0:03okay we're on hello jonno and hello again Eric hello hello I was going to say special guest Eric but you've been here once before so I'm not sure you count as special well maybe twice regular guest three times as
0:20regular you're still special to me though well that's that's sweet of you to say it's okay I can handle it um yeah what if you have you had any um I mean I know uh well let me first ask I the last uh lesson um jonno and I were looking at using solvet with external apis which was
0:48actually I thought really fun and interesting because it feels like there's a couple of things coming together there which is with these ridiculously powerful external apis now you can do so much with so little but they're all a bit different and you can't possibly remember all the details of all of them but with Sol totally suddenly you just say like oh I paste in the URL to the docs for them and please run that we
1:16were like let's create a picture let's create a poem let's whatever and yeah I don't know that felt pretty pretty fun to me what what do you think about this coming together of capabilities there well you know we've modular architecture of you know code reuse is like the dream dream of computer science going back to its Origins and we've only ever sort of gotten it you know for I feel like there's been like brief stable periods where like a platform stays stable
1:45enough that people actually start to do proper code reuse and I thought that the internet itself would dramatically increase the amount of code reuse and of course it has to some degree you think about like you know npm or pip or these code repositor that's obviously a huge huge step in the right Direction but the like the kind of Achilles heel of that whole system has always been that once the second you agreed to depend on some external resource you are now on the hook for the entire life of your program
2:13to keep up with changes to the interface to that thing and like there's a shocking number of situations where honestly for being real we don't say this out loud very often because you're not supposed to say this but like honestly the overhead of having to do that dwarfs the benefit of the code use compared to if you just written the darn thing yourself think about like how many libraries you could have just written yourself if you wanted to now it can't possibly be true all the time surely there must be sometimes when it pays off but anyway it's a very big deal to lower
2:43the cost of the low lower the overhead of being able to use one of these things uh and keep up with its changes and kind of automatically adjust you know revision by revision and especially with a live service even a static Library it's bad enough but for a live service you're being Auto upgraded against your will to the new version whether you want to be or not so you have these breaking changes and I think that's going to be a very big deal both because it's just going to make life a lot easier but also because um there'll be a lot more
3:12situations where you might actually really go through with code reuse and actually do it in a in a way that you wouldn't otherwise and with these new Services since they're so intelligent there's just so many small situations I mean ghostex we we were talking about last time was such a perfect example of like almost too small be worth doing by by by it' be very easy to talk yourself out of it it's like all right it's like a very small quality of life Improvement but you know the operational overhead of doing that is too high but
3:41if we if we lower that operational overhead we can start to apply this to lots and lots and lots of situations we might not otherwise have been willing to do and of course we will discover as we talked about last time because of the fundamental uncertainty of all this sometimes what you think is just some tiny little quality of life feature actually turns out to be like a whole company in itself you know I don't I don't know if ghost Tex is be the next Facebook but like one of these days one of these little things will turn into a massive opportunity that you wouldn't have otherwise even been willing to explore yeah so I'm pretty excited about
4:09it yeah yeah I mean there so there's a there there's a couple of interesting examples which we're GNA drill into in detail in the coming weeks and one of them uh is actually the best example of something that just happened yesterday which is as you might have seen um an answer AI post from haml was on number one or number two on Hacker News and it was about
4:39Devon um and uh that um article was written in solvit or an internal version of solvet um and so the kind of the story of that's quite interesting because the first version of it
5:01wasn't that inspiring and we were literally thinking uh maybe there's nothing here and then we're like well let's go back and use our new process which we'll be talking about but it's like you know we we record an interview with each other we talk about what's why we're excited about this or what's interesting about this or why we're disappointed about this and we chat and then we kind of get AI in a very non- agenty you know stepbystep incremental approach to help us right about it
5:31um and to yeah see that post where were literally thinking ah this is so dolls to probably not even be worth posting to us both being like wow this is a good post I think people will like it and then up on the top of hacken news is yeah really interesting to see how this approach outside of just coding is uh you know giving us big opportunities um do do you do you see that as well like kind of writing and
6:01stuff like that Eric yeah yeah and what I love about that story is people have heard that llms are good at certain things and not good at other things so like they're good at summarizing condensing you know translating reformulating excellent uh you know as soon as you want them to be your agent then the problems of hallucinations you know in yeah how many times you've tried to build one of these ancient things where you're like okay go go do the thing and it's like great here's my plan to do the thing you're like no no please do the thing it's like
6:30okay my step one of the plan delegate to another agent to do the thing you no no you're actually the agent being delegated to so like the human ability to dodge accountability and to never never take responsibility for anything is like baked right into these llms so so therefore like people are kind of like stuck they're like well I want to I've been told that AI is supposed to do the human things for me but it's not actually good at that and I think this approach you know the dialogue engineering approach even in any kind of creative anytime you're cre a
6:59deliverable whether it's Pros or code or you know I'm sure this would be true in you know in multimedia also um by taking human ideas and creativity in An Inconvenient form and then asking the llm to summarize it to help you condense it to help you formulate like then then you're really playing to its strengths and your strengths and the final deliverable is much more likely to be you know that the yeah yeah yeah the AI does not have agency what it is it's
7:28much more like a cybernetic uh augmentation you know Steve Jobs just call us the bicycle of the mind that was what he was talking about all those years ago you know did we really ever truly live up to that Vision in all that we've accomplished with computers tilled down like sometimes yes but a lot of times no we've introduced more cognitive overhead uh at the same time as simplifying but here I think we have a chance you can really take an idea in your head and render it into a deliverable uh in a reliable way and that's I think that's a very big deal it's actually interesting how far back
7:57that cybernetic idea because um I I um was actually looking up the atmology of the word cyber because my mom and I were complaining the other day that cyber was the word or word of the day and I was like cyber is not a word it's a prefix and I looked it up and it's like oh no now as from like 1990 it became a word and not a prefix I was like and my mom was like not even sure cyber than anything so I went back and it turns out cybernetics in the 1950s was the
8:26original source of the so called word says the New York Times cyber um another example which we're going to talk about uh quite a lot in future lessons is the we've talked about a bit it before is the CIS admin stuff and you know actually I don't think i' might not have even mentioned this to you yet Eric but as you know probably the folks listening don't we're planning
8:54on creating a fast HTML hosting service um to make that super easy and one of the fun ideas we had was that every time you pushed a project to your fast HTML hosting service we will install a copy of solver in it for you and so you'll get like not just bash buddy but um sorry shell s but you'll have
9:20solver there turning you into a SQL light expert and a web app debugging expert and a you know system administration expert um which yeah I thought it's kind of interesting so I don't know sorry to throw it at to you without any prep but what do you think about this idea of like you know well I'm I'm very excited for admin I'll tell you why I'm excited
9:48about it um and it's gonna sound kind of silly but you know we talk I think we might have touched on this last week that like when you first pitched this idea of using this for CIS admin I just thought you were totally crazy and I was I can't possibly right I had like I've all this I've built up all this intuition and skill and knowledge about the cloud and deployments and whatever over the whole sass era so you're like I'm going to just going to throw all that out and go back to the old uh bare metal hosting thing and just my my n reaction that can't possibly be right but then I was like well but nonetheless explain to me what you're
10:17thinking about and of course this is like just the ultimate you know MVP thinking it's like if you actually work from first principle sometimes you discover that you can just do 80% or 100% less work than you were previously thinking and it unlocks all this amazing things so it's one of the you know uh great examples of thinking from first principles but what really reminded me of is there was this time I when I first became a computer programmer and started doing client server work you know in the uh you know 90s early 2000s for every time we would try to solve a problem
10:45with a database or um you know a proper proper heavyweight system there would always be some engineer on the team who would be like this is so wasteful I could solve this problem with the flat file you know why are I want to need a database for this and like I feel like it was like a huge part of my job was just like whenever that would come up be like I know I'm sure you could but then like you we have some random file and some proprietary format that nobody can can read it wouldn't be properly replicated and it wouldn't be secure and what about this and have you even thought about how to escape the
11:14characters and you I just like I got so good at be like no so I know that is true it would be dramatically more efficient if we could do that but know somebody like we could just use a mem file we don't need to We R it to R the like I know I know that's true but like what happens when that computer crashes it's like you just like there's this massive backlog of stuff that you're not not thinking about but the thing that always bothered me all these years of doing that was like they are kind of right about one thing which is that in
11:42order to solve this massive number of admin problems we are accepting an awful lot of overhead both computer overhead like we're now taking a you know you know one millisecond operation and turning it into like a 5-second operation that's like that's pretty painful we're also taking something that could have been done on one server and forcing it to run across a hundred servers and we're having all the attended Staffing costs related to keeping all those servers running all this other stuff and of course with Cloud we Outsource that to somebody else but that's that's just an economic
12:12Arrangement where we're just we're now paying more for the computation that we could have basically gotten for free Port advertising the cost of all that stuff and the thing that like really blew my mind about this example now look this is just one little example I was like oh is it possible that like with solv it you could actually answer all those objections like instead of using them as a rhetorical weapon to prevent you from doing the thing which is how I used to use it like could we actually solve all those problems and like I don't have any
12:39ideological attachment to the cloud or to doing it to data I don't need us to use this database or that this tool that tool I was like oh this could be incredible and like now several times since we had that conversation someone has pitched me on a startup or something that revolves infrastructure and I have this voice in my back of my mind it's like if they took the class could we just do it for free like could we take we have to go raise $100 million I'm like could be could be but
13:08could you also do it more or less for free and like I don't know like I think it's really a really interesting open question about just like again forcing us to really question our assumptions about what things cost and therefore what things are worth pursuing I think it's a very exciting like tantalizing possibility I shouldn't say that we've like solved devops forever you know have a lot of work to do but these early examples I think are very promising of a a category of problem that we used to just accept as very heavyweight very expensive so we promised last time we
13:37chatted to you Eric that today we would kind of Zoom do two things one is zoom in a bit more to answer Ai and the answer the other is to answer folks questions from last time so we're going to do it in that order um particularly because I got a bit of a segue here which is and think about all these things that we can now do none of which I particularly remember us ever saying a year ago these are the
14:07things we're planning to build and do I think it would be fair to say like as a company we're all about like kind of having long-term Vision goals values some some underlying thesis about how we'll be different but we we've never had a planning process and in fact we've never really even had written plans
14:35um I guess I quite a few folks have been asking about this it's like what are your thoughts about kind of setting values or having a North star both at an organizational level at a department level but perhaps also particularly interestingly at a at a personal level is this something you do that you recommend people do that you have suggestions for
15:03you know how to help because I feel like I feel like we do it pretty well but I'm not sure why so I would have trouble helping teach people how to do that better oh sure yeah I think I think it's very clear why so that's that's easy great um um yeah so so this is definitely the most common misunderstanding about Lean Startup Bar None since I first WR the book and it's one that has really driven
15:31me nuts because and I didn't I didn't I used to feel really bad about it until I realized that they used to have this problem with lean manufacturing like going back to the 1950s like people have been having the same confusion and problem in their understanding which is there's something about going faster or I don't I don't want to speculate what the psychology of it is but anyway um people think and I was like this too when I first discovered these tools I was like great I've I can take all the creativity and the vision stuff out of Entrepreneurship and turn it into a crank that you turn and you just get an
15:59automatic result and I I so many people have have got fallen into that trap they hear about MVPs they hear about split testing they hear about some aspect of this like great perfect now I can take the human judgment part out of it and make it more mechanistic and I I wish that was true that would be really great but as it turns out I have a lot of experience with this if you if you just AB test everything and just do whatever makes the metrics go up to the right you will be selling pornography or a psychic hotline like in no time because of course the things that have the best most immediate gratification are the
16:27things that are the most fraudulent and so just you know if that's what you want to do I want say but yeah most people that's not what they want to do I've seen a lot of people fall into that trap I found that trap earlier in my career I know just what that feels like well they're just the most hypy or the most grifty or like I don't think it's grift but certainly hypy like everybody's jumping into agents now because it kind of seems like yeah it's just kind of like whatever you have to do it's like you're no longer really thinking you're just you're just reacting to what you know what's in the moment or what yeah like any people who
16:57have customers will sometimes just be like whatever makes the numbers go up that's what we do and you know Corey Doo has this very colorful word he calls it inchid ification is what's been happening to all of our consumer products and platforms but like this is a big part of the reason why it's like a large just like a large segment of these companies are just dedicated to just like whatever it takes growth at any cost like just and the problem with that it took me a long time to understand this is you can always always always make your okrs go up by sacrificing the
17:26company's trustworthiness okay because that is a really what's that okay Oz oh sorry objectives and key results is like a a fancy management trend for for not just having analytics be used for products but having that like the analytics goals of the company be subdivided subdivided until every individual person uh has a metrics Target that tells you whether they um whe they're performing their job or not like Peter teal was very famous about assigning every person in the company
17:55only one one and only one metric that's your job to just make this number go up up and there's some real advantages to being that analytically focused but one of the downsides is what I'm saying that you can um there are these intangible resources you can trade against to make your numbers go up and like I was talking to a company just the other day that the CEO had kind of stepped away and he was coming back the founder coming back to like clean up the mess and one of the things he found was the company was sending on average I can't remember the number now I think he said we sending 12 emails a day promotional
18:23emails a day to every customer so they're just like burning all this Goodwill customers are like getting so tired of hearing from them and he's like how did this happen we all know exactly how it happened every individual email you add like makes the numbers go up a little bit in the short term and no one's paying attention to or measuring the long-term cost so I tried to address this really carefully in the book to Lean Startup in fact part one of the book is called Vision it's literally the name of part one of the book the whole section is about vision and in fact uh a
18:52common criticism of Le startup is That Lean Startup is anti- Vision or doesn't doesn't have um you know there's no definite plan and all this stuff and it's just like I remember once being in an interview where someone was ACC costing me with this and I like had a copy light like I pulled out my copy lean start I was like let me just check what page this is first addressed in this manuscript just in case because I was like maybe I only put it in page 200 and so no one read that far but it's like on page five in the introduction like very clearly says it right up front so you so you know I've tried tried really hard to make it clear the scientific method does not work if you
19:22do not have a hypothesis so if you go into your chemistry lab and start throwing around chemicals just like see what happens you are not doing chemistry something cool may happen and you may you may yet make a discovery who am I to say but that's not the scientific method and vision is kind of like that critical ingredient that the more specific you can be about what you think is supposed to happen the easier all this other stuff gets so you know people's like oh if some metrics is good more metrics is better if some MVPs is good more is
19:50better if an MVP should be small then smaller is better but no none of that is true what matters is are we learning the most critical thing that we have to know in order to figure out if we're on the right track so from a values Northstar perspective the stronger the founders vision is the easier this gets because then we don't have to like figure out a bunch of stuff we actually Steve blank one of my mentors had this great phrase he said the job of he called it customer development the job is to at first the
20:19first job is to find a customer for the product as currently specified so it's not like figure out what the product should be it presupposes this is in the old days of startups this is how it always was you wouldn't start a startup unless somebody had this crazy crazy idea that they wanted to make a thing so that's the first thing is to really understand that that that having some very clear idea about what you're supposed to happen is really essential where we get into trouble and the reason we need Lean Startup is people who are so are good at convincing people that their idea is
20:48good create what people used to call the reality Distortion field uh phrase it was coined for Steve Jobs that he when he when you were talking to him he could make you believe anything up is down left is right green is blue like whatever he just had that persuasive ability which is a wonderful very helpful ability it's one of the founder superpowers on this planet to get people to believe that some impossible thing that does not exist is going to exist like every startup financing is actually a miracle if you really think about it because you're asking people to pay you money for a slice of ownership of a
21:18thing which does not exist like wow very cool so that's all good when you're trying to get resources when you're trying to get people excited when you're trying to keep people busy Founders are so good at keeping people busy the problem comes into what if they're wrong so like at what point do you step outside the reality Distortion field I used to tell people can we just double check just real quick double check that we're on the right track so that's what we're trying to do is find that balance of of vision and and testing and of course that's that's actually where the concept of a pivot originally comes from a pivot is a change in strategy without
21:48a change in Vision notice that if you don't have a vision you can't pivot by definition and if you change the vision all the time you're not pivoting I don't know what you're doing just running around and and in you know that that just coming back to that Steve blank idea like people in the class might recognize a taste of that that when they signed up the first thing they saw was a screen that said in the largest letters do not
22:15join this course if and then a big list of reasons not to join the course which because I guess that's basically what we were trying to do is bring in people who want what we're selling you know yes it's so counterintuitive um and and you know for those who know the theory of early adopters if you don't know if you only heard the phrase early adopter but you don't know where that phrase comes from you can read a books like crossing the chasm or actually was called they were called lead users or lead customers
22:45before before they were called early adopters and anyway the idea is simply that um you have to be able to find people who have such an urgent need for what you're offering that they're willing to buy it from you you like I it's like you could you could buy it from Microsoft so why would you buy it for me you know no name person right then there has to be a reason now all of you are here because at some level there was something about the solv it experience that was you were willing to to take a leap of faith on it even though you never heard about it it's not
23:14like a friend told you oh man solve it is the coolest thing ever it's not like you read a review it wasn't in the newspaper you didn't see it on TV a celebrity did like think of all the reasons that normally persuade Ordinary People to do a thing in with an early adopter we have to find a reason for them to do it in the absence of all of that validation and the reason we start startups with early adopters is a it's actually not almost impossible to start with anybody else so you may as well just accept that reality but the second reason this is the more subtle reason is um by finding the person who has the
23:43most urgent need for what you're offering and trying to give it to them first if they refuse you learn the most because if you are going at if you're going after any other psychographic segment you can always convince yourself that like well maybe I'm not talking to the the right customer but if you're like this person has like like imagine you're selling a new Med a cure to a disease like a terminal illness and you go to someone you're like hi I can cure your terminal illness and they like no thanks something has gone terribly
24:11terribly wrong in your thesis then of course the what's wrong is probably they don't have a terminal dness or your thing doesn't cure it or they don't believe I can think of a lot of things that could go wrong or like um my the example I used to use all the time is um teleportation most entrepreneurs think that their product is as good as tele portation so like therefore there's no need to Market it or sell it because it's just it's that good like who would want to have to Transit from point A to point B when you could teleport right like incredible and so we and so like part of the problem is that people's products aren't as good as teleportation
24:40but the other reason I love that example is I'm like okay let's imagine you actually invented teleportation okay hi walk walk up to a random person on the street hi I invented teleportation would you like to be the very first person to have their body physically dematerialized and rematerialized across the street uh you know it's like it's actually one of the if you could invent it one of the hardest products to sell ever like precisely because who on Earth would be the early adopter for that thing and
25:09like if you wouldn't buy if and I plenty of entrepreneurs are like oh I would never be the first person to use teleportation I'm like okay but teleportation has this almost infinite value and you won't even try it now how how valuable is your thing relative to teleportation like well but I I make your database like 45% more efficient I'm like I know but it's not teleportation and you can see the downside so like so that's kind of like that way of thinking just like to structurally understand who's the customer what do they want all that stuff is just is incredibly important
25:37this um we call it the anti-s sale what what Jeremy did with the um do not take this course if this is it by the way yeah do not take this course and so uh it's interesting because um Eric in the chat said the stuff you're talking about is similar and I can see this to BJ fog's Behavior design Focus only on those that can do what you want and are motivated to do it so like yeah don't take this if you think AI can replace programmers don't do this if you're only interested in building
26:06things by writing prompts um yeah and that seem to work quite well because for what is a really quirky strange new kind of [Music] tool that requires a certain amount of patience and it's being built around them as as we do it people have have done what makes great they they think it's an opportunity it's it's awesome to get to be the first to try something it's awesome that you get to that your feedback is heard by the development team so it actually like it's a very
26:35logical thing but um you have to find it and so anyway to go back to the answer AI story what I've always thought was compelling about this company is that Jeremy has such a compelling vision for what is going to happen with AI and what is likely to work so I can't tell you how many Founders have approached me and be like what can I do how can I make a company with AI and I'm like well how much do you know about AI they're like well I know I know agents are the next big thing or I know you know I heard that AGI is around the corner or some other vague statement about a it's like
27:05well it's going to be very difficult to come up with a vision that is likely to be true if you don't know what you're talking about about the underlying technology and you know some and that doesn't mean it won't work like a lot of newbies come into a new field and they get up to speed and they figure it out or sometimes ignorance is bliss so you know be somewhat distinctive too right because if it's the same thing everybody else is coming yeah if you're doing the same thing everyone's doing like what is the point of that um in the VC what they call that being non-consensus and right
27:31um but in in Jeremy's case for answer AI I thought it was very straightforward both the logic of the thesis of what what we were doing just made sense to me and you know it helps that he and I've been talking about it for ages so it wasn't like like I discovered it one day but like we've been talking about it talking about talking about it like there were historical analogies to draw on there were um you know parallels with other things that have worked in the past in other historical eras and there was also like a we have a very shared understanding both of where the market
28:00is going and also um a shared set of values about what we're trying to accomplish and the need to do it in a long-term way in a you know in a way that promotes human flourishing in a way that um you know resists these gravitational forces that we've seen drag down other companies like we've just we've been around long enough so like so that those anyway the reason I bring this up kind of now when you're asking me about this Lan startup stuff is my experience has been whenever someone is getting confused with some step of the process of doing
28:29entrepreneurship doing testing doing validation doing a testing AB testing or something like that the the error in thinking is always Upstream so like you know and when I do this in corporate settings we actually like teach people a formal process with like documentation and charts and two by two matrices and the kinds of stuff that you need to convince your boss in a 4 500 company to do a thing like I you can I've read a whole book about that you can read all about it and there it's like we actually formally identify what we call the leap of faith assump and we write them down and we actually have a
28:59prioritization scheme because like in in in you know with a 22-year-old entrepreneur you say what are leap of faith faith assumption some you be lucky if you get three in the Enterprise you what leap of faith assumptions you I've gotten a spreadsheet back with 800 things that must be true for this to work so we have a prioritization system okay what what's one of the most risky one of the most we look at the business plan we do a sensitivity analysis we call Innovation accounting this this whole thing anyway what are the leap of faith assumptions and we would pick one assumption and like what are the metrics
29:28we might be able to use to brainstorm to measure if this assumption is true and then pick a metric and now let's brainstorm one of the MVPs we have an MVP scoring system that scores the which MVPs are more or less likely to and um anyway I don't mean to make fun of that process it's just you know if there's any Enterprise people around here they're probably like oh where do I sign up for that yeah it's it's really very effective in an Enterprise context um but the reason it's necessary is that um people are used to
29:56working in this very formal way and we have to get we have to get them to unlearn a lot of formal infrastructure to get them to the to the realization um and so if they having trouble like a very common team I'll meet a team and they'll say like we're trying to we can't we can't come up with an MVP it's impossible and you're like what's the current plan our current plan is it's been five years and $125 million to build this awesome product after which we'll build a massive you know sales and distribution system to then sell it to lots of people in the eighth year of the plan and you're like you can't think of
30:25any possible way to shorten the timeline you know lower than five years are you sure no can't think of anything and you know I used to think oh man that's you're you know was something wrong with these people but like it's almost it's very easy Once you like okay let's back up what which leap of faith assumption are you trying to test they like all of them well okay no wonder because you're trying to test everything of course it's it's very difficult they're like what metrics are the most important here's all the metrics that have to be important and so anyway when we back up back up back up back up you know I tell the story in the book to Startup way did
30:54this in an Enterprise context where we took a product that was going to take five years and got it down to six months and what was fascinating about it was the executives in the room who watched me perform this magic trick were actually not very happy about it because they were like wait a minute a minute ago we were going to make like billions of dollars and now you're GNA sell one unit and how much money are we gon to make and I remember one of the engineers in the room was like oh no sorry sir we're going to lose money on the first unit and he's like looking at me
31:22like see you know this is so stupid and so so anyway so so it's it's like by identifying the error in thinking Upstream everything gets easier Downstream and so in in the case where you have a Founder whose vision is exceptionally clear and who has a commitment to the truth is not just ideologically attached to their Vision or or view or is defensive about kind of you know insisting that they must be correct this is actually very very easy to do so to me the The Lean Startup application and answer has been very easy so much so that we don't really
31:52need a lot of the formal methods of Lean Startup we don't use a lot of the jargon internally we don't waste a lot of time on you know decision analysis whatever because it's it's not difficult to do but again that's not because we don't do it it's because it's like deeply embedded into the structure of the organization such that it doesn't need doesn't need the training wheels anymore that's how I think that's interesting I mean that the the non-consensus part which we talked about a little bit about last week is is so difficult you know and you've personally put up with an
32:21experienced all of my doubts Eric and one of the things um you know I did early on in fact the very second post on our site is I I found the one guy in the world who was a historian of industrial R&D labs and he's great he's so you know he's passionate about it Terri doing it because he he just joined um uh I think it's like the Eric Schmidt new on new thing feature house that's
32:51right um uh and he spent a few weeks studying what we're trying to do and linking it up to you know the great industrial R&D Labs of the past um and I felt like that was necessary for me because you know as a Founder I always feel like I'm worried about my own reality Distortion field and like the idea that this thing that everybody knew
33:19in the late 19 century was the way to push forward a new broadly applicable technology being electricity like surely if that was a good idea to do for AI everybody would be doing it and the fact that I think it's a good idea and nobody else does I assumed meant that I was missing something you know so to kind of bring in a third party like that at least for me was very
33:49um encouraging to hear him say like no I I've been thinking for years that everybody's doing this wrong and they should do it exactly the same way that Thomas Edison did you know um so do you want to describe that structure I feel like for people listening we said oh Enterprise we have to do it this one complicated way but we have a different way and then we say some who looked at that way thinks it's good like what what is that structure that's so suited to The Lean Startup that's so suited to the
34:17R&D the fact the fact that we are an R&D company is itself almost unheard of and like Eric and I have had a few big investors reach out and say like oh apparently answer I is that new right you know okay how do I give you money and we talk to them and like they're like okay you know what's your MVP and what's your customer forecast and we're like we don't know what we're doing we're trying a bunch of stuff to try and figure things out you know and we're hoping over time that things will
34:45kind of turn up and they're like I'm sorry what uh can you send us your forecasts please um so yeah we're an R&D company which you know means we try to create this Loop where we build a thing you know or we try to build a thing that we think might be helpful you know and then we're like well can we do that really fast and then see if that thing actually is helpful and then it's like well not as helpful as it could be because these things aren't solved and
35:13you kind of do the research to do those things you know in our case we started even earlier which is like surely AI is useful seems like it should be useful doesn't seem to be useful for many things yet can we what's a way to figure out if it's useful like well we should hire a bunch of like really open-minded tinkerers like jono to try and build things and see if they can build something useful and see if they can build something to make it easier to
35:40build something useful um and then as Eric said one of the key Parts in that process also was to try to do this in a way that like it seemed almost certain in a normal company structure that we would not be able to keep going with that even with the enthusiasm of early investors at some point it's going to be like look none of my other companies look like this you know I I'm not comfortable with a company looking like
36:06this so you're going to have to pick a product come on be you know time to grow up so um you know Eric showed me a lot of stuff about how to make sure this vision and values persist you know and so one of the things I actually wanted to to hear from you about Eric is like really simple you know like the PBC right so you encouraged me to make this a public
36:36benefit Corporation and since that time a lot of people have asked me about that you know founders of us and I've always said to them oh my God I can't believe not everybody is running pbcs like it seems better for everybody for everything all the time you know could you like talk a bit about why you encouraged me to make answer yeah so this is we're we're we're deviating from one stream of knowledge and now entering into a new one and and I warn everybody
37:05as you get into this space there's something very strange going on here which is like it is an Open Secret that there are just these practices out there that are just better than what everyone else is doing and it's like not like it's a secret everyone everyone who studies these things and who practi like people who really are in the know all know that this is better and yet there is a huge amount of energy invested in convincing people that the conventional way we tell people to do things is the only and best way to do it which I always find so funny because it's
37:34already a contradiction since it was the only way there no no re to justify it as the best and vice versa so you can ask yourself as we talk about this why is it an Open Secret I think it's a really interesting question but to get to the to the basics um I always ask every founder to go through this hypothetical exercise I did it with Jeremy I've done it a gazillion trillion times uh who do you think is the most most evil company in the world you you can as you're as you're thinking along with me you can fill in whatever company you want it's funny the companies that
38:04people say when I ask this question there's not that many choices usually one of of a handful for me it's always Philip Morris I always I always think like tobacco is the most evil product that's ever been made in history but you could you can pick your own thing my you know my father was a pulmonologist so I grew up with the idea that cigarettes are evil so it's easy for me to think of cigarettes as a a totally evil product but you could pick a different company let's say that that company calls you up one day and says we would like to buy company from you right now cash on delivery for $1 more per share than it's
38:32currently worth would you like to sell it now some Founders are already like no answer is no more commonly Founders are like well what do they want to use it for okay I'll do the hypothetical sure like what's it for I'm like oh yeah to sell cigarettes to children that's that's their explicit plan they sell you right up straight up that's what they want to do what's your answer now every I've never found who was like sure I'll sell I don't care it's like of course not of course we're not for sell make excellent good now let's go pull your Charter for from Delaware it's a it's a public document let's go read what it says here oh look
39:01here did you know that according to the Charter that has your signature on it that you yourself signed most corporate governance experts in the world would say that you have what's called a fiduciary duty not just to consider this offer but to take it and if you don't you could be sued by your investors for a breach of the duty and every founder I'm like that first reaction is like that can't possibly be right you must be mistaken it sounds so incred incredibly stupid it can't possibly be right and I I've had to do with many people where I say look call your a lawyer ask him this
39:31question straight up and then call me back and they call me back like he betrayed me how could my guy my guy my buddy my lawyer sign me up for this horrible thing how and I'm like look he thinks he's doing you a favor he thinks he set you up for success but on the other hand you didn't tell him this is what you wanted did you and they're like well I never even heard of it so anyway so people get so alarmed like oh my God this sounds like an intractable problem how will I ever fix it and then you're like well actually thanks to the some wonderful activism
39:59that was done in the Delaware Chancery Court uh you know this is not like 20ish years ago you can make a two-page legal filing in Delaware to just say that you're not going to do this it's not like some complicated we added like yeah we added like one paragraph to our chat it's literally you I mean it is one of the easiest legal filings you will ever make and it obliges you uh once every two years to make a special report to your investors to let them know that you're still doing this I it's like couldn't be the EAS iier process to uh
40:29do it's called becoming a public benefit Corp or PPC because this is a common point of confusion this is not the little B with a circle around it that you sometimes will see at the farmers market that's called being a BC Corp which is a related but different thing and that has a whole different process for getting it that's more of a trademark thing all public benefit Corp does is allows you to to say legally to your investors the purpose of this Corporation is to make money by doing this so instead of saying we'll do anything to make money we're just saying
40:58we make money by doing this you in in answer's case we are trying to discover uh beneficial uses of AI you know for the benefit of all Humanity I don't exactly remember the exact phrasing but like something like that Society beneficial uses of AI y yeah Society beneficial use of AI I should probably memorize it um but like the funny thing is these things doesn't come up that often because we've internalized that into the mission of the company we don't have to like consult the legal agreement but it makes it clear now sometimes people think oh well I guess there's the one weird trick to be a good company
41:28from now on I'll just go do the filing but remember the filing only changes the basis by which an investor could or couldn't sue you it doesn't solve this problem in any kind of like comprehensive way it's just like a complete no-brainer of a thing I just want to mention by the way that that that sentence so just for people not familiar with this a public benefit Corporation part of that one paragraph is to say well what is what is the benefit what is you know what what in your Charter you have to write down in a
41:55sentence so that's what this sentence is and although as Eric says we don't talk about it all the time it it does come up like like because I quite like to go back to First principles as you know and so quite often in our regular company chats we'll have a conversation like maybe we'll come back to this in a moment like okay solvit seems good we or be able to use that to do a better job of making products that we can sell to
42:24people tell me again why would we provide this to anybody else you know and teach them to use it and the starting point for the answer for me is like okay well let's think what's what's our stated public benefit to you know help create more and better societally beneficial usages of AI would releasing Sol it to the community help us to do that oh we think it probably would
42:52so okay cool you know it's it's it's not bad it's kind of nice to have a bit of a you know single I I think it's enormously helpful it's like Jon o can confirm this it's been enormously helpful in recruiting Because the best people want to join a company that stands for something and that that's actually trying to accomplish something good and people who are the kind the caliber of people we want doing the kind of AI research that we are talking about can work basically wherever they want so nearly didn't reach out to Jonah because
43:22he had told me he he would never let any company hire him and I yeah exactly a shame because I would like to hire you and he was like oh no no I mean the company other than answer AI yeah there's there's an Asis there but it was a very fun I had to like unpublish the blog post I published the day before of like why I'm not going to take a full-time job for an a company I'm like oh okay fine yeah yeah and not to listen I don't mean to embarrass jonno but but as a Founder that's what you those are the people you really want to hire you want people hire people that are
43:51completely unhirable not because they can't get a job but because they can do whatever they want and now you're like look that's so rude no right yeah like I'm I'm really serious about it it's actually like really really really important and most startups struggle with this because they're like well why should someone like that work for me I can't give them more money than Google or open I I can give them I can't give them more this and that I don't have as much Prestige I don't have this and I like but so what do you have and a lot of startups um forfeit their most
44:20valuable resource which is like their moral authority to do something incred like to be trying to do something incredible um and to take that seriously like you know when you're hiring someone just out of college maybe you kind of fudge it a little bit and be like yeah yeah we use the rhetoric of changing the world for the better but there's no substance to it but when you're dealing with someone who actually knows what they're talking about and is like a little more sophisticated you have to be able to back that up with real changes structural changes so we we call that institutionalizing your intentions so it's not just like oh I intend to be a
44:49good person well good hooray like think about I think about like if a friend came to you like I intend to go on a diet and become a bodybuilder it's like great I'm so glad to hear that but like what what's your plan like well I ate a really healthy meal this morning it's like well do you have any more is anything else you anything else to tell me about that right so most companies are like that they say all kinds of good stuff but they don't really care and again the thought exercise I like to have people do is think about the last time someone who worked for a company apologized to you you know you were on the whole your call up customer service
45:18because some bad thing happened to you and the person's like I'm so sorry Mr so and so like I will be sure to take and like do you actually believe that the is sorry that this happened most of the time you don't even think the person talking to you is sorry let alone the and it's like what would it mean for a company to have an emotion and the emotion is we actually do care about something the thing that we're doing most companies are basically sociopathic they don't really care about anything
45:46but enriching themselves and so you can tell like you can tell as their customer let all know their employee and it's a really powerful thing when you can put together a company with such integrity that it actually genuinely cares about stuff and ideally it cares about the people who make the things too not just the thing itself so like you we've tried to build an an company that has actual genuine feelings for its employees for the customers that it touches and for the wider world like we are genuinely concerned with the broader implications of the things that we make and that
46:16feeling is not the customer is always right and we do anything for them and whatever do you know what I mean like we yeah you know another very intentional thing that the grp watching now will have seen is that for 3 weeks on the homepage after you logged in the largest button you could see was the one marked cancel because we were not trying to make something for everybody we wanted everybody to have a chance to see whether the thing they we made might be for them and if it wasn't
46:45we we're not going to change us to be the thing that person wants we're just like okay yeah you tried it now you can go yeah again like by analogy to like what makes a good parent what what a good anal like how what what makes a healthy relationship like it's not about just being blindly obedient to what the other person says it's about well I hope people who are listening have enough experience with this in their life to know what it's about well so there's no reason why um companies can't have those same attributes and
47:13again this is one of these open Secrets where most people when they hear this for the first time are like no that's not true companies have to be this way they have to be bad we all know that I've been told that so many times oh so many times I've been all the time but like but a there are these count there are these weird counter examples where there are these companies that like aren't bad you're like well if it can't be done how come they can do it and then people have these all kinds of weird just so stories about why they can do it and you can't but like the other thing is that if you
47:42go to the people who study these matters professionally they're all like oh yeah obviously everyone knows that it's like well no not everyone knows that in fact only people in your Naros specialty know it but like just you know we could get into it there's a lot of research backing this up that basically trustworthiness is a strategic asset and source of competitive Advantage for companies uh and so by by trying to do something extraordinarily Difficult by really committing yourself in a values way to do it for real by having a
48:10long-term uh commitment to that thing and by doing these other kinds of structural practices like being a PBC like that that actually makes the company more strong not less strong and therefore you like our kind of like modern idea that like ruthlessness in capitalism means like I squeeze and exploit to the maximum and if I don't do that someone else will do that more than me and displace me out of business has it exactly backwards we're like racing to the bottom to destroy ourselves and the evidence is all around us I don't think I have to decide it but if anyone is interested I'm happy to answer
48:39questions about what that evidence looks like it's very strong in the chat bit Cloud 2 said that refund button freaked me out I think people were worried about accidentally pressing it was so big but don't worry if you had it would have offered you a confirmation I was like I said to my team I was like okay no matter what we do we will have people saying I'm sorry I accidentally pressed the button and even though like after you pressed it it popped up a big scen saying like Okay if you press this second button this is what's going to happen you'll be out you
49:09can't take the course anymore you can't use this anymore we can't restore anything it's like we're going to delete you it's like are you sure you still want to press a button we got like quite a few people saying like oh I accidentally pressed the refund button welcome to humanity we Tred happens every time hey um I I did want to talk more about solid actually because like um like
49:37this it is an interesting question because uh when I rang you Eric and I said like you know uh jonno and I want to teach a course you know how to solve it with code course um do you think it's going to be a month do you think we should build from scratch a new platform and use it for the course based on our you know
50:05kind of internal prototypes we had created uh and you know thinking that you would be responsible enough to say don't be ridiculous that's the stupidest thing I've ever heard but no you didn't you said what a great idea Jeremy uh and particularly because this would be a great opportunity to to to create a the you know the the founding Community around this this thesis this idea of how to work with code and how to work with AI
50:35um so like could could you give your perspective here on like what's the point of solve it for answer AI what's the point of solve it for these people who are now using it and what's the point of this you know Community which will help us as well think about like in a you know whatever A month's time as we unfortunately wrap this up how do we yeah what next what next make sure we C
51:04my mind quite quite a bit well you know I it's I'm glad that we're talking about this in the context of the founding of ani story because to me this is a very straight forward application of the first principles so we've talked already about the public benefit of bringing this idea to a lot of new people so that that's like good in and of itself um there's a strategic value to answer AI also which I think is important to talk about um i' to we have this concept so so when Jeremy says that we're an R&D company um people most people don't
51:34really know what that means so let me justar lab yeah yeah like today the way that research is done in universities and most research Labs like really doesn't make that much sense and um is not really good from a historical perspective not that there's not good research being done but it's not the best structure for research to be done because today research and development are separated into two separate activities and so there's like research and developers and they do two different things and every once in a while they throw work across the transm to each other which is not I mean can't say that's bad much of the modern world is
52:03built that way but there was an older idea that research and development should be co- um codependent done simultaneously by the same people um and so um there are huge advantages to doing that again there's plenty of research on this point but I also I've I've worked in actual research Labs including the successor to Thomas Edison's actual research lab at at GE I've seen what it's become now and like again wonderful work being done there but this um this connection has been generally lost in the research landscape so we really have
52:31this idea that we should be using the the commercialization of the products that are being researched as a way to inform the research and just to give trivial examples of why that would be a good idea I have worked with many many researchers who have discovered something that nobody cares about doesn't matter or more likely they discovered the wrong thing so like when we're trying to do applied research we're trying to not just learn things in the abstract but trying to learn something that is useful to accomplish some goal so like I I never forget one
53:00of my seminal uh R&D experiences was working on a a piece of equipment that was going to be used in data centers i w go into all the reasons why but like this research team had spent years to make a dramatically more energy efficient piece of equipment that would go into Data Centers and um yes it was much physically larger than the previous version but it was much more energy efficient and they went to people who build data centers for living and I insisted I was like look we I know you've been working on this thing for years but we just need to double check that the attributes that you're
53:30optimizing for are in fact the attributes that the customer cares about they're like what else could they possibly care about I like I don't know I've never built a data center in my life but let's whatever it is let's go double reality D field I got let's go double check and of course I wouldn't be telling the story except that the they the people who build data centers don't give a crap about Energy Efficiency because they don't pay for the electricity that's somebody else's problem but what they do care about is the footprint on the floor that's actually the only thing they care about and so all this research had gone to waste building a product that was was 10 times better in a dimension that the
53:58customer didn't care about that's like so classic it's actually really sad and if you tour a modern research lab um that is any good you will have that little like tiny like mournful violin sound playing in the background of the soundtrack as you walk around as you meet researcher after researcher who's like yeah I have a cure for cancer on the Shelf here I have a you know this is like infinite energy for free here's Nan materials that would allow us to make a bridge from America to China it's just like one extraordinary breakthrough after another like have any of these uh
54:26been commercialized like no they didn't want it they didn't want it they didn't want it they didn't want it it's just like it's unbelievable the potential that is sitting in these Labs that is never commercialized now right if you meet the people whose job it is to commercialize that stuff they'll be like those eggh heads you know they wouldn't know like just like you know everyone's pointing the finger like this whose fault is it really does the thing actually cure cancer like like so anyway so our concept with n was to build uh a laboratory that does both the way that Thomas Edison basically invented the modern electrical system uh through this
54:55mechanism because it really forces you to direct the research to the things that are the most important so we call the collection of deployed research infrastructure the substrate the thing that allows us to build um uh AI applications faster easier and in a self-maintaining way um that's that's like a thing that we need to build so we have to build stuff and ship it so building solv it is like part of that substrate or we hope it will be yeah
55:24yeah we we there like something we we need needed ourselves and it was also a great way to test out we had done a lot of research in the preceding year of things you know much of which has been open source and not we only build applications but I was like look this's a great opportunity to see if the if the pieces of the puzzle actually come together into the development of um of an application and you you are using the living result of that you can judge for yourself whether in fact it did or not but I think it's pretty good for a month's work and the question was really like are there some people in the world
55:51who find this useful and you know and if so what do they find it useful for and yeah it wasn't like it as many as possible I mean we only opened up signups for 48 hours much to the sugarin of many um tried to convince people most people not to sign up um certainly wasn't for money because we did it at like 10 times less than by fast AI courses which were much less breakthrough than this well and I use I I use this I used solve it all the time now as an example
56:21of MVP thinking applied to AI applications now take connecting back to our previous lesson because I want each of you who are listening to this to ask yourself like really really ask yourself if the announcement about solvit had been new whizbang code editing platform now available for $200 would you have rushed to buy it I don't know I think by making it into a course where the customer is guaranteed to get some utility and where
56:49they're going to actually learn how to do the thing um you have a much greater chance of attracting the right kinds of early adopters to this product um and you get to really disentangle the question of do people want the benefit that is being offered from The Limited nature of the software so again you could imagine a situation where we had lots of people sign up for this course and lots of people cancel after two weeks because the product sucks that would have been extraordinarily valuable learning you can also Imagine a scenario where we put this out there for 48 hours
57:18and nobody signed up you could imagine a lot of scenarios like a lot of things could have happened and we had a pretty specific idea of what we hoped would happen and we very pleased to see what has happened happened um so that's kind of like so we talked about like the public benefit then this is like the Strategic Benefit then from a third thing is is why have a community I think is also really important we also see this as the promulgation like in some ways the product here is not solv it like solv it is like the container the petri dish that allows this community to
57:47form around it and if you look at the history of Technology platforms and developer tools that have been really effective in changing the world they all have this this passionate community that that pushes forward the envelope of what is possible with the thing and also evangelizes it to other people and brings them into uh the way of thinking so the community is like a living organism that grows um and metabolizes the world into itself and so for this we you know we have this revolutionary idea about how programming should be done in
58:15the era to come and so even if we could just use it ourselves privately like the impact of that compared to having a thriving Community um is really different and then and then of course like from a qualitative feedback point of view a community is dramatically better than any kind of split testing surveys you know just every conceivable kind of other substitute it's much better to be living embedded in a community of people who can tell you directly what they think and show you what they make and fact Tech how do you how will you know what
58:44kinds of people are using solv it how will you know what they're using it for how will you know like what are where is your Telemetry what are the metrics like well I I know them we chat you know like like um it it's it's not complicated you know it's like uh you know and again that is made possible by the decision we made Upstream of this as I Mo through the from first principles right because we made the Upstream decision not to have too many people using it that's actually
59:14really important analytics are for situations where you have too many customers to know them individually right if you have 20 million customers you can't ask them all what they think it's imp no human could possibly do that you don't live long enough so have to use statistics um STS are a very imperfect way to determine what people want and you know it's better than nothing of course but in a community setting in a limited setting we can get much better intelligence and of course as the solvic community grows we probably will have to trade intimate knowledge of every customer for more
59:43analytical stuff that's like a natural thing that will happen but you don't want to start with that if you can avoid it so is it it's not just for the feedback and so forth like there I'm having trouble I'm sure you'll be to capture this much better than than me like I I know from Fast AI like there was something super valuable about having that community in that case it was really the first community of kind of AI research practitioners you know and and nowadays I find most AI
1:00:13researchers I come across are fast AI alumni um and there's a certain kind of shared understanding of how one does AI um yeah I don't know how do you yeah what are the other things you kind of see coming out of of this community because another question we had um during the week was uh from one of the students how can we as individuals or companies of One support answer ai's
1:00:42goal of having a long-term positive impact on society oh I love that question oh that's such a powerful question actually I don't I people will think I I planted it but I didn't I really didn't um so this is something again of these open Secrets um people in today's world feel powerless it's very common to feel like we are um caught in these big forces big platforms big this big that all these billionaires you know controlling the media or whatever you I've heard so many
1:01:11theories of what's the root cause of why we people feel that way and we don't have to get into that today but I think that's not true we are much more powerful than we think and it actually is a people make a concerted effort to make us feel powerless it's actually like by Design design again you can think about why people would want that um and the way you know that's true is think about all these powerful people and all these super powerful corporations whatever how much energy they spend trying to be loved and
1:01:39approved of by the public like how much like you know if you have a VI billion dollars you could spend your time on a private island you're doing nothing you don't you certainly don't need to be out there trying to get people to love you and say you're doing a good job um so like so why why why why and so the the the approval that people give is actually like a very important gravitational force that shapes the world in ways that people do not fully appreciate and that's true as consumers it's true as Citizens but it's also true
1:02:07as employees and it's true as customers so by when the reason we talk about something like this as a movement as a community rather than as a product is because we're really consciously trying to get this idea to spread and so by acting in line with the values of this Community you you amplify it without having necessarily to do anything else so people like that sounds like virtue signaling yeah doing things that are
1:02:35virtuous is good like and signaling that you're doing them to other people is a very pro-social thing it's actually the foundation of civilization like I don't know I don't how virtual became an insult it's good so like by living in line with these values you are doing the thing like if you want to also write a blog post about it and go on social media and recruit other people like you can do that if that is the thing you feel so inclined or still somebody might call you a social justice Warrior Eric what if you're fighting for True social justice that would be just so embarrassing I mean what could what could be worse than Justice let me tell
1:03:05you um it's it's bizarre so now I really lost my train of thought and I'm like oh all things we're not supposed to talk so we have we have users so excited and saying like how can we contribute oh yeah how can I help so so yeah so by by embodying this ethos you spread it you naturally signal it to the people around you so for example let's say like you know there's been a lot of talk lately about the how long will it take for the for there to be a oneperson billion dollar company to be
1:03:33created and maybe that happen in one year in five years or whatever imagine that was you imagine that you used solv it to create something new and that like created this incredible amount of value in the world you don't need to tell anybody that you use SOL it to do it like it will be a massive piece of news it will create tremendous gravity as other people seek to emulate the thing that you did but like let's say you also Ed this to become a nurse and you took care of people in a hospital in a better way or some other
1:04:01like deeply intimate personal kind of caring that didn't you didn't make a billion dollars but you nonetheless use this EOS to acquire like that also has profound impact and gravitational pull to the people around you so that's and I think that's the hardest thing for people who are in the kind of company of one or who are truly an entrepreneurs you have actually a lot more control over what you do so like so for you it's it's much more can you overcome the mental bias that just we all have been conditioned I would say even indoctrinated to do things a certain way
1:04:30and now we're asking you to do something different it can be difficult if you're in an organizational structure you have to find ways to integrate that into the dominant structure so like I I just I just gave you one funny story about that um for a for a Time many years ago there was a group of Google product managers that had a secret meeting in my living room where we would secretly discuss le start in the Google context Google is like at that time was the ultimate not
1:04:59invented here company and you could not go anywhere in the company if you brought in an idea from the outside it's ironic for a company that's supposed to be so open sourc but like that meant Google publishes their open source that other people use not the other way around and so these product managers were like really struggling to get Google to do Lean Startup stuff and part of the reason is they couldn't tell anybody that it was LE started they had to come up with with original Google isms to explain why they were doing what they're doing it was it was totally hilarious and it was just like this this
1:05:28incredibly cumbersome way of trying to like and they thought when they first approached me they thought I was going to be really angry about it because they're like sorry we're not giving you any credit we're just stealing your ideas and rebranding them for the Google context like is that okay and I was like come to my house and we'll talk about how to do it better like I don't like that doesn't matter it doesn't matter whether you call it this or call it they call it whatever you want what you need to do is find ways to actually live the thing so that you can demonstrate its success and then if later you decide you want to be an Evangel for it which of course is wonderful you'll be a much
1:05:58more effective evangelist if you can point to examples of how it's actually worked in your own life it's so much better than like oh I heard Eric and Jeremy talking about this thing on a video and you should watch the video too like no one cares but if you could say look I was able to accomplish something with this that's a much more powerful testimonial a couple of people actually asked about this in in the in our last chat about uh being at bigger companies maybe I just pick one of them as being the most strongest version he said one experience I have at my
1:06:27workplace is that all our internal users essentially push back and tend to reject every change no matter what so we need weeks of training and getting on board and somebody else said yeah large dogs struggle with bureaucracy and communication challenges yep yep yep so for folks I did write a whole book about that we haven't talked about so go check it out it's called the startup way yeah no problem just it's called the startup way
1:06:56it's funny because most people in the startup World know me really for Lean Startup but if you go into the corporate world I'm actually known as the author of this other book um that's as famous to them as Lean Startup is to to entrepreneurs it's it's funny like these groups don't talk to each other um but for me it was like to to to same intellectual challenge are just like what what what happens to startups as they get bigger today's Enterprises are yesterday's startups and I I mean I used to ask people you know people a lot of people start big start startups because they hate working in a big company and I used to tease them a lot i' be like if you hate big companies so much why you
1:07:26trying to create a new one and they're like well my big company's not going to do that stuff and it's like oh really what's your plan and it's like yeah it's like the person who's going on a diet you know I ate a healthy meal this morning so therefore I'm not going to get you know overweight later it's like might want to have more of a plan than that so I wrote a book about it um I will say the in an Enterprise context it it can seem very daunting to do this things because there's like a dominant way that the company's always
1:07:56done things there's already some level of bureaucracy you know and and some companies are too far gone for reform honestly and if you work at such a place like if you can do solve it you can get a better job you don't need to stay there it's okay but if you want it is like people particularly people at big companies who've been out there for a while I think really don't understand how many opportunities are out there yeah yeah really you should don't yeah don't feel trapped um and we can talk about that if people want tips on on how
1:08:25to do that but like U but a lot of people stay at big companies because they're afraid of being fired and they accept a lot of abuse in the because the idea of being laid off is too painful to consider but they also don't want the hassle they have a life they don't want to go like searching for other jobs so one thing I found extremely liberating in my own career I think we been maybe talked about this last week is like trying to figure out what you stand for to the degree that you would be willing to be fired over it I I work with a lot of corporate Executives okay I work with CEOs and
1:08:53boards and very high-end executive VPS or whatever if you get them alone privately and just ask them about this they'll all be like man I wish we fired people around here I wish it was easy to fire somebody like people don't get fired nearly as often as you think that they do and so one of the things that can be very liberating is just to be like look I'm going to try standing for something and let the chips fall where they may if it means that I get fired like don't worry you got if you're if you're here I'm very confident in your career prospects really truly and and like it really is like office space you
1:09:22know that great movie where the guy like gets hypnotized and never gets DEH hypnotized and he's toally just doesn't give a about anything and he goes into work and he's just you know he just it's like I don't even know what we're doing here you know it's like what do you mean by that well I don't know it seems like our management has no clue when these reports are pointless and I'm just wasting my time with them people are like oh my God he's going to get fired but like no quite the opposite people so refreshing this truth telling tell us more yeah now don't be an a-hole
1:09:52about it you know don't that that that's part of the movie that's not realistic but what is realistic is to to then try to find an executive sponsor who actually cares about results so like all all corporate transformation begin with an executive sponsor who cares about outcomes and some people listen this are so cynical about Executives they're like Executives don't care about outcom they care about riching themselves and again if that's how you feel I really would consider finding a new job because there are people who went into industry to do
1:10:22something to actually accomplish something there are people who really care about that and believe in it and are really frustrated by the same bureaucratic stuff that you are trapped in in fact many of the people who promulgate these bureaucratic rules themselves don't like them but feel trapped and they feel like they have to promate them for the same reason you feel like you have to obey them so if you can find someone you can say look I used to call this the golden sword for cutting through bureaucracy which is if you could find someone who actually cares about results you say listen I'd like to
1:10:50make you a deal and the deal is as follows in exchange for you give me the latitude to pursue this better way of working on this project or in this in this limited circumstance you give me the you have to be very explicit about the latitude that you are asking for I need you to wave this requirement or I need you to get this person off my back or promise me that this thing will happen or really this if you give me this don't ask for money but ask for anything else if you do that what I will offer you instead is not just like a
1:11:19blanket guarantee that will get a lot done in 18 months cuz who knows but I will give you very fine grain detailed report wordss that are scientifically sound where you can see you can witness our progress for yourself and decide if you want to keep going with the experiment and shut it down at any time and I have been astonished at what people have been able to get done with this simple formula but finding the person is not easy and and and I know this because you know in my shameful past I
1:11:48used to work in management consulting and um made you know a couple of really big companies one really small company the universal thing amongst all the companies was the first month of a two-month project there would be at least two people in the project team working pretty much full-time on uh organizational mapping and their job was not to figure out how to solve the client's problem their job was to figure out who in the company had What feelings about the thing that they're trying to
1:12:17do and how did they respond to new ideas and which people tried to keep everything the same and which people were into change and which people's promotion kpis were based on finding new ideas and like that was the job and and then would come back and be like okay we have to find one person so same deal as what you're saying we have to find the one person who's going to be our cheerleader and then we have to show that person that we can help them succeed in their career and then now you've got somebody on the inside on the
1:12:46team and then and then suddenly everything becomes easier because they're like oh yeah all right this is going to be a great project great project look out for that Eric guy you know he's totally anti- Innovation you know he's going you know but you know the good news is he's going on a three Monon sabatical next month that's the time we're going to come that's that's the kind of Intel you have to have and I I will say that um most this this information is more known than people admit to knowing so like once you get
1:13:15people who get to a certain level in these organizations have what I call the little black book they actually do know like I always s with next say listen if hypothetically the CEO went to like a a tech a retreat business retreat for a week and came back with a cool new idea they're always like oh God You Know What's it gonna be this time okay but and said listen your life depends on making this magical thing happen for me under these very adverse circumstances High state or what who would you call to assign that
1:13:45project to would you send it would you put it through your normal product development process would you do your normal assignment process every is like of course not that's like career suicide you can't put that you can't ask a normal person to do that they're going to run for the hills like well who would you ask everyone's got this little notebook and they're like well what you really want is joeo certain person three levels down for me over here and you're like who's that guy like he we've tried to fire him like 10 times he's impossible to work with everybody hates
1:14:12him but doesn't mind career suicide assignments he actually really likes projects like that he knows how to get this kind of thing done call soand so so like there is this like hidden network in most organizations which is like how you when when something must get done who do you call and sometimes it's a fancy VP and sometimes it's like you know it's like the office space guy with the stapler except if he actually did real work like there are those guy like in the B there's just a guy in the basement you have to know to call I I've been an actual basements where the
1:14:41actual guy actually is there and it's like oh I've been waiting for someone to stop by and ask me to do this project for a long time it's like no one thought to do it I don't leave my basement so I would never have any way of telling anybody to do it the swing line the swing line
1:14:58stapler um you you um mentioned uh metrics um quite a bit and and that's something I'm super interested in like it's something I still find surprising about our relationship here Eric which is from from like a somewhat unnuanced view or from a surface view you would think that we would be butting heads about metrics all the time because I think a lot of people feel like oh
1:15:26Eric ab tests exact percentages specific hypotheses and like well else I'm the guy like who you know my wife Rachel wrote an academic paper the problem with metrics is a fundamental problem for AI you know we we're it's kind of known as anti metrics oh yeah but actually we don't at all and and I think what's interesting like so we had quite a few questions from last week you know one I think one which maybe really gets the heart of was like
1:15:55okay what about Apple you know design should be based on taste and intuition you know and that's very much me like with with all due respect to the people listening I don't really listen to customers you know like if if you had said like hey what could you do to have a better you know to improve you know to have a better AI help with your code like oh well I like Claude but I don't like the serif font you know maybe it could be San serif or you know well what
1:16:23about cursor I want the latency to be a bit longer on the ghost text nobody would have said like oh trick my brain into writing less by using a smaller text box and put it further away from the main thing so I don't edit things so much and I'm all create a dialogue and you know try to trick my brain into thinking this ISB you know yeah to me this is very straightforward like if you want a pastry you don't ask an electrician it doesn't mean that electrician's not good at their job it doesn't mean a pastry Chef's not good at their job but like
1:16:53the customer's job is not to design the product that's your job so so asking for people to tell you what to do is just a confusion of roles it doesn't mean that the people aren't smart they're probably really good at their job but that's not their job that's your job so how do so so I mean some some some other questions in this General category of like well you know what about when metrics are really jumpy what about the bootstrapping problem when you kind of need experience from the thing to know how it's going um you know what
1:17:21if there's a really long lag like I I think a lot of people assume that maybe you think that all metrics have to be like download counts or something but I think that's natural that's natural it's it's very hard um because the need for analytics like analytics are an exhaust it's like one of the last steps of the process and people think oh M metrics is better but I don't think that at all we need the necessary metrics to accomplish
1:17:49the goal so I think the big part of the problem is that most people have learned metrics and a lot of people even have learned statistics from a certain very narrow number of use cases that are not that helpful for startups presidential polling right where the goal where everyone knows like there's a relationship with sample size and accuracy and we in presidential polling or you know parliamentary polling we're trying to we're trying to find Fair fine grained changes in people's in something
1:18:19very nebulous and so we need these big sample sizes and it's quite it's quite difficult the other context is tend to be like people who write a lot about metric and talk about it tend to be big company product managers who are dealing with millions of customers and again you're like You' got a new version you're trying to figure out like big statistical changes that are relatively minor what people forget is that statistics and sample size you know whether you're a basian or a frequentist either way um the fundamental thing we're looking at is what is the
1:18:48underlying signal to noise ratio of the phenomena in question if you have a very high signal to noise ratio you do not need fine grained analytics to find out if it's true um so for example um I don't know if you told this story last time uh but early on in my career uh we were building a product and one of my co-founders jobs was to bring in people for usability testing and we were really proud of ourselves like we're going to be customer Centric and we're going to get feedback about our product and so we like I'll never forget
1:19:17as long as I live we brought the first person in to use our product and we sat down and we explained the value prop and we said okay you need to like log in with your social credentials here and download the product and here's how it goes and they were like no thanks they were like I'm sorry I don't think you understand how this works we're paying you to be here so you kind of have to do that and they're like no you can keep your money it's okay so like our product was so bad we literally couldn't pay people to download it that's how bad it was now the first time that happened I turned to my co-founder
1:19:46I was like you are the worst how on Earth where did you find this person of all the people on planet Earth you find get me another one second person comes in you'll be shocked the same thing happens again and now I'm like looking at him like you really are bad at this like come on can you please find me someone to use my product for me third time happens happens again now any of you who are Engineers know exactly what's coming next I was like well is that a statistically significant
1:20:14sample and like yeah it is when the signal is very high you do not need a high sample to get it so yeah like we don't need for the stuff we're right we don't need massive samples it's very clear the evidence is very clear and when you have that kind of direct intimate connection with the customer you don't need statistics you don't need sampling you don't need that stff that stuff is for it's like useful in very specific situations now I've been in those situations a lot and written about them a lot so of course I'm associated with those techniques and there are times when you do that is really
1:20:44important so I'm not here to say like you should never use it it's just this is like there's a huge Suite of tools in the toolbox you can use to get through the build measure learn feedback loop and part of the the art of Entrepreneurship is you because you get older you get better at figuring out what is the right Tool uh the right tool for the job and people who want to make it about analytics or not measurement or not metrics or not are like trying it's like you know I don't know what the right I'm trying to come up with the sports analogy but it's like you know is
1:21:12basketball about passing the ball or not it's like that's not helpful it there there is some passing of the ball in basketball yeah some teams passing more some I see a lot about it like we could make a whole theory about it but like then you're like well what about tennis is tennis about passing the ball it's like no it's not it's like well see see if you believed in metrics you believe in passing the ball all the time it's like no you do the thing for the thing you're doing so so it really is about choosing the right thing for the job and also recognizing that that thing is going to change as the product evolves
1:21:42so like I said we have a thousand people in the solvic course or whatever it is we have a certain size Community we have a certain set of prct makes sense if we had a 100,000 people it wouldn't be the same if we had 100 million people it would really really really not work do it the way we're doing it now like we obviously would have to make changes as we go but it all goes back to what question are you trying to answer and when you're trying to answer the really really coar Grand questions will anybody use it do they like it does it comp like that's very different than answering the fine Grand question which are also important but just different of like
1:22:11which button is more efficacious how much should we charge for it you know we definitely saw it this week it was great you know um my old friend Ric who's well known for his meta learning book you know took a Time time but finally tried solve it a week ago he was like comes on the Forum like hi I'm new here I'm going to give this thing a try see how it goes and then within the over the next week he basically posted every day with a new thing he's built he's never built before
1:22:40because of so like okay yeah that's very clear it's not really confusing we don't need to AB test it we don't need to like it's like well what did the Google analytics say about a session it doesn't matter that's not that's not important we have the information that we need we're able to get it Jon I think there some questions you wanted to kind of ask both of us or is that right or ask Eric or tell me you know I I don't want to monopolize things yeah I'm trying to go back through like one one one theme that I keep coming
1:23:09back to even when I was reflecting on last week's sessions with you Eric is like this this idea really applies a lot more scales and topics than I was thinking right so so you were talking about like okay measure eval like this little loop it's not just the like the signups that you're talking about or whatever it's also like you're talking about the trustworthiness of a of a company Long Term right like if we're trying to attract the best talent but then we are
1:23:37breaking against our ideology and providing stuff to the military that that's not like that's not showing up in your weekly metrics and in fact you can trade off those longterm things for these shortterm things um so yeah I I guess
1:23:53like what maybe like as a as a homework for people watching and as an exercise for me and I want to go like assign this to my buddies as well like what is what are some ways to start reflecting on that and defining those Loops right so like what is it that I what are the questions that I ask myself to say like what are my core values but also to find these places where oh I would have thought measure applied to like the speed of my code or the number of downloads or whatever but like you know it also
1:24:20applies to the company values and the the data they working or you know like all these different places yeah I don't know how do I interrogate that space I guess is the the best question like what's the process for finding all of the places where the same kind of idea applies yeah it's fractal really um and if I don't if anyone seen the old Ames Movie powers of 10 those that know what I'm talking about right where like camera zooms into the microscopic and
1:24:49then zooms out to the macroscopic and like um I've been thinking about this a lot because one of the questions that comes up a lot in my work is like what is an organization like what is it what are we actually doing here when we do work and there are so many definitions out there it's actually kind of interesting and people tell the story of a startup like this you're like well it starts with a vision an individual person has a vision and they bring this new idea into life and from the vision you know they acquire resources and then
1:25:18from the you know they got resources then they they enact some kind of strategy the strategy gives rise to a business model the business model gives rise to to a product the product uh is used by customers and so it's like Visionary at the top customers at the bottom and at every one of those levels there's some thinker out there who is to said like Peter Ducker said the purpose of a company is to make a customer so like oh but like if you go to Harvard Business School you say what is a startup they have a definition it's like a startup is the um uh an organization that make like makes plans based on
1:25:47resources not currently controlled so the acquisition of the resources is the thing obviously like people who are in the like Cult of Steve Jobs are like it's all about the vision so anyway so there all these different ideas about what it is and you can even tell the story completely in Reverse people who don't like entrepreneurship or who kind of have come from more economics perspective would be like that's backwards the first thing that happens is customers have a need the market has a need the need causes a product to arise like now that the product has arised somebody has to like we have to build organizations to service the
1:26:16business model that makes it possible to make the product that business model requires the acquisition of resources which requires some kind of strategy and like in and like some body has to be the one to capture the economic rents having the privilege of of filling this hole in the marketplace and that's you know the entrepreneur is like the the trickster the huster who shows up at the last second to claim these social profits and it's like that's ex it's the exact same story in reverse and I think what what's really fun I like to do this is like you draw that diagram top to bottom and then
1:26:45the reverse one around and and ask yourself like which which story is true which level is true which definition is true and it's kind of frustrating like well these are actually all true true and as soon as you have that realization then you're like oh this is not a line but a circle if I connect if not notice that everything starts with the vision and ends with the customer or vice versa but like the true iteration the true thing that makes a company possible is that spark of connection
1:27:12between the Visionary and people like the first employees you pitch the idea to the the co-founder who was like that's stupid or like oh that's really intriguing and talk to Founders who have successful companies it's very rarely the first idea they ever had well how did they know the first idea was bad how do they know the second idea like there's this special thing that happens in that connection that back and forth no matter how linear you want to tell the story that's uh that's how it goes and to me when you make this circle it actually like raised the question for me
1:27:41I was like what's in the center of the circle this is like a this is actually a border of a space what is being circumscribed and the and the obvious answer to me just came to me one day with this is a being these things are alive they are super organisms they're as alive as you or me so we don't own them or control them we birth them and then we become like their organs their cells and that's very spooky to think about because then you're like well who's in charge of this thing but it's
1:28:11really important to have this realization in order to answer your question because think about if like you know if you're a doctor and you come to a doctor like I I need you to I need to diagnose me I have a problem and they're like and they're like a a bone specialist and you're like what is a human being ask a bone specialist like it's a skeleton with the stuff wrapped around it but like if the skeleton is broken it can't function and so they go I have a problem is it a broken leg if it's a broken leg I can help you but if you're like no my pancreas is out of whack they're like that's not you got to
1:28:40go find someone else's department so you go someone who studies the endocrine system what is a human being it is this incredible message passing network of these endocrine messages going back and forth hormones and all this amazing stuff if that system is out of whack the thing dies and like I go to a neurologist think about all the people who study the human body each one of whom has an idea of what it is and none of whom are wrong and yet all of whom are wrong because of course none of those people can tell you how to be a good person that's a much more holistic thing
1:29:09and so the same thing is true for organizations at every level of its beingness there are these systems they operate the same way there are these questions of the true north and the values and the iteration and the feedback and that's that's real and so when you like it can be very daunting at first to feel like that's just like it's like a dizzying fractal that's going all all the time but I think it's actually very liberating once you realize it because then instead of asking the question where are all the places I can apply this idea which is actually an
1:29:37infinite set and very overwhelming you can ask yourself like what am I doing right now what like there's a reason I'm doing what I'm doing is a good reason do I actually like and whatever that is like whatever you are in the present moment you can start to apply this thing to the thing you're people do not people don't ask that of themselves enough no not nearly enough most people are completely unconscious on total autopilot doing what they're doing and I do not know why and all it takes is a second to just wake up and be
1:30:05like wait what am I doing is it what I want to be doing so like like the first thing I think a very common misconception with these ideas is you're like oh I'm doing this thing therefore I'll use these ideas to make the thing I'm doing better but I would say the more better question ask is do I is this what I want to be doing and then if it's not what you want to be doing you can use these ideas to get you from here there like okay what do you want to be doing and now let's figure out how to make that a reality it's a bit confronting it's a bit scary like I've been doing this for years so I assume it must be what I want
1:30:34to be doing yeah well economists would say that's your revealed preference and it actually is what you want to be doing what you think you want to be doing is backwards and now we're into spiritual topics that maybe are too far fied for this i g to say that that Economist view that an entrepreneur is merely the vector that happens to have the the customer requirement has been channeled through reminds me of this thing and drive to survive the Netflix
1:31:01F1 Series where like the chief engineer is is talking is saying to the camera like all the reasons that they're doing so well at Mercedes or whatever and it's like yeah you know our Aros systems this and you and Louis Hamilton's looking over from the side EV it butts in he's like excuse me you do realize that I'm more than just ballast in the car right I I operate the vehicle the GU like yes also the driver yes Tech technically speaking that is true
1:31:29yeah something jonno was talking about during the week with me which might be worth discussing we should wrap up soon I don't want to take up all your day Eric is like just Jonah was pointing out to me uh in the nicest possible way what a strange company we are um we um you know when people say like oh this money you've raised you know what are your plans for scaling headcounts like well we're going to hire 10 to 12 people and
1:31:58then we're going to stop hiring people you know uh how are you managed what do your management structure look like we don't have any managers uh cool you know which you know sort of paperwork did you pick for your legals and stuff like oh we created customer for letters you know what are your policies we don't have policies like we're weird at every level um and obviously that's very intentional um um but like yeah I guess I was kind of interested in your thoughts there and
1:32:27also like I guess maybe it's not that unusual for really early starts with one or two people to think they're going to do everything differently but never seems to happen in practice and in practice the fact that we actually do things differently seems to seems to be different um most people assumed we would have capitulated by now yeah that is the the typical pattern yeah so so I don't think the the the um label of different is
1:32:55useful because most things are totally incoherent so being different from them might be good or might be bad I don't know like which things some are people assume that the status quo isn't incoherent like it's very natural assumption I'm saying it's an open secr among the people who study every aspect of the status quo most of them spend most of their time being like why are things this way this makes absolutely no sense and that's have to say that there's not a lot of good stuff also but in so incoherence doesn't mean like it's a collapse but it has has these problems
1:33:23so to me I think what we're doing is completely coherent that's why I like it and it's not even a a framework I would necessarily recommend anybody else it is a completely bespoke derivative from first principles of what we're trying to accomplish so like take the headcount thing that's probably the most controversial thing that we do um when I tell people that because we want to be do research and development we want to hire people who are worldclass at research and development they have to be
1:33:52able to do both things better than basically anybody else on the planet and people look at me and they're like well how many such people can there possibly be on this Earth who could meet your standard of hiring and I'm like we think maybe 10 or 12 and they're like 10 or 12 million like no I don't need 10 million where Happ 10 or 12 it's coherent because like we have set this Target and then we can accomplish it and so many things people like you can't get away with doing that I'm like well what if we only had 12 people like oh sure in fact Jeremy you remember this we were talking to to a
1:34:21vendor who was helping us with something and they were really having a hard hard time understanding it and I and I finally had to be like look how about we agree that we'll do it this way for the first 12 people do you think you could live with that and they're like oh for 12 people who cares and they just like they couldn't process that they we were just going to do that and then St they just assume that we would have to make this work for 100 people and it wouldn't work for 100 people you can't do that and that could this company exist if we didn't have ai to make the people super humanly productive I don't know would it work if our if our goal is to build AGI and we
1:34:50had to B like there's so many elements of the strategy that like locked together like um like a puzzle where every piece fits with the others that it makes sense for us um and I wish every company had that level of internal coherence like basically to me the two dimensions of a company are like is it coherent and does it have integrity so like is coherence is just do the people inside does it does a plan make sense does it all fit together are people aligned to some common goal Integrity
1:35:17Earle early Google did right like this is something that I find like I admired them because they did everything differently to everyone and you know they recruited differently using like you know billboards with coding puzzles they did they made money differently by not using banners but instead having small auction-based ads they did their IPO differently based on economic first principles and at every point I heard everybody saying like no they can't do
1:35:46that that's not how you do that thing yeah they used you know they create a distributed compute differently and what's interesting is as a result of that process what nearly everybody been has taken away from it particularly at startup says oh we should do those specific things the same way as Google we should hire like Google we should do distributed compute like Google but they didn't take away the message of what you're saying which is yeah we should go back to First principles and find out
1:36:15what crazy things nobody else does we should do yeah you know yeah I mean for those know the phrase cargo culting that that we have a term for that behavior um which we don't need to get into the historical origins of that or anything like that but the um what I think is is interesting is um this is controversial even among people who agree with everything we've said so far still like if you if you listen to like I was founders of stripe for example I said like we're only going to innovate on a few things and everything
1:36:45else we're just going to copy paste we don't want to take the risk of doing things differently outside the things that we really know about and I just think that's too bad like I I think that is it's understandable and they've had a lot of success with that so who am I creating a new Global Financial system is difficult so that yeah so you want to take on extra difficulty and again it's all about whether the thing in question actually matters for the purpose of the corporation I'll say one more thing about purpose because this is I know we have a lot of product Builders here and a very we're going through an ESG
1:37:15backlash this whole anti-woke thing it's like the idea of stakeholder capitalism was like hot for a while now it's bad and people you know these fads come and go a very common misconception about purpose is that it's always about some kind of high concept social thing like we we have this very high concept Mission around you know um promoting the public welfare using AI it's like very high conent I was talking to a Founder once and he was like look I don't care about this stuff I'm not woke you know whatever I just I just want to build a
1:37:43company that that makes great products and that is the great like for the engineers that work there is the greatest company they've ever worked at just like because we're so focused on building awesome products and I was like oh interesting is it your experience that that just naturally happens in most organizations that they kind of settle into that culture and he's like oh no every place I've ever worked was horrible and he's just like explodes at me even the companies he himself founded all got dragged down by this gravitational force I'm like okay so but but earlier but he's like but anyway yes
1:38:12that's all true but I listen I just want to make a lot of money I like that's not true I can prove it to you right now that you don't he's like I just GNA do whatever it takes to make as much money as possible I was like that's not true you are only willing to do the things that make money that also allow you to build great products he's like obviously that's what I just said it's like no that's not what you just said what you said was you're going to create an alignment between your business model and your purpose and when that model is turning in alignment with that purpose then you will do whatever it takes to
1:38:42make money that way but any other form of making money you are not willing to do he's like what other forms of money money are there it's like well for example what if someone said let's make the product shitty on purpose cuz because we can make an extra I would never it's like well then you do believe in corporate purpose actually and it call it woke if you want to I don't care call it what you want but it's actually very important we live in a world now where people have like forgotten why organizations exist and so many organizations have descended into
1:39:11bureaucracy or sociopathy because we have a confusion about this root cause like what is the actual reason we make an and maintain these superorganisms what are they for what's their moral compass what's their true north so yeah I think this is you know we started out just being like Oh it's a better way to make AI you know programs a little faster but the implications are much broader I think that's like this is it's it's been funny Eric because in the end our business model is stupidly simple but seems to confus so many people so
1:39:40much which is like people like what's your business model well we want to build products that people like enough that they purchase them for more than it costs us to make them and like so is it like kind of a you know data oil offshoot reselling thing or S of a vir it's like no no we want to build products that are good they're valuable enough to people that they pay us more than they cost us to make like you remember that remember that investor who
1:40:08was like is it B2B sass then like no no it's not anything it's just it's just making products that people buy that's and that's like socially that's like socially beneficial it's like okay it's like I thought you had some kind of like charity thing or something like no no we want to like how do you know if this thing is providing a societal benefit well people are paying us more than it cost us to make it to use it so it's it's analytics are so simple
1:40:38it's actually very easy thing to measure yeah J did you have anything that we didn't get around to before we I think that's I think that's good it's been a been a really fun Deep dive how about you Eric anything you wanted to WRA to say thanks thanks for having me thanks thank making me your non special guest I I really appreciate it it it feels like I I don't know like this for the benefit of the students I I feel like I've just been able to like sit in on the coolest couple of lessons here so thank you so
1:41:06much my pleasure I hope we got everyone's questions answered and if we didn't we'll figure out through them and I think pretty much pretty much everything okay good yeah feel free to put more questions in the Discord and we'll we'll figure out how to get them all answered if people have them looking forward to all the cool stuff people build