done it's just been convex so good job on building convex oh thank you yeah it's it feels like a sensitive topic but we're here to be honest right so like I actually really don't like what happened with with like the versel and Theo thing like my take on it is that was a symbiotic relationship that made a lot of sense Theo really liked versel and and continues to like versel and so the that's the right kind of sponsorship to have right rather than the token sponsorship that's like oh I'll read something because someone paid me or whatever and so it feels like it's a shame it went that way where um I I think it it felt like it was kind of related to versell had that blowback for a little while and couple people's bills ended up being high and it was but I I don't know the entire environment around it but it does feel like that was maybe not necessarily the best outcome where like rather than a Creator being likey look I'm going to pick like six companies but these are the ones I really use right um and kind of turning more into a standard like okay if that is not going to be viewed as authentic then we're going to go to ad reads or whatever so and I say there someone that would have loved for if Theo to had a relationship with our company when there were only six companies and we weren't one of those companies but I still really respected that like I respected being shut out because we weren't one of the companies he was using and that was what the bar was you know so I'm not sure we you know I'm not sure that was the right outcome for his audience but here we are so I mean I don't know why but developers hate being like marketed too it's like they they hate being shown solutions that make their lives easier for some reason me I'm like bro I just want to I just want to get something done quick so if you if I do a sponsor video with a product I'm like this actually is gamechanging I'm going to use it one thing I had on this topic for sponsors like if you are a company and you want to reach out to a YouTuber my biggest advice is don't be micromanaging I've worked with probably like 20 plus companies over the past couple years some of them are great some of them are like hey we trust you make the video publish it and we hope it does good other companies will ask for like edits after edits after edits you're be like oh I said this one sentence at Tim stamp 543 and like can you please change that and re re-record that if you're that type of micromanager like I would rather just not even work with you like I don't need the money have a full-time job right so I want to work with people who have a good product and also just trust the process and if you don't like the views that I get on my channel like you have a responsibility and go see like what's my past history if I'm only getting 5,000 views on my videos why did you reach out to me you know you can tell me if I'm full of [ __ ] when I say this right but like I don't think we've ever really asked you to change anything for the most part and and that's pretty consistent with the people we work with here's here's why I believe in that we're talking about all this Creator stuff right is that like you understand your audience and your voice way better than we do and hopefully you know you're only working with us if like you have actually paid attention to the product and so have some understanding of it so I think imagining that the right outcome is somehow the the like the the company changing the way you would talk about out the product they made I think it's very seldom the right idea like the whole point is to be a part of that place that happens there that's already established between like the Creator and Their audience so as opposed to somehow trying to like correct you know like I I think it's very seldom I would ever say that like uh like I would feel comfortable saying no no no you said that the wrong way right there's some little things are just like oh hey that one thing like it's actually easier if you do this instead of that but the truth is it's also Fair that like let's say with convex like our user would be confused at that part too and probably do it the way you did and so I also think even though that's more of like it like a tip for next time if you use convex again I also don't think that's like stop the video right because it's very fair that oh yeah you probably would solve the problem that way because like we didn't make it clear enough or whatever so so yeah I think for the most part the best my the best call in my opinion is like if if you if you really are saying I I like what's happening on this channel and you've really assessed it deep that deeply and I I like the way this person works just kind of let them work and it should be extremely exceptional to like try to control the way that they're presenting you at least that's my but that's my take definitely I mean the Creator knows their content Their audience and not only that is like I think a lot of companies look for views if you get a thousand views but all your views like 50% of the people who like truly are true fans of your content and they go and check out the service that you just promoted that's probably worth a lot more than 100,000 views and it's just people who click the way you know like you can make a nice thumbnail and have a nice intro but if they're not actually going to go to your service and sign up the views don't matter I've had one company say that I was I did like a really really great job with like the conversions for I don't know why like their product but like I looked at my views I'm like these are just like basic average views so I guess chasing the views isn't it can be good but I think it's just getting someone who has a a strong connection with their audience and can actually promote something is probably better to look for you you mentioned the thing about marketing skepticism this is obviously something we ran into because like a lot of startups we're just a bunch of Engineers I mean at the end of the day we're like in this whole world that we're in as much as we are like hypothetically perpetrators now we are also like victims to the same degree right so just because I've been in the industry for a long time it just became very clear to everyone I don't know 10 or 12 years ago or something that software Engineers are a really valuable customer right like software Engineers have like a lot of purchasing power and all this kind of stuff and I think because of that there's just been a ton of products launched at in the software space in the last 15 years and I do think that like a lot of software developers do feel like it's pretty saturated in some ways with like so many new things and so it can be kind of frustrating and the skepticism can be pretty high the hope would be that that is actually the if there's a glass half full version of it it's like that can be one of the like Virtues Of following you know creators you trust on YouTube is because it can act a kind of validating mechanism for like among all the crap that's out there that's you know just buying Google ads or whatever like these these new things are actually pretty interesting you know so um but we'll see it's definitely it's definitely uh it's something that's in the water and what's very funny about us is like again a group of Engineers and start of a company it's on one side we're kind of trying to say like no actually convex actually is pretty good check it out but on the other side we're we're just as much like oh my gosh like stop telling me about it's definitely hard I mean I don't run a company or anything but when I try my little side projects it's like I could try to Market those on my own channel but my audience doesn't care about these little side projects I'm working on so marketing is from I've learned on this YouTube channel and just like over time is like marketing is almost like half of the equation you build a great product it's it's awesome but you have to get people in the door and then get them to try it and then they're going to get hooked if it's actually a great product and that's or like even if you had to spend 30 minutes to try a product Engineers are so busy or reluctant to even try it and it's like holding their hand and getting them to try that first time is like the hardest part in my opinion what I mean what the hard thing you've seen with marketing um convex oh it's incredibly difficult yeah it's exactly what you described right it's uh just getting someone to try it I mean some of it don't get me wrong some of it's like our fault I think there's some things we've done wrong like we haven't ever had a good you know this we've never had a good way to describe convex and not great like convex is a you know we're actually working at fixing this right now but we made this like new kind of thing we probably should have just said it's it's like a Firebase we just never never really did and so it's like is it a database we'd say well it's a little bit more than a database is it a backend the service yeah but it's kind of a newer one it's a little different so I think that hurt us but even if we had had a really clear like quote unquote category it's still really really difficult just especially if you have something new because someone has to be willing to learn a little bit right now like to try something new and and yeah that that can be a challenge for sure but yeah the marketing part is is definitely hard I think it's part of why like one of the things we have happened because like people that do use convex for for us I would say it's usually like the commit point is like half an hour to a day and they're like oh okay this is pretty and I think that that's why like sometimes if somebody comes up and says like oh we got to get more people using this this is better than people know or whatever honestly the we always say is like your voice matters so much more than ours like word of mouth is everything with develop developers I think right and and um that's what we've heard from other project leaders and Founders We Trust is like the thing you're trying to do with all your like distribution marketing I guess is just build enough of a head ofest steam That Word of Mouth can take over because I think like every time convex gets a big surge in like on a given day if we get a big Surge and new signups a lot of times just because a bunch of people say cool stuff about us on Twitter it's not because there was some single brilliant Insight from a distribution perspective it's because a developer did something really neat and is like hey this thing's pretty cool so I think that the word of mouth thing is essential yeah I mean like I remember when I was going through X and uh who who does tan stack Tanner is it Tanner Tanner yeah Tanner made up quotes about com I'm pretty sure you guys got a lot of traffic from that one because like you have yeah yeah it's just stuff like that yeah yeah because when you have people who are like considered domain level experts in web development and they say something about your product like people listen a lot of us don't have the time to try other products and even I'll even be honest like when I did that first sponsor video with convex it didn't click right away I'm like okay this is just another product and then I started like building with it I'm like okay and then you guys started adding like the file storage in the vector searching with the AI stuff I'm like okay this is like a no-brainer at this point like there's just so many things that are included out of the box which is why I used it for like project planner Ai and all these other tools because I'm like dude I can move so much faster using it yeah but no marketing is is is hard and I think it's it's newness and differentness is one of the big challenges but but yeah lots of we're not the only company that's uh fighting against that and trying to figure out how to cut cut through the noise I'm kind of curious like to hear a little bit from you on like AI right so AI Coen for example is starting to be obviously a really big part of people's jobs and for you you're kind of an educational role I I could see it impacting you in a few ways right so one is that like some of the folks that are learning to program now like are are using AI to to code and so that might change the way you think about teaching people because they're going to be tool assisted or whatever I could also see it I think you even did like a poll with your I think you did a poll with your audience recently where it's like is it distracting or not if I'm using you know cursor or something well I'm coding right so I how do you think about the like AI as a coding assistant things like you know um co-pilot and cursor and wind Surf and these kinds of things and how it the kind of impact it's happening on on creators you're doing educational content yeah this one's a tough one I feel like last year I kind of had like an existential crisis when I started using cursor and like a legit just would ask it to do something and it modified like 10 files and it just worked like one shot I opened up my web browser and everything just worked and I'm like what's the point of my Channel now you know it's like I'm sitting here and I could spend two hours building the same thing that cursor just cranked out in two minutes but I guess I just have to remind myself that like I'm coming from a perspective that I know how to code I've been doing it for a while and so like when I see what cursor generates I can read through it and but yeah yeah yeah it seems good a beginner they have no idea what this code's doing they just accept it blindly and then they run into bugs and they copy and paste those bugs back in the cursor and they just keep doing it and they can end up spending hours trying to just get cursor to like fix itself when an experienced developer can fix it in 5 minutes you know so I'm like trying to balance that and I'm trying to figure out like when I make tutorials how much of autocomplete do I do how much do I slow down and explain the things that I do accept and I'm still struggling with that to be honest with you because it's like from my perspective coding with AI if you're not coding with AI it's just like you're kind of I don't want say doing yourself a disservice I find much a lot of value with AI and like productivity has gone through the roof especially when I'm doing like infrastructure code with terraform like I just say give me an S3 bucket and it spit spits out like five different resources automatically sets up the permissions I'm like I don't have to go to the docs anymore like it just kind of cuts out that whole middleman of going to the docs reading the docs and when the AI gets it right which most of the times it kind of does now it's like you just basically sped up your whole process so yeah I don't know I guess I'll just keep on playing around with it I've been building this game at at night I'll just like promp and say hey just do this refactor this and um so it's yeah I don't know do you use it a lot at convex like when you're building at your rust engine or anything else like do you find it really useful or is your system so complex that it reaches that point where it it causes more bugs than I would say the deeper you are in the stack the less you can blindly rely on it but even deep in the stack using it is a kind of like superpowered auto complete is still there's still a lot of value in it like the team ran into to a um a pretty obscure like rust bug the other day like and it was one where it was like just I guess it's an implicit criticism of rust even though I've been building projects on Rust for 10 years but is that the team ran into something and really good Engineers on our team and they did a you know all rust experts on Deck kind of a thing and we just basically had to get everybody that knew rust really well on our team which there's people that know rust really really well and it still they had to stare at it for a long time and then eventually copied and pasted like the section de Claud or whatever and asked for some hints and or maybe uh gp01 or GPT 401 or something like that and this really great description came out of probably what the problem was and like the solution wasn't quite right but like it actually identified what the problem was and we had these four really Top Flight Engineers that had been programmer rust for years all it was one of these things with concurrency and pinning and one of these things is kind of a disaster in Rust anyway but um but so yeah I mean honestly we use it we're using it more and more all the time and like convex users use it a lot and so but certainly when we're working on like our Web projects and we're using convex we use it a ton because it boy is it incredibly productive at that layer I think it's pretty much everyone is using it very heavily when we're building like applications on top of convex these days which you know a number of us spend a lot of our time doing so yeah it's I think I think your the point you raised is we think about that too right it's like you kind of look at it and say you know I just got done in two minutes what would it take what would take me two hours but also too that's a little bit predicated on the fact that you kind of already know how to validate all that stuff because of like all the bad code you've written in your in your past and like all the architectural decisions you've done and like and so you kind of put your reps in so that you're able to validate like uh this stuff so that when it lands if it doesn't quite work you know how to adjust it or whatever and I think that it's one of the big questions we have actually the team was talking about the other day is like how the educational process will work in the future to be able to let's say take Junior Engineers today and how how do they build the fluency they need to be expert curators with these tools you know instead of maybe just blindly accepting what they're giving them and I it feels a little un uncertain and unknown right now how that's going to go yeah I will say like it's good hearing from you all cuz I mean I look up to you all as like really intelligent Engineers like the the amount of work you done at Dropbox like I don't know anyone else mess with how much data was that like pedabytes exabytes of data yeah it was ex it was exabytes of data a lot of data I've never worked on a system that large like I I would say that I've kind of worked on small systems with like not that many users but hearing that from you that you all find Value when you're coding out like lower level rust code I think it's good because there's a lot of people online like if you scroll through X people are like oh it just it just generates slop and I'm like I don't understand how you all coming this conclusion are you guys actually using this or just like jumping on the bashing bandwagon cuz I think one thing that you have to figure out when you're engineer is like when do you reach out for help because I have seen people like get stuck on a bug and they stare at their screen for two hours they don't even go to Google they don't go to stack Overflow they just keep on trying stuff and I'm like have you even googled this error and then they do it and then the GitHub issue comes up and they fix it in two seconds and like getting used to knowing when it's better to just reach out to Ai and like hey just throw some ideas at me maybe one of those ideas you throw my way will like get me to solve it really quick at work we do a ton of pair programming and assemble programming so we'll have like two developers at work just working together and you'll be amazed how fast you'll get unstuck with bugs because someone else like will catch those bugs when you write them and vice versa and now it's like you have cursor that could potentially just catch the things as you're writing them as long as you're like diligent of like reviewing everything you accept speaking of like videos and AI I have this like it's it's queued up I was actually maybe get record cut of it today but like I was trying to summarize a few like a few takeaways on like a recent project I did with using cursor and like one of the things I found is that I actually I tweeted about this I'm really bad at Twitter but I tried anyway so I said I think there's a temptation to say Okay I want to add something to my project and so if you let the llm loose on it you do it in a way that's not carefully structured sometimes what the LM will do is it will need to change six files and in leave code within code you've already written and stuff like that and the more that you're having that happen the more you have to slow down a lot because you have to look it's a fairly detailed relationship about this compared to that you know and the whole point here is to be able to go really fast right ideally I don't really have to read its code very closely right because and so the thing I found that like I kind of like is instead of treating it like a second pair of my hands writing line by line next to me is just treating it like another engineer on a team we like they're doing this module I'm doing that module or whatever and but with the the the iteration cycle being you know uh you know a minute instead of a day it's not like I look at the pr tomorrow I just look at it right now and so then the way I would translate that into how you actually like generate code with it is a lot of times the task I like to give it is just say like I need a component that takes these props and does and because if I almost give it like a really modular pure kind of extension that's in one file of my project then it the chance it generates that correctly is very high and and it's almost like hard to get that code wrong like if it if it renders it looks right it's it's right like definitionally but the more it's interleaved it changes with things like State Management or concurrency or whatever sometimes I have to look a little closer at like did to get the semantics right so so yeah that that's that's one of the things I've done to keep like the the kind of total cost of development like faster with AI is like try to shape some of the things I'm asking it to do as like modular implementations with a narrow interface like a function interface or component interface because it tends to be really successful with that because it's a very pure function like inputs and outputs in instead of just like yeah here's a whole code base and like I've got this big feature I want to add but I haven't told you where to put it and then I've I've kind of got a lot of reading to do with that point because it might insert lots of things in lots of different places yeah just continue on that topic because I think it's a good topic the thing I have found curs are really good with are just AI tools especially ones that can actually index your code base these are the ones I'm kind of I think are the most valuable is when I start with like a starter kit and it already has modals and dialogues and alerts and all this stuff already there when you ask cursor to add a new page that shows a modal that shows a red error and does validation and you like actually give it those things in your prompt you'll be surprised how accurate it just generates a whole page you click on a button it shows the dialogue you click on submit it says hey you didn't fill out these forms like it just adds all that in and then it also like will add in the backend and make sure that I me you need to verify the validation in the authentication gets set up but it's just the the more of a solid foundation your codebase has like if you're following good patterns you're decoupling your components I feel like cursor and other AI tools just do better at generating the code because you've already started at like a professional level and it's just learning from what you've written to just add on to it you start from scratch that's when you start writing to really bad bad code that gets generated and that's also why I'm talking about the G of the G of four stuff is like if you can know the lingo like the AI can get a very specific implementation like okay I need you to a couple these components and use a decorator pattern and it can just do it ver that's a good idea yeah that makes a lot of sense yeah you want to give it like clean attachment points and if it doesn't have any attachment points yeah I mean you're thinking gang of four is like there like the Good Foundations like good attachment points and the Gang of four is like relatively specific description of the like the the way you want to attach or yeah I agree if if it doesn't have any of that then there's no idioms it's extending so sometimes it'll kind of do anything and you know that anything might be good or you might go oh man like I've seen it change code I didn't tell the change like it opens up files I did not tell the change and it starts modifying stuff I'm like chill my code's not that bad you don't need to fix it for me that's kind of why yeah me from like a like a pain perspective the attitude I have about it it's just like I just know if you put it in this file and it exports this one thing and this is what the interface is and don't touch any other files I I it it's a lot less angst for me wondering about those changes to things that were already there as opposed to like yep new thing that's the functional interface off we go so that's really tracks though I think with what you're saying on like if you've got these idioms that you like and it's extending them then you can be much more confident about quickly accepting the work it does yeah be like instead of being like I like that idea though start just if you just give it an interface like here's my typescript interface Implement something that works for this I bet you it could do a really good job just cranking out the functionality for that right cool well Cody thank you thank you for coming on the podcast how did the first your first ever podcast go was it okay I think it went pretty good anything else we should have talked about that we didn't talk about man I I I could probably have so many random content creation tips but uh if you have any tips you've seen our content so you can be you can be uh brutally honest like what do you think we should do to get more people watching convex content to load up your channel just kind of it live you know I would say like the podcast thing you're doing is great I don't know many podcasts that are actually like at a at a lowlevel like database type of talk so I think that's really good that you guys are doing that I think just trying to get as much content out as possible so that you can actually like I I treat my channel as like the agile process right you don't know what you're doing is right until you get it in front of your viewers so the only way to do that is by keep on publishing content and then you'll get those signals that hey we're on to something with this video like I'm sure if you went through your Channel right now you probably have a video that has like 100,000 views in fact I think you do it's like the one where you have like the nice animations like conx is a new type of database I don't know whoever made that video but they need to make more videos like that because that's a signal that something you know connected with your audience one of the thing I think we haven't done good enough at is just kind of getting the volume up to learn more about what works and what doesn't work so we're we're actually we're getting more videos out hopefully these these days there's this uh new new person on our team Mike who's starting to put a lot of good AI convex and AI stuff out there yeah I think just finding consistency I I think I haven't ran like a company channel so it's a little bit different the the thing that I see with YouTube is typically people come for the person so I go to Theo because I want to watch Theo I go to primagen because I want to watch primagen I go to convex his channel I see basically all your engineers making content which isn't necessarily bad I just it's out of my Niche like I don't know how you'd you juggle that but I think people come and they'll they'll probably watch certain people that they connect with more than others and I think also if you can make sure that the recording style for all the videos is consistent so like let's say you're doing live coding make sure everyone's webcam is in a certain location with a certain type of style like the thumbnails you guys are doing it seems like you guys are experimenting which is good you need to figure out which thumbnails perform better than others we're starting to in I was say in the last six weeks we've started to try to standardize on some formats for repeatability like and like finally making our version of like like the internal guide and notion on like here's how we make a video basically so and which is just say so far yeah what you're saying makes ton of sense it's all stuff we want to get better at yeah like if you go to your channel the ultimate typescript back in like that one has 118,000 views from a year ago like I would just say look into that you don't have to make all your videos that because I think probably you put a lot of work into building that and then also just like I guess in at the beginning of the video like that's like the most important part like you've seen the drop off charts like you got to get them hooked in and typically a hook is like here's a problem that you probably ran into because I'm a I'm I'm a developer just like you and here's a really quick fix now let's walk you through how I can implement this like just some type of hook that keeps them around um when I started doing my channel I feel like I didn't really have a strong like first 30 seconds is like the most important and then someone who's actually interested in learning they'll stick around for the rest if if you got them hooked in all right Cody thank you so much for jumping on this was great yeah thank you for having me all right thanks everybody we'll be back soon with another episode of databased [Music]