Is Open Source the Move in the AI Era?

hey there I'm Parker x i LED product and technology for a startup that did well over $50 million a year in sales and after that I did a bunch of different SAS products to learn how to leverage AI the ideas didn't work but the execution was there and now I'm scaling an AI Services Company where I help Main Street Implement AI things that actually work so on this channel I do a daily upload where I talk about what's going on and some strategy stuff I answer some questions and let's just jump into it so yeah there's Q&A there's some of the AI news and you can get AI news anywhere but this is filtered from somebody that actually builts the products it's not that I am just this media man a mystery man with a media company uh so it should be a little bit more helpful and I only talk about news if it's relevant or meaningful then some strategy and then the building so first question came in and it was about my SAS product that's been on on hold because I've had trouble getting it to the Finish Line it's a multi- agentic platform that frankly just has so many things going on in it that it makes it really hard to get it to work consistently and if I can find how to get to YouTube Studio there we go basically the question was hey this map thing looks really cool well have you thought about open sourcing it so yeah um see where is it in here open sourcing it it looks like a great project so this is a good question because the open source Community could be really helpful from a business perspective and a cash flow perspective to be able to get it out I think it would be really important for me as the owner of the project to keep a hold on the quality obviously everyone in OSS wants high quality but that would kind of be the fear is that you'd have just lowquality work getting submitted I guess there's really no downside to that but what I wanted to do is actually put together a little research on this it looks like it's still going through but I prompted Gemini deep research to figure out you know I'm contemplating doing this I want to incentivize developers involved by doing a profit share now this product's not out but I think what could be really interesting longer term is taking the kind of goal-based compensation that's really popular in obviously sales is you know fully commissioned but goal-based compensation can be percentage of sales it can be cash bonuses but in this case it would be okay if we have a metric that we're trying to raise and in this case it would be conversion to a subscription How Could You attribute what the developer did to conversion and have it be isolated so that you know that with some level of confidence that it is um helped by the thing that they're pushing I've had this personally kind of pitched to me as a product manager because you can get a lot of like you can basically like see the changes directly so every week at Delivery Dudes the company that I talked about at the start of the video every single week we'd have this scorecard and you obviously have your funnel and the type of people that go through your funnel are either first timers so fto first time user experience or eux which is existing user experience and what we found is by basically splitting these two across all the different platforms forms so you'd have a funnel for web and mobile web you'd have a funnel for desktop you'd have a funnel for iOS you have a funnel for Android all these different funnels and each week in your scorecard you're breaking the metrics down by the number of people that are coming in to the number of people that are accomplishing your goal in our case it was selling the food getting them the food delivered so across all of these you'd have eight different conversion metrics and your job is to reduce the number of steps so Android and iOS were clones but you'd basically have all these steps where the top of the funnel was login slash sign up for a first- timer so you want to make sure they're getting in there really fast and then you'd have the uh basically like a a home screen so home would be something like this where you have them logged in but you're not quite sure where they are so you need to be able to get their location data but if they don't do that then you're kind of out of luck so it's like how do you get them to do that so that they can get the best um suggestions you also have to make sure that you can show them stuff in case they haven't typed it in so you kind of have like these dummy datas of some uh like featured restaurants and whatnot but you have the homepage and then after that you have the sorry the restaurants page after the restaurants page it's the menu page after the menu page it's the cart after the cart it's the checkout and then and you might find that oh we split test something with checkout two because checkout one has no payment details on it it's just address book and then checkout to and so you have all these steps right so for a first timer you'd have four five six seven steps and this same thing would apply to all the different funnels a existing user experience would have obviously less steps let's say like 10,000 people are coming in and the supplies the open source thing and the profit share that I'm kind of interested in just ideating on but if 10,000 people come into your Android app and they're first- timers and only a th get through that'd actually be awesome right you'd have a 10% conversion rate what we found across mobile was it was somewhere in the camp of like 10 to 15% and then you're doing this across a week so you know Monday Tuesday Wednesday Thursday Friday Saturday Sunday and it just is that average so we know that in the food delivery space and also in ride share that you always see a big spike towards the latter part of the week and then it comes back down and the lunch business for businesses and corpo business is kind of the stuff that like keeps you running in here the steady Ed because they're running it on someone else's credit card but that's a an interesting aside just so you guys kind of understand where my head is at because every week as the person that owned all these metrics you know you get together with an executive team and then you have an owner column so it just say Parker like 10 times and then you'd have you know week one week two week three week four week five across 52 and you'd get this trend line where it's like oh did it go up oh did it get down but an example of this would be if Parker gets it on the one that generates the most amount of cash the one that's the most important to the business let's call it iOS because I think we did like half a revenue on that if we could move iOS two points 2% not two points but 2% then what would that mean to the bottom line how much is that going up so that's just something that I I haven't seen in SAS I think it's probably because the majority of SAS companies are Venture backed so the incentive structure just isn't in the line with something that you could do here but that'd probably be one of the stickups is like you know VC prevents creative incentives especially around profit and then I think the second one would just be attribution so if you had more than one going at a time or you had one go in but it had nothing to do with the actual effect that it had on the business because some other thing happened that might be what prevents it but I think this is be worth testing and at least at minimum it keeps me interested in the idea of making it OSS so to answer your question yes I think that this would be an open source piece of software I needed a little bit more thinking on it it would definitely move things forward it looks like our results just came back from the deep research that I had so evaluating the open source transition and Inc of structures for an AI powered SAS product and again I just gave it um a prompt that basically says I'm contemplating doing this here's a one pager about what the thing is and then I said I need to find some examples of successful open source Technologies and give me 10 examples uh with their origin stories Founders go to market strategy pricing struggles Niche so I automatically see Linux okay so prospective transitioning a proprietary status product focus on goal setting person development through an AI coach to an open source model presents a unique opport Unity for inovation Community engagement that's a big one too I think that Community engagement alone is worth doing open source because it goes into this macro Trend that's going on where as like let's take you know 2000 we have 2010 2030 and let's say the left or the Y AIS is the cost to develop software right so if we take this and we just plot the trend that's going on it's just been crazy right like you can also look at the salaries so I'll do salaries as well salaries and the salaries really like just this is going to be correct for a pattern but obviously I'm not going to put numbers on here because we have a dual axis but the salaries go up and up and up and up and up and up and up and then we have zero interest rate phenomenon and it gets to like Q3 of 2021 we kind of hit this local maximum we hit this like Peak and then it's slowly started to come down and I think it's just going to continue to do that in terms of salary and then we'll find some comfortable number because the job will change for coders and then that will stay there and the amount of like basically the cost for development was very very high and then it slowly come down and it's fighting against the salaries but there's something going on where the people that were high salaried are worth it because they're getting more leverage out of the tools and the Technologies and the libraries and extractions but I think that that cost developed software is just going to plummet so point being is that as software becomes more deflated in its cost and it becomes more of a trend where if you're a company and you know you need to get Salesforce well you may think twice like we were a company doing 50 million a year and when I went to get Salesforce they wanted 30k implementation they wanted more money than you could ever imagine just to have a basic St so we ended up like rolling our own and we rolled our own CRM with a bootstrap budget we did it in a few months and that was around here so that trend is definitely going to go crazy where people will roll their own stuff there's going to be lots of like Niche specific slash like Boutique software that solves very very specific use cases so I think that as this trend goes on then the other Trend that's going to go on like if we thought about just the power of let's call it like brand which is community the importance of brand Community has always been important right but it's just going to go wild like I think it's going to go parabolic because it will be what differentiates it it's really like you know deodorant is a commodity but people that have a really strong brand and they speak to the audience of oh you don't want aluminum in your armpits like they're the ones that win because they've branded around that they've told a story around that for a really really long time so I think community and OSS is going to be such a GameChanger so I've kind of talked myself into it let's take a look back here so Community engagement this report analyzes the landscape let's see let's look at some of the examples is there anything else particular focus on the challenges of implementing a profit sharing based model work on conversion rates it might not always be conversion rates but that's just the most obvious one that I think of like you could also do a model of a price per feature and just make it dead simple with small medium large t-shirt sizes as they do in agile so we have Linux obviously they did organic growth through developer adoption it's free they monetize via distribution and support services don't know what that means on the first one there was some nerd battles foundational OS so these ones I want to find one okay how about WordPress it's a little bit more similar free open source big community of themes and plugins WordPress might be a good comparison because WordPress basically like it's kind of like a puzzle piece of like a not maybe not puzzle but like if you had WordPress the value that it has is obviously like ease of use but then the customizability with themes and plugins extensions all of these things allow it to then easily connect to all these other services whether it's like CRM or CMS that you like more but that's what makes WordPress really helpful for people is they can connect it to all their stuff really easily like my like 78y old dad has a WordPress site and he's like oh I can figure it out so it connects to all the stuff it's uh usually like bidirectional and then if you ever wanted to leave this thing like that's going to be a huge pain in the butt you already have all the stuff on there's entire companies that just focus on migrating away that makes it really hard to move so I think the comparison would be with map if you're unaware it's basically a five inone productivity suite for become the best version of yourself so you would have your notes all the daily drivers you have to-do products or uh projects you have your health dashboard which I believe needs to be in front and center in the stuff that you're looking at every day otherwise you won't think to open that app that's Gathering dust you have your calendar and then you kind of have all these little like sub agents within here so maybe an editor summarizer you have your prompts you have the things that you are naturally doing like whatever task that you do in one of these manually has its own agent and the problem that I had when I was doing this before was that it's just so many agents that are working in unison that all ladder up to this one that's like you know future Peak Parker like that ideal version that it made it really challenging to kind of orchestrate all these things but how it ties to the WordPress analogy is that you have this going on this big old piece of software and there's a bunch of third parties that you want to be able to connect it to because you'd have this operating system of your life going on but guess what like I want to have whoop plugged in to the health and I rolled that myself that's fine but then oh I want to have apple Health done as well it's like okay well we need that one okay got it and then it's like oh wait no Aura what about Aura what about straa have you thought about getting um a gar and that same thing can be said for basically everything else here right so this could be kind of the bread and butter for what OSS folks could start with is something that is just connectors and then if that does work then we gain some confidence in the model we feel like it's good and then you can start doing actual like core feature development because these are all pretty useless independently like I don't think they're generational uh companies but I think together they're really powerful but getting that off the ground is a challenge oh so thanks for the question next up is some news so not a lot but the kind of core concept here is that China keeps releasing stuff that's better than ours so by do this was a company that I was watching maybe like 10 years ago because what I found when I was designing a lot of e-commerce stuff like we're selling food but it's basically Ecommerce you're doing Commerce via food was I'd look at the most popular website lists in the Commerce category globally and BYU was on there like I don't know how I don't even know if they are e-commerce I think they've pivoted or something but maybe they were a search engine but the long story short is I came across this company and you can just see that like they just ship right so they they launched this thing Ernie which you can't even use if I went to the website right now I'd have to do a Google translate on page and then the translation between cantones and uh English is rough like I played around with it last night but the point is is that uh they just released something that's on par with GPT 4.5 which just came out but it's 99% cheaper so curious to see how uh the stock market's going to look tomorrow when the world wakes up but the last time this happened I think there's like a $600 billion do crash or uh deletion of Market on the Monday following the Deep seek R1 launch and the reason that happens is because the top I believe it's something like if we if we think about kind of Market distribution if you have this big old blob that represents the S&P 500 S&P 500 which if you're not familiar it's a basket or an index of the top 500 companies in the US and so if you look at this I'd say that the mag 7 which is the big tech stocks they make up I believe it's something like 80% I don't want to necessarily look that up I could but it's like something like this where it's like those companies represent the vast majority of the market Market in its entirety and what's interesting about that is those mag s companies are all pitching this story of ai ai hype right so the question is like does this mean you know that it's necessarily true that these companies are worth less money that their market cap is worth less when someone comes in and totally changes things up because the big ones in there is Microsoft like two trillion dollar company and they had a deep partnership with open AI you look at the other ones like meta and what they've spoken a lot about you know they pivoted heavy until using llama and all these models and they're open source but they're way worse than open AI from uh basically every perspective maybe not MH gen but when open AI says it takes you know billions and billions of dollars to be a to build the you know GPT 4 and then a company out of China comes along and they're just like I don't even know if that looks like flag and all but if you have the us over here and it costs us like 80 billion doar in seven years to get it done and then they come out and they're like okay uh it's going to be six million and it's going to take us uh maybe six months part-time like that just totally boggles the minds of every single investor so then what you see is just like typically if you have let's call it uh Day Zero and we'll flag this that's a flag apparently we flag this is an event where deep seek launch and you could just take this and say like any big Chinese launch it's typically like if this was launched then and the market reacts like three business days after right so let's say like the normal if the price is steady and it typically goes like this just a big dip and then maybe on the fourth or fifth or even sixth day after you have another event that happens which is a news cycle of all the investors getting on new shows and saying oh this is not true like there's no way they're lying whatever and you know it's kind of this like recovery tour to or even like someone call it like a smear Campaign which is kind of what's going on with another piece of news where open AI just basically came out and said that China bad like these models should not be able to do anything and they're so bad and they're State operated well guess what like that's incredibly ironic because you're closed Ai and you effed over a lot of people along the way maybe even killed somebody so don't demonetize me YouTube for telling the truth next up is the idea that you basically want to just focus on what won't change so when these models get better and better and better you need to ask the question of like what won't change over time so that you can be positioned to win so this is a really good good question to ask and it's what Jeff Bezos said so he looked at Amazon and if we look at Amazon in 200000 that's when they kind of got things going but he asked a question like he first of all he saw this trend that was going on so he said you know internet growth is going bananas is going up and right and that's the thesis for why he quit his job on Wall Street which I feel is really similar to what I'm doing right now where AI is going bananas Main Street has no idea about it I need to be a part of this full-time like this has to be the thing but he basically said internet growth is going bananas it's something I need to invest my life into and it will change every single day but like the Technologies and the pace the wars law like it's going to go crazy but won't won't change if he's selling books to start is customers over the Long Haul will always want the following so what won't change they will always want lower prices they will always want faster delivery times faster shipping times they will always want more selection so these guiding principles brought him from being the most tumultuous Crazy Fast changing grounds moving every single day environment to making it so that I can order these markers from Amazon basics for you know 15 cents or whatever and get them tomorrow in the morning if I want so those same principles apply to where we are right now and drive what I'm doing with these videos by putting little internet trees little statues on the internet that will live forever is what won't change is people like basically what won't change in AI era what won't change in the AI era is people like doing business with people they like that is a fundamental truth so in my case you're selling services if you're selling software if you're selling anything you're much much more likely to do the business with the people that like you and it'll also on your side person the vendor the the seller of whatever product or service it'll make it far more pleasant and enjoyable because it's not an uphill battle because there's something going on with these videos that I'm doing now and it's called the parasocial relationship so I don't know you but you know me when you reach out you know how I think so I think that this becomes more important that you are able to be human you're able to connect so we know that people will always continue to like doing business with people that they like they'll prefer it they'll actually take a hit on the service they'll pay more they'll do all sorts of things to do business with the people that they like another thing that will not change is businesses will always buy something that either saves them time so they they businesses buy things that either a save time B save them money see sell more stuff so you either are saving you are making or you are getting them their time back now one thing I will say and this is what I want to lean into in the services is you have more leverage as a seller of a product or service if you're close to the money and so if you're just focused on cost savings you can only save so much as is relative to your total operating C like cogs can only save so much whereas there's much more leverage in selling more right and that's this other bezos's quote where he talks about baseball he's like the coolest thing about business and we're going to draw a little baseball field is when it comes to invention and it comes to selling said inventions in baseball you can only hit a home run that gets you one two three four runs right if you had loaded bases but if you have something that you're selling that's high leverage and it's a new invention and it's likely using AI you can run around this thing multiple times and get crazy wins so there's just it's exponential so these are some of the things that I really think about whenever you start to get anxious about all the things that are changing like this is what I come back to is you want to be likable you want to make sure you're leveraging the technology to do some of these things next up let's talk about offer creation we're going to do a little bit of strategy so I started a paid Community then I turned it free and I was wondering well what is it that I would even put in there I need to open it up but I also recognize that you're going to get a lot of like low quality people that don't engage because they don't assign a value to it it's just another free thing it's like if you got some free product when you're walking around outside like you likely just go to the trash can and throw it away but if you paid money for that same thing you wouldn't throw it in the trash can because you have assigned value to it with your dollars so now what I'm thinking is okay well what is a good offer and this applies to you and your business of like defining what the offer is and then defining a price so this is just something where I'm thinking okay well if I have this paid Community what are the things that I can do in it I think it's just obviously having you know let's call it like two posts a day about things that are working what's not working some sort of summary about this right here right so I'm doing these long form pieces of content there's going to be an hour of content per day on a daily upload Channel I want to be able to find the most interesting tidbits in here that would be helpful for people that are trying to it's basically for Main Street to know how to leverage AI so I think these will be ultimately automated with a human in the loop but some sort of like nuggets so I I promise to them that hey we're going to get fresh update in here if it's not helpful I'm not going to just Spam just a post but I'm going to try so there is like some sort of filter there I think probably maybe like a monthly Q&A q& a maybe a [Music] monthly mini course so just a couple videos that I basically do stuff like this except instead of posting it on YouTube it's behind a wall other things in here could be the prompt library and this is only valuable if you use the things because this is there's such a high quantity of prompts on the internet but think about the person doing it are they using this to make money probably not they probably are just posting it to get likes so probably a prompt library and then there'll be a sliding scale of pricing so start at 19 a month and then lad it up to 29 ladder it up to 39 and so on every call it 50 members so that's the offer I'm thinking about and there's also pieces you can do to gamify it so one thing I need to figure out is like on the other strategy bit with school is what does the uh gamification look like what can you unlock but it does start with these kind of core things and then we can have one-on-one calls as an option for purchase they'll immediately get access to two free courses that no one else gets which is one is on product management and scaling a company to $50 million a year other one is on just like an AI 101 slash third one I guess R3 is just on chat GPT so then I'm thinking in terms of the strategy to get folks in here and get them aware obviously I'll have it in the description of every single video but what I may end up doing is like if I have a master class then there might be some sort of spot that's in here it's like hey like just so you know check this out and then hey just so you know check this thing out and then at the end if you want to learn more you should go and do this so that's what I'm thinking I checked out this other platform it's called and all I gained out of it was that they are uh they like cool marketing the product is way better in terms of the quality of it like the UI looks better there's a lot of modern stuff but the ux sucks like there's an ism in the UI ux World which is i' rather have a chair that's really comfortable to sit in and looks really bad than a chair that looks amazing but sucks to sit in so UI is the thing that you see ux is the thing that you feel this product makes me feel old it makes me feel dumb I even talked to some people on your last night and uh it just they got some work to do there's also just the monetization side of things is they lean very heavily into this idea that you never would pay $99 a month for it but it's if school's 99 a month and there's no processing fees for payment well it's pretty quick that you start getting hosed if you use because they charge 3% plus 3% so that's 6% right and so at what point do you cross the threshold that you're paying more than $99 a month well it's around like $1,750 so as soon as you start making any s months of cash like $1,700 a month is not a lot of money at all so it's like if you're making 21 ,000 off this platform annually then you've already gone beyond what the cogs are not to mention that trouble are not trouble-free but that school has doubled the traffic so your chances of having discoverability are much better I also think that the I'm not sure if they're VC backed um VC round pretty sure they are yeah okay they raise 17 million do in a seed round they raise a lot of money I don't think that school raised any money so I think it's less of a risk long term to do School uh other piece of strategy is just the separation of Niche by channel so what I've committed to is this power list by Andy forella which if you're not familiar he did 75 hard and his power list just says what are the five key tasks that you need to do every day over the next 90 days and one of those for me is doing this daily upload what I'm realizing is that I talk about a couple of different things across my brand so I talk about the AI for Main Street which is a much larger market and by Main Street I mean like noobs like the Gen the general public I talk about AI for developers very different it's still in the same niche as AI I guess you could call AI a niche but it's also tough because you could say health is a niche but that's it's not like there's Functional Health there's traditional Western Health there's working out there's diet there's all these things so after a couple weeks of doing some testing on this stuff I really think these two need to be broken out where it's like the Main Street stuff would be on one channel right so you have the Parker Rex is the main Channel and those types of videos I think would be for getting the most views it's the widest audience it's stuff I'm looking at this board it's like five ways to learn AI like that'd be much different than the developers who are like dude like I know this I'm not like trying to learn this they're going to want to learn stuff about development tactics and using AI coding tools and then the third one is the you know the Parker XX daily which is just for people that want to see me building in public it's also uh like a much more loyal audience that you be garnering and you get the questions so I'm thinking about the strategy for this where it's like okay I've started a second Channel and know that this content wouldn't work on the first one do I need to have a third one just to have the separation of it it doesn't take much more operational drag because I'm just recording the same quantity of videos It's just separating it out but that's kind of what I'm I'm thinking would work and it plays into this last part that we're going to walk through which is called the content scenarios and what I haven't seen is anybody do a good job making a Content system that doesn't produce spam so I drew this out yesterday and I'll show you in a second but I could reply like basically if I unlock this content uh pipeline where I take the long form piece of content which this videos about 40 minutes right now but if I can do it for each Channel where I have long form go in and then it turns into uh you know the YouTube shorts it turns into IG reals it turns into Tik Tok so you have it broken into those videos and I could probably use Opus clip the one I'm thinking of using to begin with and then you have the second prong which is written so you have blogs and this is the one I'm focusing on each one of these is an Automation in itself blogs you have LinkedIn you have Twitter you have Facebook and then you have images images video and text not red and that' be IG that'd be what's another one Facebook so some of these aren't actually represented by like the final post that you're making but it's like I'm thinking in terms of the actual pipeline so what are the like for for images on IG you know you'd need some sort of Automation and then it would come back together with the text that you have right so then you end up with the actual you know post that's going up or the the image or the video so I think that this is all possible as I get these systems working and they'll they'll definitely have like stop gaps in them where I'm like no don't like go forward until I take a look at the review and that would be done through something like uh click up so if I look at the one yesterday went through this and basically broke down uh where was it yeah all the different steps so I made this this automation I think the video is going to go out on the dev Channel but it allows me to go into Premiere Pro right after I do this video I uh will export it to a drive the drive is being watched the drive fires an HTTP request quests that then hits a postprocessing audio service called aonic aonic does two things one is it processes the video to make it sound good and add chapter markers a summary the transcript from open aai using whisper and then it posts it to YouTube and then I need that stuff because then I can use that for text so then it it stores it in Google Cloud uh Google Cloud bucket with four files it'll have the HTML show notes that has markers summary transcript has video it has uh subtitles and then ultimately it gets put into a database using just literally Google Sheets so I can see it all there and that Google Sheets is being watched by clickup so I can know where the status of this pipeline's at and then I use chat gbt with prompt to basically turn those into blog posts that are written in HTML and then that gets posted back to the database so I can see it make sure the output looks good and then it'll go to super base there's options for other stuff like as I expand it I might want to have image generation on the blog post using a perplexity style thing with flux 2.1 you can use a summarizer uh voice mod with 11 Labs where it could actually be trained on my voice so when you go to the website there' be a little button at the top where it not only has the YouTube video then it has the um or no it have the generated image from flux one it have the YouTube video that The Source material is from and then it have an audio version of log so these are all things that are possible it's just a matter of getting them dialed in I do think this could be a product on its own so I'll be the beta tester of it I want to leverage it so that I can get these digital trees planted everywhere on the internet and um see apparently my thing doesn't want to click but for now it's just doing the two um and I'll end with this so there's each of these automations that you can do with make or nadn can be really their like I'm I'm starting to learn it's kind of like its own app where you don't see it but it's this magical thing that's going on and Hemingway editor is a really popular text editor that lets you get a on the writing and the readability and basically the quality of the thing that you're writing so I was able to pull together a deep research from Google on how that works and that would just be another step so I'm thinking a lot about like how to kind of position these automations to be a little bit more complex than just like crud plus AI I think that if you could put in an actual node that has like a Lambda in it or something like that you'd have to probably use n8n which I found out yesterday it allows for stuff like this like you can do shell commands in there you can run um lambdas which would be sick because then you would be able to get much higher quality stuff so that is it for today if you guys have any questions let me know in the comments below you can comment on here check out the school it's going to be paid by the end of today I got to do a lot of work to like let people know in there that it's going to be 19 bucks a month um but you should check that out if you don't want to buy anything from me that's cool too you can just keep getting the free value but you should like the video that's your way of paying me hope you have a good day and I'll see you in the next one