we forgot onboarding and got crickets

00:00 · cold open

00:00we covered it within like maybe like a week we covered like every single page every like little setting everything and then i was like all right guys come on back to a platform and provide us more feedback crickets crickets absolute crickets

00:19 · we're back (four months later)

00:19speak here from 2x engineers and i'm joined by the wonderful the mighty the destructive and hyperbolic darren how are you doing today sir day been busy hot busy with a lot of the oh yeah hot in southern california but definitely busy definitely busy um yeah it's been it's been actually a crazy amount of time because from the last time that recorded this uh i think it's roughly around four months close to maybe maybe three on the edge of three but four months since we recorded this last time we we were saying a bunch of things i mean uh watchers can even tell from the length and look of my the roots in my hair it's been a while since we recorded last because my hair is longer but it's still on the bottom yeah so let's let's comb

01:14over a few things that have happened since um i'll how about this we do like a little um back and forth we say one thing that we can think of that's like a big thing

01:24 · azuki regionals recap

01:24that's happened i will start we went to azuki regionals fantastic oh yes we did we definitely did it was it was a fun time i got my crap yeah you did well you did very well matt thanks and i did uh how'd i do i did okay you did okay but like at the same time let's let's just be real let's not get too deep into this because i want to i want to kind of cover all the things that have happened your your brain shut off and you know it yep yeah it was like around like round six or seven around six or seven literally where your brain just didn't do math anymore i just like i can't i should have i should have brought like a snack in the mid mid section or something i don't know i don't know why people start regional things in the morning and then just continue to a late afternoon they

02:19should just start them like after lunch yeah like one to nine just going from like one to like six or one to seven yeah right because all the big games are your like first four or five right that kind of like basically determine like you're standing in the leader in the leaderboard so then it's like okay you're not starving most likely because you're finished lunch and then you just go through like six seven hours and then between the six seven hours you then decide if you want to scoop because you didn't do that well right which a lot of people just dipped yeah because because they're there to play for the cards yeah play for the pricing yeah like if you're a competitive player it's like what's the point yeah like just you rack up like four losses just dip yeah you just dip they just scoop and then you just go eat dinner right but i don't know

03:15there's another another thing probably because of staffing right i guess but then you know court uh the company that was for the card shop that was sponsoring it was basically core tcg down in pasadena so you know staffing was always an issue probably right to figure out who's going to staff it yeah it was my first event ever in a we'll call it in a regional sense because i've went to like local tournaments before but i thought it was pretty well run but i'm also not one to really know if it is i i thought it was fine i just think that the tables the table numbers shouldn't have been on like this piece of paper there yep and it was like super low you literally have to go to like every table just to look at it versus like having a little stand yep and then like if they're i know core does multiple you know card game hosting things and it

04:14was kind of strange that they just printed like paper like why not just get like a laminated paper and put it like a metal stand and just reuse that on like every regional yeah and you don't have to wait or print paper every single time yeah and the tablecloth situation was kind of weird yeah but that's okay yeah overall fine judges judges did judging um but that that's that's my item i want to talk more things azuki but like what's another one of your big events that happened in the last four months i think the the launch of

04:47 · big & pando market arcade are live

04:49big big big it's just big.com check it out it is out we got yeah primarily it's a platform to help people grow it's around the creator and small business space there's a lot of uh like like tooling in terms of like i guess like a beacon type of thing so where beacon and all those other toolings don't shine big allows a mcp connection to allow your ai agent to connect and to basically go out there and like tell big like which automations and things you want to set up just through mcp or cli that was like one of the things i was kind of getting annoyed with with like mini

05:44chat type form all these different platforms and a lot of those things just were like you got to pay in order to play i think the only competitor that doesn't do that is beacons but then beacons doesn't have all the i would say core functionalities that i would want as a creator or small business which is the reason why i decided to work on big with you matt yes yes and not just that because um whole you you mentioned it very slightly but you are you are in technically both categories you are creator and small business owner right yeah they're like these platforms cater to one or the other typically and even in our space because we're just starting we want to build it

06:40from the ground up like big has a a vision that is larger than itself but we understand that it is starting small and we want to work with the small creators and large creators as well as small businesses and large businesses so we want to build them to build in features that fulfill their needs and we will build it according to like what they want as well as what we think is like current serves like serves this thing yeah and is agentic forward right right so it gives you all the bells and whistles um we're currently working with some i would say mid-sized creators and small businesses underneath the seven digit i guess revenue but you know just primarily focusing on small businesses first just because

07:36you know you know that's kind of you know that's just like a e-commerce business where you know i built that to also utilize big at some point so everything kind of the tooling and everything like that is like for my own businesses right so anything first is because it's a direct need as myself as a user so the feature set and also the i guess the feedback form also allows users on the platform to see what we're building and also decide what we're building as well they can put their own suggestions and i believe they could upvote and downvote and maybe we could do something interesting with people on the platform to decide right we can maybe give away like oh if your feature gets selected for this

08:33month maybe we'll give a month free for like your monthly subscription or something like that i don't know we'll think of something if if you just decide to use that feature if it brings value to the platform there's no reason not to reward our users for it especially if they're involved yeah yeah exactly i mean even in that same vein because we just talked about azuki the panda market sim is actually also officially out because it wasn't before we were talking about this one oh yes right yeah that's why yeah we were still we were still building it um actually yeah i mean it was it was usable when we were yeah but it was just usable yeah the reason why is because like i didn't want to release it and like allow people to come on to it until it felt right then you know again like we're we kind of dog food or like we test our own products right like i try not

09:30to release anything until it feels right just because i you know we worked at plenty of startups where like they're always like oh yeah speed speed speed yes speed is important but also execution and at least having a good feel of the product on top of speed on top of right yeah so i'd rather have a good product than have something that just releases and then it doesn't feel right or i can't use it right so that's the main thing no it's super fair um yeah we've definitely been burned by that more than enough times both together and separately so oh yes that's a that's always a fun one let's see what other fun things have happened in the last few months um i know there's another thing that i

10:22 · custom builds for creators

10:22mean there's no i mean it has technically started but it hasn't grown to anything large yet but we both we ourselves are are partnering up with particular creators to help them expand their business with custom solutions yes and that has also happened yes that's happening and and we are bringing them onto the big platform yep essentially it's a big big build right so we do big custom builds in terms of uh for small businesses and i guess your medium-sized creators are basically small businesses at this point day and age right the creator economy every mid-size to large size creator that decides to monetize is basically now a small business right and all these small businesses in today's age are trying to figure out how

11:17do i operate operationalize my their businesses to get things streamlined right to basically free up more of their time and also potentially not have to hire uh so many people va's virtual assistants that is or um other assistants or um other assistants or assistant directives or what they call executive assistants right so if they know how to use ai and or if we could streamline streamline or automate a bunch of different things and provide services to help grow and make their businesses bigger what big um that's what the goal of big is intentionally yeah so big itself as a platform is meant to be used by everyone but there are people that have very niche

12:14and specific uh needs and we do want to work with them to help them expand as well so yeah we have we have we have at least one person for sure right now that we are building something out for um pretty pretty curious to see where that goes and even in that one um this one engagement that we have we've already expanded on what three new integrations that didn't exist before yeah but i think like the reason why we wanted to put it in big is that you know if actually these ai agencies or other services-based business in the software automation or ai space will do like custom build-outs ones um the reason why i or we've decided to not do one-offs is basically because small businesses

13:12i've just noticed just like talking to a bunch of them they don't have the time and or resources to go learn how to manage or maintain these yeah types of systems right so if you work with us big and you decide not to do the one-offs and you know bring your business onto our platforms you know people use it go high level they use like a bunch of like different places it's kind of like that where if you come to our ecosystem basically is that we provide you the services to allow these integrations but then we also do the upkeep right and if there's any issues or problems for you as one of our customers are subscribed or onboarded to bigs platform then we basically have our brains that are there

14:11to always make sure things are running smoothly for you and your business yep and that's kind of basically like you always have i would say a seasoned team of engineers at that point and i think that's yeah i have a say in like where the product's going too yeah i think that's a good trade-off right we don't charge too much i think we're pretty affordable so yeah a lot of the pricing in the space is um i have big question marks whenever i see it it's it's pricing based on i believe what the software used to cost correct which is not that problem but the problem is people outside of the tech space and or tech twitter or just tech in general they don't have the bandwidth to go in there and to learn

15:09basically how to be a new age operator yeah with ai and also understand the tech side

15:19 · opus 5.5, fable & tiny models

15:20basically uh software development yeah there's um i mean this is like also a neat segue and the topic to cover but the ai models in the last four months have skyrocketed because this is a this was a pre-fable era when we recorded yeah it was it was a pre-fable era and pre-opus 5.5 yeah pre-opus oh my god 5.5 so so where we are so six two so six and opus 5.5 yeah and sonnet 5.5 yeah but 5.5 and six we did have opus five though we did have that for a little bit and that was uh yeah everybody i'm still using it but i mean five versus 5.5 is basically like what was it four and 4.5 yikes like when you saw when you saw four

16:17you thought four was good and then when 4.5 came out oh my god that was a game changer yeah and then now it was like the same thing happened where like five was like yeah questionable but it still did its work i had to force my models to be 4.8 i didn't like it at all it was like that bad yeah if i don't i don't know if there was like better token usage on five or more or less the same i don't really know i don't remember but yeah i just we're here yeah we're at 5.5 now so it doesn't matter and it's only going to get better over time yeah when uh when people say like when they're still against these ai models for any type of generation like image video uh code word or text right everybody was like it's not that good it's not that good i'm like okay but it's

17:16pretty good like before 2026 2025 it was okay it was okay right it was okay yeah and that's that's already saying something if it's just okay and this is only year one or a year half not like year one yeah like just imagine where it's gonna go in like three to four years it's just a matter of updating the model and getting those like incremental 2x 3x productivity boosts just like that without like changing any of your pipelines or workflows or whatever you might change your workflow to even be more like diligent in specific aspects right but it's like night and day when people don't have their pipelines and they have to start from scratch to figure out how to

18:11use things versus like having your automations having your workflows having your pipelines having like your like orchestration pipelines for all your agents right like having your skills and processes is like crazy yeah but sometimes even so like when you have specific skills sooner or later maybe the frontier models might not even need those skills they won't because it's just naturally so good like like right now there's plenty of skills right that you needed to remember like a specific thing right what happens if it gets so good right these frontier models where at some point in the future you don't even need the skills like there is a future where that could be possible yeah very very useful and we've actually

19:09we hit a very um a detour we hit a detour in the agentic um uh adventure of the world i'm trying to bad words but now we've entered the world of like having specific models that do specific tasks and i want to note on uh jeb jev and laia oh yeah district they strictly just classify things they're absurdly cheap as well as absurdly fast yep i haven't used them personally i think those are like certain models are your specific use cases that you can or you could place them into your pipelines or workflows i could see the use cases but we just have it implemented or you know spend the time to integrate with all those things because we have just so busy with so many things yeah

20:04i mean i've i've built my pipeline my my local um local hosted pipeline software it's called right now it's just called flow but i'm still working on it and i'm trying to work specifically with laia because i i am still very much bullish on the one day i'm gonna just need my own local models and i want to just be prepared but i i'm just i'm very excited for where where things are going i mean also shout out to pewdiepie with ajax i'm gonna try his model sometime soon yeah he was he was saying that's a little brain dead or what is it called like not not there yet but it's it's there yeah but it's not the greatest model but it's there he was saying it's a little dumb but it's okay it's better than nothing yeah and it's only gonna get better i can only imagine what it's gonna be like in six months he was trying to train it with i think

21:01chat gvt data our responses yeah yeah band band band one and then maybe band again and then he was just saying on his video how come you guys can train on other other people's work but we can't train on your work so he was just kind of throwing shade that way fair man we're we're hitting this weird weird world where even in the game space um so there's two major things that happen in the game space and i'm not sure if you've seen this but one big one they've almost effectively completely ported over the functionality of the ps5 to work on an x86 so they're just getting around needing physical anything and just having it and we all know this is with ai two people are taking um complete games and like merging them with other games like oh yeah i see i saw

21:58that yeah i saw that the elden ring related like moddings yeah they're like modern games together i was like amazing but um the whole uh yeah the bloodborne crew they're like uh some guys are making a proper port of bloodborne apparently it's running at 120 fps it's like almost ready beautiful so like gamers are gonna be eating real good soon which makes me very curious uh for for where things are going but yeah i'm excited i'm more excited for the indie gamer space yeah versus like these triple a agree um triple a audi bot companies yeah shade on ea yeah but yeah yeah i mean it's it's all weird i mean i started um started my own game last night literally just because i had way too many tokens that i needed to

22:53burn i was like well i might as well start and this is where i have to say that the ai has gotten terrifyingly good i thought i was wowed when fable came out because it was good at design like just hands down it was pretty pretty dang good at design without many references opus 5.5 i have been like running it into the ground almost every day full design passes making have it having a build full things features my i feel like it's tickling my token usage right now and it feels like it's better than what fable was like because i haven't even touched yeah in like weeks yeah i don't think you need to touch fable unless it's like super mission critical like very intricate detail that you need to like have that fine comb to comb things through yeah yeah because like funny yeah funny story there was a

23:51 · the $200-plan token burner

23:52community member in a different community where uh she's not an engineer okay she's kind of like a tinker five coder type of thing trying to like operationalize her business okay and then she was asking in the in the ai support channel i'm token issues and i'm like oh you're you're probably on the twenty dollar plan right okay yeah that's that's kind of normal right like every two three hours you'll run into like token this use no this this chick was on the max plan the 200 i'm like oh well oh you're probably you're like on the you know the 5x or like the mid the mid-tier plans for like these frontier uh companies right no she was on the max max i was like wait max max you're running into token issues there's no way that you're running yeah i was like what are you doing that

24:47you're maxing out your tokens when you're not building custom software and you're not pushing like work on like a startup like you're just operationalizing your systems what is what are the tokens that you're consuming yeah that's basically what i was asking here and then i was like reading her operations and what she was doing i'm like this doesn't make sense how are you still maximizing it's not like you have like thousands of clients that you're like doing just like operationalizing like your whole business for like thousands of people and i was like that's okay it still doesn't make sense what are you doing and then she said she was on astra and i was like wait astra you mean you mean the highest tier model astra and then she's like yeah

25:44i was like no you don't need you don't need astra girl like like you're burning tokens like like you might like between soul and astra in terms of defining and understanding context from a text yeah or like a paragraph of things yeah that needs soul probably that like anything text based you don't you don't need like a high tier model you just need the mid to low yeah you don't need anything highlight for that so she was basically just wasting tokens i thought you were going to tell me she's like fable ultramax but effectively it's the same thing yeah but just on a different model i don't think it's just ultramaxing but maybe medium effort or something i don't know whatever effort it was she didn't need that action i basically told her like you only need to

26:42use it once to get that mission critical definition of how to do things and then after that switch it to a lower tier model and basically give the handoff from astra to your lower tier model and then let the lower tier model build out that automation or workflow and then from there you don't even probably need the mid-level model to do the workflow no automation and you just say like okay i need a new definition so you could be more clear so you don't make this mistake but then again this is this is the thing where like people are utilizing ai and they don't know how to use it and then it's like a learning process

27:33 · coaching vs building, and why big exists

27:33which oh i guess we could segue right now matt into the whole like teaching yeah teach not not like i'm sorry um i got building versus coaching yeah coaching versus just building and providing services right like for example if i were to help operationalize her business she wouldn't have these issues and or have to think about or learn how to do these things but she kind of wants to be more hands-on right and there's not that many people that want to be that hands-on as a business owner right but in like what i thought was what business owners because this is what as a business owner myself this is what i would do i would have to learn how to utilize ai to as fullest

28:30potential so then i could then hand off to an assistant like a human assistant to then manage the ai at that fullest potential sop wise or workflow automation wise to basically supercharge my employee basically but then that has to come down top down which is like basically was changing the whole tech industry in my opinion like if anyone's a small business owner out there i think they need to pay attention to what's happening in the tech industry yes there's a lot of layouts that are happening but why are these layouts happening it's because now that ai is in this in the scope of place right it's so new management doesn't know how to use it and if they want to be able to manage their

29:29employees they themselves need to know what the capabilities are for ai to then become like kind of ic slash manager types to then actually be able to manage people the right way right so then you kind of either having too much middle management where they'll just remove a lot of middle management just basically lay off a bunch of people and engineers because they don't need that many people because then you have ai now right so then that's the whole thing where like engineering product design project management all those four industries are basically combining into once super teams where you just basically need a team of maybe like four or five to be like a supercharged ai team yeah like

30:26and then one manager of the humans that basically manages like a fleet of agents that know how to do things at like at like a very uh competitive i competitive level yeah that's i mean that's yeah i think it's hard because even i guess not in the same space of teaching or building even when it comes to growing big because we've we've had people on the on the wait list for quite some time that have stated that they were excited for it that they also want to support they also want to help oh right right yeah yeah yeah yeah so what happened it wasn't launched it it was basically i did one i built it for myself yeah right and i asked you if you wanted to help me build it and like basically do

31:23like like basically a startup yeah right and basically grow it and also become like a service-based business as well but then and i wasn't the only one that had these issues i was asking around like the business community hey do you guys like my any chat you guys have any issues with it they're like yeah and i'm like great or they're asking me because of my knowledge because we've been working at multiple different startups and my background right they're asking me is it possible to build this this is this and i'm like uh do you have where's your data at and they said like oh it's that many chats i like all these different places i was like if you don't have your data internal to your own operations and how your operations runs internally how can you basically supercharge that with ai if all your data is scattered right so then i was like okay

32:22there's a need for this so then that's why i built big because i built all these tools everybody wants these things right if you wanted to get at that next tier ai level you have to have these things in-house to basically supercharge like an ai agent to basically understand how your business runs understand your clients needs and wants at the same time to then have it execute at a high level otherwise you don't have you don't have the context right if you don't have the context how can you it's human person right like if the human person doesn't have the context of all these different things how do you expect the person to execute or deliver like a substantial result that actually moves the needle right otherwise you just kind of get this swap that is very generic and or

33:19might not even actually help the customer or your client at the end of the day right so that's the the reason why we decide that's the reason why i decided to build big because even for myself that's what i need right in order for my if i wanted to bring like voice agents or ai setters or any other stuff for my own business for my own stuff i'm gonna need that so i was like okay let me just build it because i don't see anyone else building it right so everybody's focused on like what's that called all these companies are focused on the internal ai like yeah and that's definitely going away at least maybe not going away i feel like it's implemented in a lot of the big boys but people are quickly understanding that they just want their ai to hit your system like i don't i don't want to use yours

34:15it doesn't make sense right because a lot of a lot of these bigger companies are basically just ripping like kimmy models or quen models and basically like um what's that called open weight training it in a way or what's that called uh i word for it but they're like training it in a way for their own specific needs for their business but then it's like why even take all that time i guess the one thing that these companies or these businesses might be scared of right is like oh if we open it up like a mcp connection to like open ai or or tropic right and allow their ai ages to connect they're basically gonna steal trade secrets yeah right and they're

35:10saying that you know like they're already they're already gonna steal your trade secrets regardless yeah security security is a security theater at the end of the day yeah like i mean like are you really gonna fight against it i mean i mean i'm fighting with a provider right now for another client they they don't they don't even have a public api so i was just like i just want to know what things are accessible because it says it is stated that you guys have an api that people of a certain plan tier can access yeah of those providers or like companies where like that's the reason why i created big because a lot of these places like they're closed gardens and or they just don't even have like sufficient apis yep right and i don't know if that's like a legal thing or what it is but even

36:07when i'm like asking i guess claude like is it legal to do this is it like to do all these things and basically it is as long as like your model isn't training on customer data which is basically plot and open ai state that they aren't doing whenever you're prompting right so i'm just not going to question that because you know like that's what they said what they said right what they said so i'm just going to use a frontier model to basically connect with you know business data or whatever so yeah why why make it so difficult unless you know you you want like something super custom tailored but even that at that point tailored model training data gonna

37:03be better than a frontier lab right can i mean it's just gonna behind but it's still gonna be we're already reaching that point where it's almost good enough i situations it's definitely more than enough if you have like super domain expertise like in like the sciences and or specific laws or something like that like law firms or whatever that like you have like intricate laws or something like that then yeah yeah probably you could train your model to be more specialized in that sense but then like how many how many places need that or how many businesses need that that many yeah in my opinion right so but also like all these other businesses that are saying that they have that domain expertise so maybe you don't to be honest to be frank

37:58maybe you think you have market share but it ain't special yeah there's a lot of that out there i mean what we're doing with big is not anything special in my opinion there's plenty of products out there that do serve what we serve but i just think that i want to be working in somebody else's ui when i could just keep working through cloud or terminal or cloud code or something like that where i'm just like do this set this all up

38:32 · agents over UIs & forgetting onboarding

38:32please it's it's kind of like this um the trifecta you need to be able to read your information you want to be able to interface with your data your your personnel people operations right and then you also want to be able to execute commands that allow the whole ship to run um and we're trying to take that approach very much so with with big where we want people to access the things they need to access use their agent or use the platform but really be able to act on the things in the way that makes sense for them and us to be to be frank like the ui is for a human like ui is for a human twice for human at this point in time more humans on the interwebs that are

39:29interacting with all these platforms tools websites so forth right at some point there's going to be more agents that are going to be interacting with just tools in general in my opinion right which is why even for twitter right they decided to completely do their api in a usage-based business usage-based api yep the reason why is because if ai using it right they're just charging per request right and it's like faction of the zen or something like that right so that's something that big might do in the future if we switch between like who's using this platform how many people are connecting through the mcps and apis if more people

40:26are connecting to the mcps and clis than active users on the ui then that clearly tells me that you know people are using their ai agents to handle the tooling plus also to me it's like oh why do i want to go click around at a website and then go figure out how to utilize your platform through the ui which is like the very first thing that they encountered as soon as we got people off the wait list yeah i told them to yeah i told them to get into it right like we're engineers we're developing the thing and we totally forgot about like the onboarding experience yeah we totally forgot about onboarding and user experience of like when people like but also like when we think about it like oh yeah it's just intuitive like you just you just click around and go figure it out

41:22right like for us we can that's what we would i don't care what because because because because i think that that's also like the gamer mindset like it's true like like the gamers instructions forget their instructions let me just play the game yeah reading what is that skip skip let me just get into the game right like that's that's the mindset we have right we come from that like you know asian era of like intro internet right so then these people that come in maybe they're not gamers they're not like tech savvy or what have you right and they're coming in and then the first thing they asked was how do i use this and i was like great and then i just screenshotted those messages and i sent it to you matt and i was like we have a problem and then you're like frick i was like yeah i just remember that i just started laughing so

42:19hard and i was like right for us that makes total sense but for other people out the window and then you know i was like i was like wait i was like man how do we do onboarding let's make a video like yeah we probably should make videos and i was like oh shoot okay wait and i was like wait a minute i remember i found a tool for this driver.js actually it's pretty sick it's pretty good it's pretty good like i remember finding it like i forgot when like i was just like you know on the you find all these types of tools and things on tech twitter yeah to be honest i found off tech twitter at some point in time and then i was like wait a minute i remember there was a a specific mascot yeah and i think they have a fox mascot for their driver dude and then i was like i remember this i forgot the actual name and i was like okay let me just let me look for like onboarding like libraries in js right yeah i was like oh i found it and then i sent it to you and i

43:15was like yeah i think this should work and then you did the first yeah you got that you got the library and then i was like all right help me just get it and i'll just like we'll tweak it go yeah i'll tweak it and go through like certain pages and stuff to make it more like intricate based on like what i can see that doesn't make sense or like just skipping like certain things and the context is like not there for people to understand so i was like okay make sure that everything is covered and then we covered it within like maybe like a week we covered like every single page every like little setting everything and then i was like all right guys come on back to a platform and provide us more feedback crickets crickets absolute crickets and then i could i could i could already tell like we them at that point because we didn't have onboarding yeah i know it's too complicated and or they don't want to learn how to use a new platform right so

44:11which is which is to be fair everybody yeah even even the people that are not using our platform and are using the traditional like mini chat type form whatever if they're coming in there like first time first use i do this how do i do that yeah how do i do this and you know those those platforms are probably creating their like little ai agents in there but sometimes you know those ai agents aren't even you ask it a question sometimes and then it just gives you like some random generic answer that's not even maybe cater to the platform and just like how is that possible when you guys literally built the platform and your ai agents internal and your internal ai agent still cannot tell you how your platform works that doesn't make any sense to me why it's not clear that's real or or

45:09or which is another thing where like if that's the case without we wanted to build that type of like ai agent into build into big where people don't want to like connect versus cloud or whatever blah blah blah and we actually wanted to provide that as like a little feature set yeah to me i would be like what was it called um like the prompt response reviewing right reviewing the output oh yeah of the model yeah yeah yeah yeah i feel like what's called i don't know what the right words i always forget words but

45:44 · ai in college & unprofitable labs

45:45right you're reviewing the output of the models right you got to make sure the output is 99.99 correct and accurate to your actual platform yep and then if it's not you should be having this iterative process to make sure that it is yeah and then these you know these freaking companies with like 10s and 20s and 30 people engineering teams i don't get how i don't get how man we're in this age where it's just like if i can duplicate myself that is a dumb amount of ability which is which is the yeah which is the power now of ai yeah to be honest like you can duplicate yourself or try to duplicate your way of thinking as much as you can and pass it on to these ai agents right like that's another reason why we decided like it doesn't make sense if you

46:40have a knowledge base and skill set to execute to utilize ai to basically build whatever you want to build yeah so it's kind of like i'm seeing it across the board of like people that know how to execute they're not may not even be engineers but they're like super savvy where they could go figure things out and they know how to find the right context and like find the right information information to then give it to their ai to make it as close as possible to like be a good engineer for the ai and it's just like you just have to find the right context and if you know how to do that for any field then that you're basically having an employee for that field right off the gate so it's kind of like you're basically i don't know i see on like a tech twitter like these people that are like from harvard university or whatever

47:36that aren't actual engineers that built like these crazy libraries that are like two 2k 4k stars on github and it's all ai and i'm like dude nice i'm jealous i wish i wish like back then like 10 years ago in college or whatever that we could have done this right like yeah coding or learning the foundation if i had ai and i was back in college i went to college for game development i probably wouldn't have never slept yeah if i i was in college and i was as degenerate as i was in college with what i have yeah yeah oh boy it would have been over yeah i don't think i would have ever slept because i was already not sleeping back then and i was

48:33writing code by hand i was already having a somnia back then i'm like dude a somnia now with ai baby oh off to the races yeah you're just cranking it's a good it's a fun time to be alive um well okay um interesting yeah very interesting time to be i don't know about fun i think it's definitely interesting yeah interesting because i was like when i thought fun i'm just like wow everything is on fire yeah everything is literally on fire my god i don't know about fun yeah that was uh i had to i had to really i had to wipe that away real quick you know what's you know what's crazy like these frontier labs are just burning like on paper like they're not profitable no i think um it's roughly 20 i think it's uh i

49:29think i saw something recently it was like 50 billion negative 10 billion like in revenue no i i thought it was like in the hundreds i was in the hundreds i was just saying one of the companies oh okay i don't know which one but yeah i think the couple yeah we're definitely in the hundreds of billions and it's all going all going down the drain well i think that's another thing where i'm like you know google hasn't been that vocal in a lot of this stuff like huh they're not very vocal in terms of like their marketing of whatever they might have like gemini ads right every so often but in terms of like tech twitter or just like influencer like push open ai and quad definitely have like they probably work with these influencers to push out like as soon as a model hits all the influencers just talk about it just to give awareness

50:26to like people to get people using their models but in reality like these labs aren't profitable like their business models aren't business models which is like another thing where a lot of these start of tech founders right like do you have a business model is i don't know we just kind of you know drive up drive up the user base and get enough users get a shit ton of users and then just tweak the the pricing right and then they're like okay yeah we have money now that's majority of tech to be yeah majority of the tech that's how tech was done but i don't think i don't think that's viable anymore in this day and age you gotta you can't just have an idea and have a lot of users you gotta have users and actually have a profitable business otherwise there's the cost is just gonna be the cost yeah yeah the 2010s were definitely about burning money and being

51:24acquired and then exiting yeah now it's or you just now it's acquisition distribution and revenue i revenue part either because freaking well i don't want to go with why dude i don't want to talk about why i see an a116 disease with their class of marketing teams yeah i've been engineering teams they're marketing teams yep i don't want

51:58 · bootstrapped marketing & the internet's missing payment code

51:59to get in that though so where is it all going where because we're obviously we're going to try to record this hopefully week weekly well we'll be able to talk on uh we record next week we'll be able to talk about la tech week oh that's true yeah we're going to la tech week we are because we're not marketers right nah i'm trying to learn as much marketing tips from from youtubers right yeah but i happen to be a co-founder of a marketing related product yeah but but still like like like your type that type of marketing product and marketing requires up front capital yeah to even access that type of marketing yeah right but if you're bootstrapping which is what we are doing like me and matt and maybe other people out there you see you see

52:56right it's either just either user-generated content by yourself as a founder or by your team as a founding team right and or do you have some way to like reward the customers and create like a growth loop where right like if you bring on customers to your platform is there a way where the platform in itself is marketable or you create some sort of marketing by just using it there was a interesting thing that i never i don't think i told you but there's like an interesting thing that i kind of wanted to like think about in terms of like big or like people that use big and we're tracking attribution we're tracking their growth right for their business what if we create like fun stat pages and leaderboards for like you know how like

53:54they have like spotify wrap-up yeah what if you have like a big wrap-up like either a quarterly or like half year wrap-up or maybe yearly wrap-up or like how much did your your business and or whatever actually grow yeah right and then it you kind of could maybe like tell like which posts or whatever kind of popped off for you and or whatever yeah like how much activity it's like oh this this form went crazy or my crazy or like this piece of content that you create like your dm funnel for pop the off right or like maybe if you're a custom build person like okay you have your custom build how much did that custom build once you built it actually drive revenue for your business like how much did your business grow after you were able to operationalize or

54:50automate a bunch of things and work for us yeah well today right like that's that's the results that we want big to provide to businesses because it's one thing to just be like oh you have a bunch of tools and everything like that yeah but if you can't drive revenue with having these things what's the point like why are you subscribing to us right it doesn't mean anything if you could like have no value in terms of like oh i'm a subscriber now to your platform but then it's like i don't know how to use your platform to this full potential to actually drive revenue and that's where it's like as business owner for big i don't want that to happen for my customers like i want to know like which one of my customers are having issues and how do we even help them grow as a business right like it's basically like business as a service right that's basically what big is positioned to do

55:46right like operations as a as a service business as a service how do we then operationalize every small business and or mid-size or large enterprise to figure out what in terms of their ugc whatever content attribution how do we actually to then say like this is what you did in order to drive revenue or whatever right so that's the problem that has to be solved which is an interesting problem i don't think we're there yet for big and or the industry in general but you know gonna figure it out in time with power of ai power of ai with opus by my side

56:41fable fable six i'm waiting for you yeah i'm actually kind of kind of scared but what in the world is fable six gonna be fable i don't even 5.1 is a weird number but i think that they put the point one because of the whole what was it legality of certain things right yeah there was oh it was any change it was the eu right like the eu required anthropic to do something in terms of legalities oh they changed the model they had to release a version for it i think no i think they had to like um watermark the the output of anthropics things i think like like you i think they're saying like we need to know what's actually being generated by ai in terms of like text all text and all images yeah yeah yeah yeah so so like the eu basically did that and was like going to anthropic

57:39and maybe open ai and said like hey you gotta you gotta watermark your stuff we're requiring it so i think that's why they put a point one update which was basically giving that watermark i think too i know but like i i almost started watermarking images for pando but then when i went through i was like what is the point of this there's more than enough waste to un-watermark things by generating like a new asset or something so like what is this yeah what is the series of hoops like we're wasting here know these these these these boomers yeah to be frank these boomers that are deciding to create these things don't have any idea going yeah they don't have any they don't even know how the internet works no they don't which is terrifying the normal people yeah normal people don't know how the internet works they don't

58:36know how what actually allows them to use the wi-fi or satellite or whatever they don't they just go to the browser right and they just put on a website and they're like all right works how does it work nobody cares i didn't even know until i actually joined the crypto domain company yeah i was in a crypto domain oh crypto domain that's what you said yeah i was in a crypto blockchain domain company that actually was like oh i need you to build a plc to make crypto domains basically be actual domains that people can use on like google chrome all the different browsers right so then it's like how do you do that so go learn this is the only time in my actual i don't know engineering

59:34career where i had to actually use hex digits and bytes cool and under and understand text digits and bytes because i had to convert hex digits and bytes into domain names in order to understand how to create the plc protocols that are attributed to that in order to send the data for a domain name between servers yeah we want to talk about weeds i got to the weeds that's kind of cool yeah yeah i got it to work it was interesting and uh you have to you have to work with um icam which is like the international like domain authority of the world there's like a non-profit that basically handles like root servers there's a whole history

1:00:31behind it yeah you gotta learn the whole history of like yeah intro very interesting yeah so i gotta learn all that and that was i don't want to talk about it i don't talk about that i want to name the company yeah but but i did the plc and then maybe like a couple months later i got laid off no like i should take should take that plc back that's all they wanted from you yeah and then they changed directions to like try to get that stuff done and i think they're just it's like a lot of you know again like boomer processes of like trying to get things to work a certain way and like i don't know you're basically telling the internet protocol of like how the actual internet works to adopt this new way of doing things and you had to get buy-in from basically the the ones that own the

1:01:27domains not the ones that owns the man the ones that handles how the protocol is built and oh how it's actually interpreted yeah how it's actually interpreted how it's used right because even even when the internet was first made there was no payment system gateway http right so you know how like http status codes or whatever there's like 400 whatever there was supposed to be a there was supposed to be like a payment specific http status codes go into http but it was never done right and that's the reason why there's like i don't know all the different payment services but there's no like there's no like just payment on rails just based off of like http right so then there's another whole thing where you can you can just have every everything into the protocol but

1:02:22that's a different can of worms that you could open up yeah one very interesting yeah very interesting very interesting yeah very interesting stuff i got into in that company like i gotta learn a lot about like red tape or just the internet you learned about the internet yeah very interesting maybe one we'll dive into the history of darren just like i brain is braining your brain is

1:02:52 · what's next: pando arcade & indie games

1:02:55braining i mean it's super fair uh but what comes next we have la tech week we have obviously q4 that's kicking off we have the tail end of this mega heat wave we have a new bubonic plague of sorts that's kicking off in russia kovid did you not hear about kovid part two it's not it's not i've seen that yeah apparently there was like a lab break uh some lady whatever broke a thing and then contracted a thing she went to a hospital or whatever was on a ah that's that's just the cia just just breaking things in literally what it is probably yeah not just some random lady trust me there's no such there's there's no such thing my tinfoil hat is tinfoiling all right there's no such thing as oh what a coincidence there's no such thing anymore

1:03:52guys yeah there almost like never has been it's not a question yeah it's definitely not there's way too many dumb coincidences uh around it that i'll i'll tell you about offline because this is not a conspiracy podcast we're just two as two ex engineers trying to do something big and bigger i was like where'd my thought go and i was like ah bigger up but also pando pando's releasing hopefully hopefully the shop releases this month i know i'm trying to yeah i'm trying to push i mean i'm even i'm even thinking now i was gonna ask you i'll i'll ask you right now on the pod the game that i'm building is like a fully like kind of offline rpg like in the offline yeah it's just like just a single player rpg i am building in support eventually for like multiplayer but like it plays is it like i can't like location huh location basically no like it's like

1:04:49literally fantasy story standard classes like rogue and warrior and you're going on an adventure you have quests like it's really really standard adventure game um similar to secret of mana if you've ever played like those kind of rpgs super nintendo i know it i haven't played it yeah i might have played like a little bit of it because i was a kid and i was just interested but i don't remember if it was it was that i mean they all look kind of similar but i don't remember yeah you're just going around it's like action combat and you know you're leveling and there's a story um yeah but that's that's one thing so it's like if that ever worked out i would totally house that on pando people would just like play a game like pando arcade literally becomes an arcade you know the whole because it's kind of interesting because like you know how the internet back then like a lot of like the web browser games yeah or the first type

1:05:47of internet games right if you think about it yeah so we're kind of like bringing that experience back that's what i'm saying because because i don't think there's that many web browser games anymore nope not really like you have really a place for it like you had a runescape back then you have like i think it was the whole like you know when the kids were getting online they're going to like this channel nickelodeon yes called penguin like penguin right and then you just play like all those different types of games right on all those different websites and then you transition to runescape and then maple story which maple story classics coming back in three weeks dude i'm probably gonna play i'm probably gonna play maple story you of hours i plan to maple story in my teen years bad we're

1:06:43bringing it back so bad make a guild so bad yeah make a guild and then we'll quests yeah oh yeah jump quest jump quest rage bait biggest rage bait should have ever seen as a kid i gotta put a jump quest game in pando for sure yeah that'll be that'll be fun oh yeah you were saying we're bringing back like the the old school arcade experience yeah just because like i think people i like people are trying to get experiences again plus also like the whole congregating industry yeah also like the whole gaming industry kind of in a weird

1:07:39position nowadays i developers always had a bad but it's just terrible for the industry now in my opinion like it's even worse than before like i don't know anyone that yeah like it was already bad before but the triple a space is pretty bad um but the indie scene is like the right oh yeah yeah the indie scene is thriving right like if you want to be a game dev or game development this you just gotta be indie yeah you gotta build your own indie studio build your games yeah build your one hit and then it might just make you a millionaire like mega bonk yeah yeah and you don't have these dumb investors or dumb publishers or whatever right then like i was saying and the other startup that i was a part of what i i

1:08:34think the next big ip is gonna be ai generated for sure for like yeah i don't think like it's basically everybody now has access to become a creative person right like not everybody knows how to like design or create like draw or whatever yeah the the technical uh threshold has been dramatically lowered anyone with real ideas real vision oh my god i just got triggered there's yeah yeah whenever i think of the word vision i remember hearing a line in a place where you and i used to work i don't remember i wasn't part of the i was i was on the back end yeah data team you know back end data team internal tools but

1:09:30we didn't i went okay when i had the yeah when i had to help build yeah endpoints for those things when i saw those things i was just like what are we building this for what what the hell what what this this is it no one's gonna use this i don't know what the i was like we're spending so much time on this yeah this is not gonna move the i'm like bro like heads yep yeah yeah it's cooked uh yep it was cooked literally yeah yeah literally yeah okay we we have we have since uh we have since uh traveled our separate ways thank the lord yeah thank the lord but well darn it has been fantastic catching up with you

1:10:25i mean we catch up a lot just not not on the pod yeah here's here's to the next episode not being another four months hopefully next week uh we'll just make this a more regular thing because hey not not just like this podcast but like even both of our content we got to get back on the wheel oh yeah the time is now yeah opus opus 5 5 maybe with these hyper frames the vinci oh five blinds now i got we got to get that set up yeah that's uh i found a completely open sourced set of all of the adobe products that someone made i'll send that over to you after this which means they're probably all mcp like ready let's be real they probably are hopefully yeah hopefully yeah if not if they are if they aren't that just means we just make it that way as long yeah as long as the gates allow

1:11:23me to un-gate the world is my ocean yeah the world is your ocean um yeah do you have any closing remarks anything you would like to say just if you're a small if you're a small creator the waitlist is still open um we i check out weekly but most of the time nobody joins because we haven't been advertising big yeah but now now's the time we will start very yeah we have a free yeah we have a free tier it's always going to be free and it's going to be a good free tier unlike these free trials or these other places i think beacons is the only one that actually has a decently good free tier so good for them i mean yeah you could you could use beacons or use us but you know if you want to support big to

1:12:18support small small developers trying to build something big but thank you guys for listening or watching hope your dinner has been great if you if you happen to be catching us during one of your meals like i like to do with my podcasts and we will catch you on peace peace peace