Episode 2 - Limits of Tools & Creativity

Fired By Robots Podcast The Episode So, hello Ivers, how are you today? Hi Paweł, doing great, doing great, can't complain Wysaer is going here in Ireland, enjoying that, yeah Okay, awesome, so what have you been up to this week? This week I'm, I have to say that I'm kind of still getting into the mode about kind of handling my own time, you know Because I was just fired recently and before I was doing all my kind of projects on side And the day job, you know, kind of gives some structure to the day for you But currently I don't have anyone to call me on meetings or basically I don't have time to stop working or something or start working on something So possibilities are endless but that's also kind of confusing because you have to do everything You can choose anything but you have to choose to do something on end So yeah Yeah, that kind of structure that regular job gives you with meetings etc. It's limiting but having some limits is actually important, way more important I learned this the hard way back in the day It's actually important for your freedom basically Because you need to like for example when I used to make music, I used to make hip hop beats Before I started basically programming even more tools in your arsenal basically Where you can use thousands of instruments but setting up limits Okay, I'm using only three sounds on this, this kind of limits can help you with creativity, you know, your creativity sparks up Okay, I need to do that or setting up even a time schedule Okay, okay, I need to spend 30 minutes on this and having this kind of limits always helped me for to create like better stuff And I think I've lost that ability ever since vibe coding became a thing because I've been building like maniac like seven projects at once Yeah, yeah, that's the tricky thing You can basically work on multiple projects at the same time and then you are not 100% in any of them And then you also do not finish them I found that I was doing live streaming on weekend, Saturday and Sunday, both days And that kind of helps because you at least you are on some kind of schedule You can't jump from one thing to another because you're kind of sharing what you are doing And then someone following this kind of process and then you basically limit yourself to one thing You are not kind of trying to prompt three projects in parallel while still, I don't know, writing script for YouTube Or I don't know, creating thumbnail, you know, all together Yeah So that's my plan actually I'm kind of planning to still do two to three times in a week live streams So I think that will help me with this kind of progress Because you have to set goals a little bit more clear for yourself what you are planning to do And another thing is I'm going to I'm going to coworking place where other indie hackers are coming together and working And the format is so that you basically on the morning you have some kind of standup Basically you're saying who you are, what are you building and what is your kind of goal for today? And at the end of the day there are demos where if you want to you can show what you built and get feedback from the group Oh, that's cool Yeah, yeah, that's very interesting Yeah, that's like indie hackers coming together every week And I like that and that also gives you kind of structure And usually I'm going only with laptops Sometimes I have iPod with me but that's limit you a lot because you have just one screen One focus you're basically doing whatever you are doing And I found those sessions quite productive At home usually I'm not that productive I maybe have progress on multiple projects but that's not that big like focusing on one thing for whole day Yeah, we spend time like perfecting our home setups And it ends up like being on your laptop in some cafe is like the most productive time, yeah? Yeah, yeah, that's also you know that quite often we are very productive on flights, you know? I do not fly often but when I do you sometimes figure out that you basically can do a lot in those few hours on the plane Because you have nothing else to do, you have nowhere to go You don't need to, I don't know, load dishwasher or go to the fridge check what's there? Is there something new? You're just sitting and working, you know? Yeah, yeah, yeah, I totally agree You like exactly like this is my like main wish issue with working remotely basically Because I've been working remotely since 2018 I basically that was one of my actually goals into getting into programming I'm looking for a remote work That's my main reason of looking for programming job And then I had to like wait another half year for another offer basically to get But I've got what I wanted and then the rest is history We I'm on my ninth year right now But yeah, that's true that working at home has has its benefits You you have well, you don't waste time for commute But I often think okay, there are things that I probably would need to do in the meanwhile And it's always something distracting you My son is homeschooled so so basically there's always always things to do There are pros and cons definitely I I still kind of remember how it was when I was working in office job when my kids were small I so I was driving my daughter kindergarten She was basically in kindergarten at seven Then I was driving to the job to be at job at about eight And then I was working until 5.30, six Driving back and picking hair again from the kindergarten again at seven at the evening And basically I saw her let's say very little because just those few hours from picking her up from the kindergarten And and I don't know at at nine at nine at latest ten she should sleep already you know to be able to be up in the morning So yeah sure you can be super productive in that office working that corporate job But I choose being kind of distracted at home because now when my still when when my daughter coming back from school I am at home I can spend a little bit time with her you know talk about how it's going and so on That's totally different experience so I will not change that for for anything Oh no I wouldn't change that either Yeah yeah but but yeah it's a especially living together with super pro productive person like my wife is She can just basically like that like click and focus on thing she focus do something then go cook meal in the I don't know 15 minutes come back again deep into work and working I I I am not that person I'm easily distracted I can distract myself I don't need even external distractions you know yeah yeah yeah also in that in that way we are like very in very similar situation I was wondering how is your new app doing right now I started to build subscription tracker so my new application I basically built it because I agreed with two of my friends that we will Do like developers challenge let's build a new application and get it it first to the fast to the hundred bucks Otherwise I will not build new applications so that's my reason why I am building it otherwise I will basically focus on my existing applications and double down on marketing features and so on But yeah I started that application and that is just a subscription tracker because solving my own problem I just realized that I have no idea how much money I'm paying for subscriptions which I'm using which I'm paying but not using and so on So I basically have probably MVP already done I was working on weekend on this application I used screen designs AI to generate user interface that was surprisingly good to be honest First time when I see it is a really kind of usable usable I generated design I think today or tomorrow I will do like final kind of push to fix some design kind of imperfections Let's say so not even kind of bugs but somewhere kind of screen looks a little bit weird somewhere scroll to file far away and so on you know like Some small things I will test with few more different Invoices because what I'm trying to do I don't want to add all the data manually because I'm lazy and why should we do if we have robots for that I just want to parse whatever is possible to parse from email that you're getting new subscription and add this data kind of semi-automatically let's say so not like Silently somehow scanning email but when you've shared email to the application application should at least do the best to try to read this information for whatever we are throwing is it PDF is it just email body copy paste something you know and yeah I just want to test with different multiple subscription emails PDFs and so on few that I tried and train at my app kind of working so I think test flight So these definitely is this week yes Oh awesome so do you use like apples is like OCR for the PDF yes OCR but I had but I have to say that I so far I didn't write a single line of code in this application I basically that's just basically YOLO application YOLO AI kind of agent driven application and in the middle of the app I actually switched from the codecs to Gemini in Anthropic and also just because I was told so on the live stream that was kind of I don't know part of the show part of the fun you know doing live stream and then I was told okay Gemini actually is very good and I said really is it let's give it a try easy and I don't know it well I have to say that you know I get this application more or less kind of working on that live stream and I didn't look into details of the code in this application I think I will do that tomorrow on this kind of indie hackers session just to go through the application look how it is and so on because currently I can tell you yes it solves problems that I'm kind of saying to solve but how is it really better than codecs I don't know yet so I can kind of say but it's not bad it's not worse than cloud that's for sure okay oh that's that's great I I saw your live stream and I was actually really surprised by the results you get for AI generated UI for using screen designs I've built like two applications like last month and I've actually used like the change GPT new image generation model I've given I have given it like a summary of the application and I ask it to generate the app UI and it has also been really good so I think like the UI mockups generations right now is at really great place in terms of AI generated ones like for me I remember when I was like building apps like two years ago something like that the hardest part actually was having that vision at the beginning on how that's supposed to look was like the unlock of building the rest of the app basically so I spent sometimes I spent way too much time and I never designed the apps I always designed with code I always started with code I never wanted to learn those those those design tools I just I just thought okay I can design with code I'll bit so let's do it let's do it this way and and but it took me a way longer time in order to create it like basically a home screen for application but I when I did then I was very easily able to follow and have the structure for the rest of the app but nowadays like you can like just describe the app and the GPT image 2 or something like that model is really good right now I really like the the theme I pasted it into codex codex implemented it codex in terms of UI has used has been historically worse than cloud in terms of frontend development but whenever you give it like an image to generate and follow the the assets that's that it it's really good right now in my opinion it's on par with cloud frontend abilities in my opinion yeah and I I really like that we moved really further but like your screen designs I'm wondering what what they are doing like if they're using like a GPT 2 model image model or they're like have trained so some of their own AI what do you think do you know anything no I don't I don't know I don't know I don't have any information anything I don't I don't know how do they do that I just know that they have huge database of images and my assumption is that they did train some kind of their own model probably based on GPT open AI kind of models and yeah I think that they definitely did some kind of training because I even kind of try to ask like you know kind of more open questions and kind of saying okay you're looking into the similar applications find this and that and so on and that basically but they have did it and did quite successfully really useful tool to be honest I'm surprised that they did and yeah they released an update for it not so far away few few days or weeks week maybe ago they released version 2 that they currently have before it was way more simple but yeah my design process is quite similar to yours so not design no design just jumping straight to the Xcode and trying to build something and then I see what works what doesn't you know start to play around with the top and say oh no actually the screen is not it you know I need to move it around but sometimes when I can't visualize what I want to build in my head then I'm kind of going to the just pen and paper and I'm kind of sketching something and then I'm looking for again for examples then there was mobbing for a while ago now I'm using the screen designs just to look for kind of similar apps similar screens and then I'm kind of grabbing some images some some for inspiration from them and sometimes I'm kind of moving to the design apps like it was sketch up for a while ago but now it's more Figma but I rarely I really rarely kind of creating any kind of design that's more like kind of mood board you know and then copy pasting some elements and then maybe creating something really simple I do not even kind of worry to make like perfect design like I would like to have it at the end I just need that today in my mind what exactly I want to build and then I again kind of jumping back to the Xcode or codex already nowadays you know previously that was Xcode because manually basically do all that kind of code writing and everything yeah yeah I have a really similar process I've I am I am still building one of one apps one app that I've had an idea for and I would say it's like almost ready but I thought okay that's an idea for app for women so I'm going to spend more time on polishing the UI than actually I've done for any other apps and I've I've I basically spent two days prompting like the animated logo I've spent some time prompting to basically have animated version of app icon basically because as again GPT generates most of my accounts and I like it one shot at basically my app app icon idea whenever I gave it to you and that's that's that was really great idea and I'm trying I think I try to recreating this with some kind animated fashion and I've actually gone gotten into really really good results so I'm way closer right now to finalizing this I need to spend some time now into like generating some audio assets for this for these apps there are part there are some parts of I would say kind of meditation things in there so I need to like spend some time to have like this audio generated and I'm always using 11 labs API for like everything music related and AA voice generation related I've used that in my cold Smith application where I added like breathing meditations there there there is some I've added that one feature to make the app a little bit more broad than just only called the planch tracker have you used 11 labs by the way it's it's really fun it's really fun API to use no no I didn't I was just kind of thinking about audio kind of generation one focus application and I wanted to add like rain sounds you know forest sounds waves something like that but yeah I somehow abandoned it today it's not like abandoned that's you know one of those days which are on hold someday someday later you know but I just quickly checked where I can find already ready to download kind of clips and there are so many kind of for example for music there are so many places where you can get music for social media like YouTube and so on but as soon as you start to look for using in the apps then it becomes crazy expensive or non-existing at all so that's yeah I didn't didn't I didn't finalize that research I should look on that 11 labs maybe I can generate something yeah yeah I would say it's in terms of like music generation it's quite off on the expensive side I have one application that has AA beat making applications so I have I have unique perspective I feel I have quite unique perspective on AA beat making and I've spent like three weeks and basically for that apps trying to perfect prompts and making that user prompts will get us good translated into into what they might need as as as possible and and basically like one beat generation usually in on average costs around 50 50 cents basically so that's quite expensive I would say that's why the app is right now a hard to pay world and I'm thinking about different ways of monetizing this and I will I'm going to be releasing this app for Android to really soon I this is an app I've I've been rebuilding in from iOS native to actually react native and the export SDK for expo have been released last week and I've already updated it I'm all over the place but generally I wanted to say like about the finalize this because I've I've been jumping through the topics there on on 11 labs it's really good and it's actually a polish startup that's I think recently they've raised some ungodly amount of money again and but that's like one of the biggest AI startups and it's from Poland basically so I'm I'm really happy and I've actually had a short interaction between the interaction between the interaction between the CEO of the CEO of the 11 lab on twitter because I've seen some other IT company shared story and congratulating them on the X and I said to them okay congratulations really great product but your own mobile app could be a bit better so if you ever need help let me know and then I found out that the CEO of the CEO and then I found out that the CEO answered me that so yeah that's what that was cool I I remember like I think they've improved a lot to their mobile app already I've I used it like last year quite a few times because I I was building my the AI video generation application where I also used 11 lab application And I've thought, okay, I can build this kind of really nice experience of generating audio using your API. Why you don't have that in your application, basically. And that was my thought. And yeah, they have actually improved it a lot. Cool. Okay. I see we have some MRR numbers in our kind of notes. I can share my, I'm quite happy to say that my MRR is kind of going up. Even I'm kind of lazy. It's already more than 4,000. It was below. It was below. It was 3,000 something. 3,900 something last week. And now it's already on four. So it's good. Yeah, it's good. And yeah, because I got also received kind of last salary from my last job. So it's getting more important every day and every week. How your applications are doing? Are you happy with that? Are you ready to share something? Yeah. I can share. I don't have, I am actually using right now my phone. So I cannot share to the camera right now. But before I'm going to, it will get to me. I wonder what, what do you know what changed for you in the last week that it grew? Or that's like some organic change? Yes, I know. I changed prices. I raced prices. But that was prices where I said, let's say, I would say a week, few weeks ago. And basically MRR is going up. I was doing, I was running multiple price experiments on different products and so on. But I figured out that it's maybe not like a rocket science, but often maybe it makes sense to say it loud. Often it is so that, yeah, you can, for cheaper price, you are getting more trials and maybe even more customers. But if you double the prices, not always half of the customers will go away. Maybe you will have a little bit less paying users, but much more money. And I tested multiple different price points, multiple different kind of offers. And now I found the price point where I'm kind of making a little bit more money from the same product and same offer and same paywall, you know. So that's, that's the only thing I changed. Yeah. Okay. Yeah. That's, I think that's a very valuable lesson because I've noticed that multiple times on multiple different occasions. I, whenever you like actually ask for a bigger price, usually nothing bad really happens. The, usually you just get more upside basically. And I think we tend to over under price ourselves. I also been experimenting with pricing some time ago, increased a lot of pricing for Vidnots and my main Vidnots applications. I, I increased the pricing across all of my applications basically, but specifically my Vidnots application. I even, and that helped me get to even higher number right now. But in terms of like MRR, my, I've had pretty bad two weeks last week. And plus I, I've had some errors on Android and I'm building this app and it's like quite getting, I'm trying to not make it as much as complicated as it is. But, but while making it, I've had some mistakes and there has been some issues with a premium entitlement. And some users complained that they like purchased a trial or, or a plan and the app wasn't, wasn't recognizing them as premium users. And that's, that, that has definitely lowered my, my MRR numbers in terms for, for this application. And, um, I think I've, I've, I have brought, I've been testing one, two, three, four different onboardings right now. So like for the past two months, I've been testing four different onboardings and I'm going to be testing two more in the upcoming months in order to like figure out what's the best onboarding idea. And yeah, so my numbers are around right now, around two and three K 2.3 revenue. And I've been pretty much decided basically on focusing on making sure I, I'm going to go with marketing. I've started posting again for social media. I've actually done some batch content basically for Instagram, YouTube, and Tik Tok, uh, for my one account and for focusing on feed notes, especially. And I, I wanted to go all in on that application, try to make it as good as it can. It has, I feel it has a good potential to grow even more and it, it didn't even scratch a surface, but like last week. Showed me that my, some of my experiments in ASO has actually turned quite all right. And I've didn't have like a, uh, notifications for all the projects I've built in revenue cut. I'm, I'm using two accounts on revenue cut. That was the issue. And when I connected my, uh, like all of them to my phone and I, I've noticed that there, there has been some applications that I've built. Some time ago are actually generating some revenue. That's not great money, but I've like today. I, especially today I look up, okay. This is revenue from five different applications and it's amounts to something. It's not like tons of money right now still, but it amounts to something. And I started, I started thinking, okay, maybe I should pump out those ASO ideas. Because I've actually had some success in terms of earning some revenue from a few of those, not really great numbers, but like maybe, maybe. I, I think that right now the ASO, um, there, the ASO, uh, app store boost that, uh, we used to have is no longer there. I think that's generally recognized or in the community, but. I feel like you need to wait a little bit longer to see if there are any results from ASO. And I've been like working on improving all of my apps release ASO for the, for the past two years, basically. I've had, I have like 15 of them. I've seen some, I've seen some changes, some, uh, improvements. So I'm kind of like hesitating right now. I have a boiler plate for those and I could pump out like those 25 until the end of the year. So there are kind of two approaches on, in the app development. So go all in on one application, do marketing push, push it hard, you know, and basically kind of, yeah. Do content marketing, we just copy what else you can copy from everyone who is, what everyone is doing on the social media and try to kind of push it all to the 10 K. Another option is just to have kind of handful of applications in the app store and how kind of smaller revenue streams from each and everyone. And currently there is kind of, kind of a trend. My bubble is talking about, oh, ASO, app store search optimization is dead without kind of marketing applications, basically just don't get downloads and so on. But I'm in similar situation like you are, right? I have also multiple application in the app store and I see that most of them are generating money without me actively kind of marketing them. And I don't know what is exactly like, you know, secret formula or blueprint or secret sauce here. But the thing is that if your application is in the app store long enough, it's getting downloads, it's getting reviews. It will keep getting them, you know, because I see, I just opened it my also revenue card kind of dashboard and I'm looking. Yes, I see sort of basically downloads and sales from applications, which I didn't market or promote for a while, you know. And of course, when you do this marketing pushing, then those numbers are going up. Yes. Yeah, it works as well in that way. But as we know, time is the limit. Everyone has only 24 hours and you can't promote all the applications and going all in on one application. It's also one thing it is scary. Another thing it doesn't really work well for my brain to focus on one thing forever. Yeah, yeah. Yeah. It's like I feel like in that realm we are very alike because I feel like our brains works very similarly in that case. Like I am also hesitant on going all in and dropping everything what I do and just focusing on marketing. I decided, OK, I'm focusing on marketing on this one, but I'm undecided what to do with the rest one. Last week on the first episode, I talked a little bit on outreaching to creators, influencers, etc. And I've re-subscribed to the Topia Press platform where I set up the outreach agent. There are there's outreach agent and then that is quite easy to set it up. And I'm basically lowballing creators trying to ask them for creating for cooperation. So I'm focusing on the on Ovid notes and I found myself a few other influencers for this new app I'm building. And that's in a woman space and I've got answers and I potentially might have one collaborator here. So we'll see. And that may be a really good before launch thing to like boot, like jumpstart it. We'll see. I'm definitely in the camp. I won't I it's hard for me to be to work on one thing and focus on one thing them. And that's what has been happening for me. And it's not great for my ADD brain because it feels slow, basically. But there are big updates, the big changes happening. From my side, I'm kind of planning to I'm releasing this subscription tracking application. I will do a little bit of marketing push on that, but I maybe I will try for that one. Some kind of a video, some kind of marketing, you know, kind of reaction stuff. I basically parsed this receipt and so, so that it's so much money in a year, you know, stuff like that, you know, kind of realization videos, you know, kind of surprised faces. Yeah. Yeah. UGC reactions. Yeah. I think that's cool work for some kind of UGC. Maybe it's the top yappers kind of something like that. I don't think I will push a lot of personal effort into the marketing for that application. I, because I would like to go back to that, my rider's log application and give that kind of more kind of bigger shot, you know, kind of do a little bit more marketing because for that, I think I should create some content myself to market it. We'll see, but that's the current plan, you know, kind of to do, to do, to release application and try to do some less time consuming marketing for that application, you know, and, and, and to do not try to promote everything, you know, five hours. Yeah. Yeah. Yeah. Yeah. That's, that's impossible. That's why like ASO is kind of like, I know it's kind of that, but it's, it is, it's really good when that, like, when you, you are able to like jump into this angle algorithm and suddenly you have some revenue. I've actually had last night. I've had some fat, my first revenue from my game application. Oh, cool. Congrats. Yeah. Yeah. The idle, idle ranging game, which I actually started playing a little bit more. So I've added some updates to it too. So yeah, I'm like split into so many things right now. I've, I've, I'm still split all the time. Like, and I think that's been a constant state of my mind being split on the, what to like put my all, all in hands there. Uh, but in terms of like generating content and AI and, and automate, uh, automations, I remember you've shared in your community a little bit about how you're using open cloud, uh, to, uh, to, uh, to do some automations. Do you, do you, do you still use, use this or not? So those are like kind of two, two, two things. Then open cloud. That's one thing. I do not use it anymore. I played around with it. What I had for open cloud with it, what I automated. I am on, I have been on Instagram for multiple years already. And I have a lot of photos, you know, with, I don't know, me programming stuff like that, you know, and my, I had a day that, oh, actually I can do. Like automated image posts when you have some text over the image and something to promote applications. Let's say I have that focus timer application, like let's say fancy Pomodoro app with app blocking and so on. So some kind of motivation stuff, kind of text over images, create those carousels, publish on social media. So technically it was working. I build automation. I basically kind of filtered a bunch of images, which are kind of more like a vibe applicable for that, for those cases. But I don't know, results wasn't there. You know, that content didn't get a lot of views. I think maybe those carousels with text are not working anymore. I don't know. I built this kind of system. I kind of were scheduling those posts automatically. I was posting them automatically, but yeah, I don't know. I didn't like the results. And to be honest, I also didn't feel good about just bombarding my socials with some kind of shit content. I bet I prefer to create short form video from my own long form videos or from this podcast, you know, when there is some value in that, when basically that's just smaller piece from something bigger that I created. And I feel okay about posting that it still feels like bombarding audience with content because you can post a lot and multiple posts in a day. And I don't know if it's good or bad, or am I annoying for my audience or they do not notice because the algorithm do not show them. But yeah, basically, yeah, that's, that's, that's, that's what I had set up. I didn't like this. And then I built tool to make short form content from long form videos and also scheduling. But that's like, again, without open cloud, I don't use that anymore. I'm just using application. I found that if you're doing some specific thing, it's actually just easier less. There are less errors if you just build tool for that. Like I built application, it just works, you know, with cloud, but it was some kind of dealing with problems all the time, you know, something broke up, automations didn't work. They forgot about, okay, where, where images are, what else not, you know, you basically kind of fixing that pipe. It's like leaking pipe. It's kind of works, works. And then you see something is doesn't work or okay, why did you grab five, five, same image posted, same image five times? If we clearly have folder with thousands of images, if we were to pick, then it's forgot where, where are, where are filtered images? And basically you write rules, you, you agree with cloud button rules, and then it's forgot it. And you have to remind, okay, what we decided to do and so on. It's, it looks fancy when you show demo and you're saying, Oh, basically I automated my marketing channel. This is how it works. Yeah. When you do this demo, it works nice. And that sounds amazing. When you are one who is kind of fixing those leaking pipes, then it's kind of getting annoying, you know? And I found that if you do kind of single purpose application, it just works more stable, better. And problems are easier to fix. If you found problem, you fix it, it fixes it. That problem will not happen again, you know? So yeah. So I noticed that you have similar experience as, as I did with open glow. When, whenever it released, I got hyped up and decided, okay, what we can do with it. I went a quite a different route. Whenever I had an idea for an app in the shower, I've, I gave this, I messaged through telegram and I made it at some point. I had so much like, this is a popular time right now on Twitter. AI psychosis that at some point I started building like five or six applications through telegram on that GitHub account. And let me think one idea survived that I actually continued working on and I'm using right now a few other ideas, just like in a short span of time, like I've built 20 repositories there. And that are like just laying there. There are a few good ideas I had there, but I like abandoned them. There's like my brain at some point was okay. I need to build all of this. And I think that this illusion of productivity boost that AI gave us basically, because like, maybe without AI, I would be a little bit more focused on what to like this one small project. Then maybe this one small project. And we, if it has been a huge effort to build another idea, I wouldn't even try that. I think that illusion of productivity, it turned me at some point into like some kind of productivity maniac. And then I, then when a claw, anthropic basically disabled usage of your, your account and subscription inside of open claw, I basically stopped because like I spent $200 basically on only on, on, on, on top of the $200 subscription. I spent another $200 for like the tokens. So I stopped using that. I was going nowhere with it. And I forgot about this AI agent, but I've been reading a lot in the past, I don't know, three weeks about Hermes agent. And I thought, okay, let's try again. And I thought, okay, let's try again because open claw was all right, but I had very broken experience in the meanwhile with it. And, and, and let's try it out. And I know that one of the issues I had with broken experience and recurring cost is having that agent on personal server. I think that's the, that's the shortcut. Um, what stands for, but anyway, but so I thought, okay, I, I don't, I'm not going to buy a Mac mini. I have a steam deck it's on Linux and I'm not using it at all. Let's, uh, let's try it out and set it up. The setup was surprisingly easy of Hermes agent is just like two commands install. Then like connection to the true telegram was so easy. I remember it took me so much time using open cloud because there was some issues and it was so easy. And I thought, okay, I decided, okay, I want to like this, make this my personal assistant, maybe research, some research, no coding, no coding happens here. And I'm using my open AI subscription inside Hermes. There's a separate, uh, uh, basically allowance inside open AI subscription. You have your main allowance for all the models. And I, I, I, I am really actually happy because I set it up to be answering those emails that I, uh, you have the, from outreach agent. Uh, uh, uh, and, uh, and, uh, and it's, it helps me read those emails and that like takes off a bit, uh, from my mind about those, uh, outreach things. And I've actually set it up also that I can, um, um, set up to save some stuff in notion. And I, whenever I have maybe idea for, uh, one topic or something like that for the podcast to update, like the, uh, the episode and write the episode draft here, I work together with it. And I have another like agent for, uh, uh, on my laptop, basically on my main machine where I work with codex on, uh, uh, generating some UGC reactions videos for my Vidnotes applications. I finally got it to like working and I actually was able to schedule them with posties. We'll see how this goes. If this will help me like get some, uh, um, some, some users. We'll see how it will evolve in the future, but I really liked it. I have, I, I have it, uh, working on my steam deck. I, I can communicate with it. I can save those ideas. I've asked it to research, for example, for more creators that I can outreach to. And on top of like top yuppers. And I, I, I had really good, um, a good experience right now with this Hermes agent compared to open claw. And I'm only just scratching the surface right now. I, I want, I want it to be like maybe potentially my sim have that help me promote my application. I'm focusing right now on like the Vidnotes all the time application, but we'll see how it will go in the future. But why you like, uh, why, like, is, is this, is there something that you like in this Hermes better, better than cloud, cloud bot compared besides the pricing? So there was problem with this kind of pricing cloud kind of didn't allow us to use it anymore in kind of general subscription. That was one of the reasons for me as well, why I stopped using it, but is there something else or that's basically the same or it's better? Basically open clock. Open clock. You can use open AI subscription in it too. So, uh, there's no, no, no, uh, this, this issue is no longer existing here, but like the ease of setting everything up is way easier for me. Like for example, like what I like it is like, it's not breaking all the time that I can quickly go click through everything inside steam deck. Like having that manual way of like fixing stuff. It's way easier for me than having that on the server. And that's really important, but basically it's not breaking. I don't, I didn't add any crazy features right now yet to it. You can do way more potentially way more of it. And I'm going to be testing this with one other tool I'm building. Basically the important thing is like using it on your device that you have that you can like go and quickly like fix. For example, when you set up posties, which is like this agent, I set up, you need to authenticate the device and it opens inside the browser. And you click and it's authenticated and it's working. Basically that's like very simple setup. And I couldn't set up email properly on open cloud access. And now I have it easy, really fun, easy thing to do here is also setting up cron jobs. Basically I have like one cron job that's happening every hour that it basically the checking my email and giving me information about the email. I have like separate email for that. I'm going to keep like exploring and definitely keep on talking on that because this is really interesting stuff. Yeah, it sounds interesting. I had, I had my cloud bot installed on, on my, on my gaming PC and that was on Docker kind of image. And that's basically like same, same, like having it on in cloud. I think technically from configuration kind of perspective and everything. And yeah, this becomes kind of tricky when you try, I had again, also same similar setup. I created email for it. And then I asked a load kind of sharded access to some Google docs, you know, to that email, to that account, which was cloud bot account and so on. And yeah, it was kind of working, but everything was so fragile and setup was kind of also, maybe that's interesting and fun first time to do that. But generally you're just wasting time on all that setup. And at the end of the day, I can have all those conversations with, with cloud code in terminal or in cloud application as well. I can just do this brainstorming and everything, you know, currently those tools are improving so fast. And that, that gap in between what cloud bot was able to do and what just cloud or codecs can do now is smaller basically because you may have similar conversations. You can allow to use computer. I, for example, created like process to cloud bot to generate invoices and basically, and that was again, kind of a little bit fragile, just kind of going to Google doc, kind of creating invoice. And that's what I was saying. I was just sending to me in email and so on, but now you can give access to the folder to cloud on PC on, on your Mac or PC. And basically you can do that locally. And, and, and that's relatively safe. You know, it's not like this cloud, but you're just installing it on your machine and basically giving access and who knows what will happen. You know, I will say that result is good enough to, for me to just use cloud and use codecs on, on my mind Mac. It's easier because all my files are here and I can basically charge them if it is needed. I don't need to somehow send them to the server, charge access, what else not. It's, it's just easier. Yeah. I don't know. Maybe I will, I will play around with that Hermes if I will have time for that, but I shouldn't because I have way more, way more, much more, more productive things to do. But it sounds interesting. And I have stream deck as well. When I'm kind of at home, I'm usually playing games on PC. So that's team deck is basically is kind of doing nothing for most of the time. So cool. Yeah. Yeah. I actually, I just installed like uninstalled one game to make some space on, on the steam deck. And that's basically it. You don't need to like do anything else there. You just are in the desktop mode mode, desktop mode. And it's, it's kind of working. I will see how this will evolve. I am trying. I have a one idea to have one more automation, like straight out of, um, to like get all my, because I don't have, uh, properly sorted out the costs, uh, like for app store ads. You receive every few days, uh, and PDF with the invoice for app store ads. And I want to, it to go through it and then, uh, summarize it for me and maybe some put it in some one place. So I know what's going on in here, in, in there properly. So that's one more fun automation tool. I'm going to try it out to set up, but for now it's quite so quite, uh, so far so good. Um, I really like it and we'll see if that evolves into something, but I've recently also started using, uh, even, uh, a codex, especially, I feel it's quite, quite smart and not lazy. Basically whenever I was using Claude, it's, it was quite lazy. I needed to ask for another thing. Like for example, do you know what ASC, uh, CLI is or not? No. It's app store, uh, connect CLI. And it's a, uh, free tool made by this, uh, one guy. He created a CLI in order, in order to have his agent to connect to app store and you can do basically key anything through it. And whenever I was like talking with Claude, it was kind of lazy whenever I wanted to add maybe, uh, some kind of like, for example, create a new version of the app and create what's new for all localizations that we support. Basically you can ask it to do. And like, especially with codex, for example, I've had multiple times for multiple, uh, my applications. I gave it like this command, build a new version, archive it, uh, submit to review. And then I like a few minutes later, my app, there is a waiting for review up with a new build there. You give it access and like you, it removes so much money while labor that you do with X code right now. And you archive automating other stuff. And that's one thing that I've been using it for. And it's been so helpful. Like I have another automation in my computer for creating basically screenshots in Figma and localizing those screenshots in Figma using Figma MCP and desktop bridge. And then you can, uh, use ASC CLI and basically do it inside one terminal. You, you, you have the whole process. And the one thing I didn't, I don't have right now properly is, uh, testing the app, but I've made this localized, localized shot like free tool to create UI screenshots. So I'm moving towards like more of the automated, this having that process more automated and I'm going to maybe record some kind of video on how to set it up. But I saw on your stream recently that you've, uh, uh, like been, I think releasing the app and creating this, uh, everything in the app store, like subscription, et cetera, by hand. And I thought I'm going to share with this with you. And that would, could be actually really helpful for our listeners. And because this ASC CLI tool, you can just like, you can even pass it the PDF, for example, with the, um, like latest revenue cut, like a report from the subscriptions and then give the context to your app. And then you basically create subscriptions with localized pricing and you can, you'd like the one last step you need to do is basically upload a screenshots of the paywall and everything else can be automated there inside your app store. So that's like going to save you, I think tons of time. Yeah. It's amazing tool. I basically open it and basically installing it as, as, as, as you're talking, I, I use it. Um, I still using, there is this, uh, uh, uh, so though. Yeah. Yeah. That's basically cool for localization and that kind of helps you to still do manually throws their application, but, uh, you can do everything much faster, you know, kind of, right. What's new, basically keywords, keywords update, all the kind of multiple localizations and so on. That's makes this kind of process faster and easier. Yeah. Yeah. If you can just do that with one prompt, it's, it's definitely better. I remember some time ago, I've used that for a while, but whenever like CCLI came out from this guy who opens artist thought, okay, it's crazy good. And there is a, I think there is some, uh, some, some work there on the Apple ads. So you can work with agent to even like help you like set up Apple ads and maybe analyze your ads and how to make it even better. That's, uh, another thing I might explore in the future, but it's really good tool. And I'll put the link to it in the podcast description, uh, later on for, for, for our, uh, listeners, because we have some listeners I've, uh, seen, I feel like the, our first episode did really well. We get around 30 subscribers, which is good. Really awesome. I think for everyone's, uh, all this is listening to it. So you posted a, uh, tons of clips on your socials and I've posted like one or two. Um, but I'm, uh, I really glad we started this podcast together and, uh, it's, uh, really good to talk with you, uh, for, I feel like we could do this, uh, like three hour podcasts. Uh, we, we won't run every week with battle and we won't run out of, uh, things to talk. Yeah. Yeah. Uh, yeah. It's, uh, I, I, I, yeah, I posted also those kind of, uh, short form clips on my socials and it's not everyone even found the podcast itself because I saw on YouTube today. I, I, I tried to found the source for the clip where it is, you know, we need to add some footer or something to those clips. So it's a podcast name to be easier to search. Yeah. Yeah. Yeah. I, I, yeah, we need to add that. I'll work on that, uh, this week and let's, uh, finish maybe the podcast with, uh, with positive vibe and talk a little bit about this week's text firings. We, we are fired by robots podcast and we talk about AI and getting fired by robots. I think we could make that a weekly segment of, uh, what's happened in the last week in tech firing. So I think on the day of our recording of the first podcast the last week, meta cut off cut around K people. So it's 10 per 10% of its workforce and the people were asked to like stay at home and they've been, yeah, that's quite a lot of, uh, employees being fired and meta is doing the best it has ever been in terms of like revenue, but they, they like reshaping, uh, the organization. And I, I still think that's not directly correct. They connected with only AI. I think like the, this is the steel Elon Musk's effect that he had on Twitter when he bought, like when he fired 80% of, uh, of the team and, and it still works. Yeah. But there are some, there are more things happening. So into it, I know slay off of about 3k employees. So almost 20%, uh, or turbo tax and quick books are very, uh, popular in the U S through botox is basically a product for filing taxes in the U S as far as I know. And quick boxes for regular auto accounting software click up also a fired like one cape above one K employees. And, uh, and weeks, I think the Israeli website building platform announced the largest layoff in since history around one K jobs and Kraken. Do you know Kraken? Yeah. Kraken is a crypto currently exchange and cut around 150 employees. So around 5%. And I have personal story here because I've applied to Kraken a few years ago and I've applied to Kraken again last, like, and I received an answer that they won't follow with me. But those numbers are crazy. So those, those thousands of people that's already kind of hard to kind of follow and kind of understand. And that's just one week. But also when we looking on that percentage of the company. So for Meta it's 20, 20, no, it's 10%. It's 10% of workflow. It's weeks 20%. Basically those, those are huge, huge kind of, that's big part of the company. You know, it's, it's, it's crazy. And yeah, good for them if they can really kind of optimize that work. And maybe that is the Elon Musk effect, as you said, you know, that, okay, you can basically fire 10, 20% of the people and companies still keeps working, maybe even more. But I don't know, those times are kind of all that overhiring, what was happening in Twitter and similar companies. I think that those persons are already gone. Maybe there is some AI replacements. I don't know what's going on. It's disturbing. It's crazy. Those changes. That's, that's crazy. And with all those people in market, they're still kind of getting questions. Oh, is it still makes sense to learn kind of programming and try to become a programmer? That's getting harder and harder to answer this question. Right. I still think, I still think that our kind of skills that we have, those programming skills are really valuable. And with getting those AI tools, we having more power, we can build stuff, you know, and make businesses out from that. But yeah, what to do for someone who is kind of pitching, choosing career. I don't know. It's kind of getting into IT and especially this question, oh, well, the, is it kind of good idea to become a programmer, to get a job as a programmer? I'm not so sure about that. Kind of becoming programmer to build, build businesses by yourself. I think, yeah, that's still valuable. And I'm happy that I'm doing that. But yeah, to some, to suggest someone to go in big tech, that's nowadays it's kind of iffy and very tricky. It's kind of tricky because it depends. Yeah. I've seen some TikToks of people like those aren't like the slacking of employees. These weren't like very successful employees, very good employees, as they say, basically. But the changes are strange and we are in a strange time in the industry. And I'm wondering where it's, where it's all going. And, and also I've seen, I'm not sure it's if it's also worldwide, but in Poland, there is the first time in many years, there is decline of user, of students applying to tech university, basically, to, to learn programming in the university. So not only programming, but generally in the joint IT, basically, there, there are less applications for computer science. So I'm wondering where, where is this going? But like, as we talked earlier, the AI, um, may, may, does a little bit more productive. Like the tool I mentioned, the ASC CLI. Yes. It saved me like basically maybe hour or two a day, but now my VNOS application has grown like quite well. I'm building tons of stuff on top of it. And I, I've learned that it's quite hard to maintain quality right now with, if, with AI coding, the making sure the whole system is coherent and there are no issues and bugs. That's not that magical basically. Yeah. But we'll see. It works like magic. If you have starting from scratch and building relatively small MVP project, then you can show, look at this. Basically you can build application in a few hours instead of few weeks and you don't need five developers to build it. Yes. Yes, of course. In those demo cases that works amazingly well. This is my new application, what we mentioned at the beginning of the episode, that subscription tracker. I didn't touch code yet. I will, I will do that probably today later on live stream and tomorrow working in this indie. The Hackers group, I will look through the code. I will find, I will fine tune design manually because that's just faster and more effective. My point is, yes, you can build application quickly and show it as a, as a success story and go to the management, upper management meetings and say, we can cut off half of the developers. See this, what this AI model is doing. We don't need them anymore. You can, you can paint this story. Yeah. If you want. And also you can give a legacy code and ask to implement few more features and it will break the. product in 10 different places, adding just one feature. You know, you can paint both of the stories. I don't know. I think that it is kind of moving forward progress. And I think AI will be somewhere in the middle. It will be able to work on more challenging projects and be kind of more reliable and so on. But I do not see it replacing developers anytime soon because there is no high jumps in progress anymore. Yes. Every model becomes better, but it's not like two times, three times, 10 times better. Yeah. And we are hearing that story that in three months, in six months, all the developers will be replaced. We are hearing this story for how many, two years already. So, okay. Or your job will be obsolete in three, in three months. But yeah, of course, AI improves or speed and efficiency of the developers, those who can build ups. And yeah, we need less work hands in the market for different reasons. Therefore, I'm not surprised that there is decline in studies. And they are seeing those presentations from Anthropic who else not that, who is saying, developer jobs will be obsolete in three months. Why are you going to study at all? Yeah. Yeah. Yeah. Yeah. It turns out like the amount of, I think like the, the silver lining is like the amount of lines of code written wasn't the bottleneck either way. And that's basically the case right now. Okay. We can build maybe faster, some things, but it's more important than ever to focus. What's really important to build. I have multiple projects right now where I'm actually cutting scope and kind of, uh, reworking them. So they are basically more focused tools. And I think like, this is the direction we should be going to. We'll see how these, uh, AI improvements happen. There is no enough power for basically everyone. There's power shortage in right now. So that's, that's, that are the issues, uh, that maybe Elon Musk trying to solve. Like we are some putting some data centers in space and powered by solar thingies. Power is one thing. Another thing is also AI costs. You know, sometimes you can find that it is cheaper to hire a new developer to do the job instead of burning tokens. Uh, I saw on Yahoo Finance Singapore. They're basically writing Microsoft is not completely stopping AI push, but the company is hitting the brakes on certain expensive third party AI tools and swifting engineers to cheaper in-house alternatives. So basically Microsoft found that they're kind of burning too much money on AI. So it's double-edged sword. It's, it's not, it's not that simple that just replace programmers with AI and it's not free. Yeah. AI costs money as well. And then also kind of, and, and when you, and when you start to implement AI in the company, when you start to try to replace those developments, that is less effective and more expensive than just having those developers. So yeah, I'm pretty sure that Meta Microsoft, they will be fine, but those kind of smaller companies who are kind of jumping out of the window together with big players may found their same self kind of in very bad situation. Yeah. Yeah, I agree. Okay. Let's finish it up. Thank you. Ivers. Uh, again, we are fired by robots podcast. My name is Pavel. The, this is my friend Ivers subscribe and comment and tell us what you liked about the episodes, what you don't like. Maybe there's something you would like us to discuss on the podcast. I will definitely read our, the comments. I'm reading all the comments and once again, thank you for watching and see you in the next one. Cheers. Thank you. See you. Bye. Bye. Bye. Bye. Bye. Bye. Bye. Bye.