Nice to meet you too dude, it's gon na be fun. Checking out your channel this morning and going through stuff dude. I watch him for the guitar. I don't know why anybody else does welcome gandalf aka coach trap.

That's classic! I got my special shirt on for you guys today. Let's go guys, oh you got merch, you do have you have merch now i just make it for myself. You know okay yeah. So before i have virgin ears, so i hope there aren't i'm just kidding what are you drinking iced, coffee, no hot coffee, hot coffee? I was writing code, so i can't i can't write code and not have coffee or some form of caffeine, so yeah i'm trying to wake up dude.

I went to a concert last night, i'm a little a little slow today, uh this band. We do this like public concert series in the summer, the city i live in and i forget the name of the band, but they were like a rock tribute band, so i thought you could forget. The name of the city was the country that good yeah, exactly almost oh, my god, so for those who are not familiar cold strap here is the community gandel for for palantir farzad is a good friend of mine. We did multiple streams together, a former tesla, a very undervalued undersubscribed, channel experiences in business, intelligence, software, enterprise software so and the volunteer skeptic though no hater.

I would just say that, so i would call it healthy skeptic. So both these gentlemen channels are listed in the description for the stream and in the pinned comments. So the first thing i urge you to do is go right ahead and subscribe to their channels. I guarantee you, you would not forget.

If sorry, you would not regret that also, you would not forget that probably not used to like promoting channels, because i never ask people to subscribe to mine so whenever i do it for other people, it's appreciated, but i do mean it. Those are two channels that are criminally undescribed and hopefully, today we're going to change that a little bit now uh to kick things off, i'm going to tell you guys how this is going to go uh. Basically, i think this is better, so i hope to be silent most of the stream, because i hope you guys will do most of the speaking i'm just here to moderate. Take some questions from the chat.

Basically keep this going. So the premise for this is we want to talk about palantir, a question that doesn't get enough attention on the internet, which is what exactly is valentine doing and what are their competitive edge compared to comparable companies, competitors potential competitors, in-house id departments, whatever that may be, And the premise for this is basically we have farzad here who has the technical know-how, the experience and, obviously the brain power to, i would say, objectively uh judge whatever is put in front of him. On the other hand, coach trap who's like the technical king of the community. I would say so um i would just start with that's.

I hope this is as much as i speak through the entire thing, because i'm just here to moderate, so i'm just going to kick it off with the first of all. Each of you, gentlemen, please, like take 30 seconds to introduce yourself, so we get some subscribers on your channel come on. That's like the first thing we want to do before. We start the the actual conversation, the concept we could start with you and then for that you go.

Second, is the the challenger like in boxing you know so: yeah, i'm uh, i'm a vp of technology for a large edtech company and i've been working in the space for about 20 plus years. Over the last five to six years, i've been working on a lot of ai ml initiatives and um miserably failing a lot of other people trying to derive business value from those things and when i found out about palantir and kind of their platform, pretty much fell In love with it, i was like, oh my god. This is the thing we've been looking for and, oddly enough, when i created my youtube channel, i created it to advanced software apprenticeships and somehow i just wound up talking about volunteer like because it just i just got super into it. I wanted somewhere to talk about, and i had the channel as well whatever just start talking about up the platform here and i got hooked up with them through my work um and ended up forming a developer relationship with them.

And now i have access to like a free, foundry stack where i can build software and um. You know hoping to maybe launch startups on there one day, but right now my mission is just to try and get as many other engineers to understand the platform and hopefully open it up to as many as possible, because i feel like that'll, really help unlock a Lot of value for them and for for the overall software community, so yeah, that's what i do. Thank you, kendall yeah. I love mr discount patrick by david.

Your turn love it. This is great. I can already tell this is going to be super fun, so i'm super looking forward to this um farzad mizbahi. I've had bi background for about 10 years, so i worked for tesla for about four years building out, essentially the the bi kpi reporting platform for service distribution and operations for the last four years before that i was a director of bi and pricing over at the Uh, i guess right now, still probably the largest uh pet food distributor in the states.

Philips performance supplies there. We built out a bi platform, essentially um from scratch, using existing tools at the time we used burst. I'm not sure if code is familiar with that, but that was one of the softwares that we used to build that out. A lot of my experiences in sql and tableau.

But i have been exposed to different bi tools, and so one of the things that - and i'm also you know - i invest in stocks and stuff like that, and one of the things that really drew me to palantir was people like tom and others were extremely extremely High on the company - and one of the things for me is that i always try to really understand the product first before i i really say, invest a chunk of my of my net worth into something and the biggest problem i was having was truly understanding. What palantir's value proposition was especially versus other tools, and so i'm super forward. Looking looking forward to this conversation, because i think it's going to help me really understand exactly what paluntir's value proposition is from a product perspective and i'm super excited to uh dig deep with uh coach, strap and tom. So thank you for having me man and thank you for the really kind introduction bro you're over there like promoting us like crazy, come on so i mean that's part of the part of the idea.

Why i think it's important to do this. I think it's a fair trade-off. You come here, you give some of your own time, which obviously is valuable to educate the community. The you know the the trade-off here is that we get you a little bit of clout.

I mean both of you have what i feel are under appreciated channels and i'm sure will grow immensely, but let's push it a little bit further. I know you both are not classical youtubers as in like you, don't really. This is not what you do. It's kind of a side thing for you all, but uh i mean it's nice to have a bigger channel.

You both will have a bigger platform and you know get your ideas across. Why not hey before we start uh, you want to give a shout-out to the adam bergman show says david divine russian is kind of like let's go, let's go everyone smash the like button volunteer to the moon. Thank you tom and coach and forza. The mid hair is beautiful and i'm jealous that's uh.

I cannot argue with the last part. Exactly everybody knows that uh okay, so we have 500 people in the chat right now, which is uh just enough to get started. So the first question i want to you know what: why am i talking first, that we had an amazing stream couple of hours, music, pile interior tesla? What not, i think you were left with a lot of questions. So if you, if you could ask like one specific question, uh from a guy who like deep into politi, what would you ask him yeah for sure? So so my biggest question would be what what is palantir doing today, that the competitors can't copy and say like six to 12 months.

Like that's really, i think that's a good way to start the conversation, because then we'll be able to sort of deep dive into okay, exactly what what is palantir doing from a mechanics perspective. I really want to understand if it's a a true mode or is it just a clever way of utilizing ui to streamlining, workflow et cetera, et cetera, so maybe we'll start there and then we'll take the conversation. What do you think? Let me try and give you kind of a holistic view that it's volunteers, like foundry's like a layer, cake right. It has like three huge layers, which is like data engineering data modeling, and you got ai and application development right and within each of those there's.

Dozens of products right so like foundry has over has hundreds of microservices that make up the platform, along with dozens and dozens and dozens of individual modules that are just full product. Suites that you would use right. So imagine like one of those suites is a full data engineering platform with version control systems with code pipelines built in that auto, deploy everything you need once you write the code, so it's like kind of like something you would spend a lot of time building. If you were ahead of technology at like facebook or something you would build this platform for your data engineers, but in foundry, it's out of box right and but that would represent several several years of custom development work where you just can't go, buy a platform like That now there are a lot of companies trying to offer that right, like databricks, is trying to build that and offer it and snowflake is sort of trying to build out and offer it.

But there's still a lot of integration. Work going in that's just one layer of foundry like the next layer up where you've got data modeling like that, is all what they, what they call their ontology system right. The ontology is what enables what we call domain-driven design in software, so it's like. We can actually build software systems that act on representations of real-world entities, and that includes their behaviors and they do that through a powerful set of custom functions, you can write to represent behavior, but they also include an event system, so you can respond to events and That ontology can implement those behaviors as if it were a real world thing right.

You want to auto rebalance your supply chain. You want to do some crazy in response to an event that comes in. You want to run a simulation. All that's enabled in that.

Like sort of data modeling layer of foundry and then above that, there's the ai and apps layer right and that's where things get real interesting because what's happening, let me kind of take us. Let me try and tell you about what's happening in sas right now, like one we're drowning in it right so like about 70 of an organization software now is sas, and so like at my company, that represents over 200 products. There are sas products to manage sas products. That's how up this is getting right.

So so what happens is is the industry now is talking about this idea of data apps, and the idea is we're going to bring the apps to the data warehouse we're not going to have sas applications that are silos where this gui lives over here and all My data's siloed over here we're actually going to bring the apps into the data warehouse and build everything on top of that, that's going to be like the new cloud. Basically, when that happens, when that revolution takes place, takes place, boundary was way ahead of that game. Like they knew 10 years ago that that's what they needed, and so they built a suite of low code tools, including like quiver and contour workshop workshop. Being one of the most impressive.

Where you can they built skywise using workshop right so like they built an entire multi-million dollar software application using their own tools? That's how powerful they are right so like, and they knew that the data warehouse would be the application platform of the future right, and so they have those low denocode tools to let you build full-fledged apps, where you don't need a software engineer anymore. You just need to understand your data and you need to be able to have the same type of skills. I would say, like a traditional bi user would have maybe like a tableau user or some other user or some other person using their low code system right. So it's not no code, but it is definitely low code, although you don't need to, they do have a lot of built-in functionality.

You know um, so so yeah, okay, so that's kind of like the three layers and how they're differentiated but to like. If you're saying what could the competitors build in six to 12 months, i think the better question is it would take them 10 years. You know like you, want to get caught up to foundry you're, looking at a 10-year time horizon for sure, like in terms of all of the apps that are available and all of the modules, the the thinking of knowing that the data warehouse would be the future. Aws, that was the thing that kind of set them apart and that's what they spent so long building you know got it got it, so it's so it and so a sort of a layman's term of summarizing that it's essentially a place where every all the data You could possibly need for anything is stored and you can use the different layers to build out whatever tools you need to execute on the business.

Is that a fair way of summarizing yeah yeah? Imagine imagine you had a it's like a big data os right. It's the os of the future for you to build all the software. Your organization needs where, in the past, we had microsoft office and we had google right so like foundry even has like a google docs thing in it, which is crazy right. Why would they do that? Well, because now that your data is being built on foundry apps, all the data is now accessible too.

They even have a chat application right so like that eighty percent of data, that of the organization - that's untapped, because it's largely unstructured foundry has an app for that meaning all of your data is now accessible right. So i think that that was that's the other thing that differentiates it is like it's truly like an operating system. You know yeah, that's what it sounds like yeah, yeah, okay got it that's super helpful and so when it comes to actually so i'm gon na use sql as an example, because that's what i'm most familiar with right. So when it comes to like actually um uh, creating the data sets uh creating the tables, the schemas, you know making everything, so they can work with it.

The joins the views, whatever you need and then the graphical layer, that's on top of. Let's say i use a tableau or even even a microsoft, product power bi, whatever you want to call it right. What what is that experience like in palantir, and would you say it's any different, any easier or any more complex? I would love to hear more totally differentiated from anything else, you're going to be using unless you build a big data like a data engineering platform using open source tech for your company, but there's there's two primary ways: you model data in foundry to get the data Into the format that the bi users are going to use it, which is the ontology layer so like everything, goes from the ontology layer when you're talking about bi users or app users, but the first one is like there's a set of there's a complete online software. Ide environment that has spark embedded in it and you can do pi spark programming directly in there and you can use both sql and you can also use spark directly right, and so it has version control built in.

So you get a master branch for all. Your pi, you create, what's called the code repository. That's a complete pipeline of all the transform functions that go from the raw data set all the way out to the ontology layer you have branching built in, so you can create branches. You can do a pool request review in there of the code changes you make.

It has a complete ci cd system that will deploy all that code. You have to you, don't write any infrastructures code. The spark system is fully managed for you. You can control all the settings of the spark cluster, so you have like the complete capabilities of spark at your disposal to build just about any damn thing you want right.

So it's like you, could the it's un unlimited what you could do in terms of like data transformations um, the other way you can do it is visually, so they have a set of visual tools that allow you to build data pipelines. I forget the name of the tool in there um for but there, but there's this tool in there. That is a visual user tool that lets you go from a raw table to something to the way you want to visualize it right. I don't use that tool personally because, like i'm, writing uh code and spark, but that's the other method, and it also is version - and this is the cool thing about foundry - is they took um a change management like platform like git and they created a whole bunch Of custom stuff, on top of it, but everything you do in foundry is versioned right, which is pretty awesome so like including the ontology like a lot of people, ask like like what, if you're, making all these changes.

There's all these downstream apps like how does the system keep up with the changes like there's a whole change management system in there as well and conversion the ontology? All your pipelines are versioned like it's pretty incredible platform when you look at it from that standpoint, but there's the so. The the capabilities are far beyond what you would have say in tableau, where you're primarily using like sql right. So you have. You have a lot of in and the fact that it has like a custom, git implementation built on it.

So you can do code review and you can do release management and you don't have to write infrastructure as code. It's a huge accelerant right because now, when i'm sitting down to like i can get straight to the data, i can get straight to the modeling. There's none of all that integration work that takes like two years to find out that your engineers didn't know what they were doing like i can't get to like one month. I've got the data, it's all done and we can decide right.

There did we screw up or not in the data model, and it's going to actually derive some value for us. I don't have to sit there like fumbling around like three blind mice. You know trying to yeah trying to figure out. If did the team screw up or not, you know got it so from like a traditional like technical perspective of somebody that will be creating um these sort of tools for the business? It's it sounds like it's uh, it's easier to work on it's much more powerful from what it sounds like.

What i'm also curious to learn more about is, for so my biggest struggle in the bi world, and i don't know if you faced this before, and this i'm really curious to see. If palantir is something that that solves for this as well um - and you tell me through experiment - because i'm super curious to hear about this yeah uh, especially in my world - you have the operations language and then you have so operations speak. You know, we all speak. English right, but the operations team or the people that absorb the data speak one language and then the technical folks, the people that generate the the the kpis, the visualizations, the tools.

Whatever you want to call. They speak their own language right and usually what ends up happening is when both parties talk right. One party says i need this and then the the technical asked the people they're like. Oh, this is what you want and then the people are like.

No. This is not what we want. We asked for this, and you have this sort of like crazy mismatch, and i feel like that causes so much friction on organizations does palantir in any way help bridge that gap, because one of the things you mentioned was that um any typical sort of like a Bi analyst or a bi uh person that would typically do visualization. Some is somebody that could potentially be uh, create more values and volunteer.

Is that is that the case with this tool? Kind of walk me through that? Because that's one thing that i'm really curious to learn more about yeah, i mean if that question makes sense yeah it makes total sense, because i live that every day you know it's like the tension between the people who actually need to make a business decision. The analysts and then the technical people responsible for owning the data and getting into a format that they can securely access it. That's a complete cluster at most companies, yeah and so the way foundry solves. That problem is that it's one, it's a holistic platform for every actor in your org, so your analysts are absolutely gon na be inside foundry using it.

So it's all that data is governed under one security model, the other is they have this program called preparation and in preparation you can do your own transforms of the data you have access to, whether that be like intermediary data raw data, clean data, whatever you want To look at and you can prep the data in the format you you want it in so like part of the premise of foundry, is that we are going to remove the engineering bottleneck right, and so we give you tools, you the analyst that allow you to, Like bypass that engineering bottleneck problem further like the barrier to entry on spark programming, is so low. You don't need to know jack about spark like all you need to do is know, pi spark. If you can write sql, you can program in pi spark. You don't need to sit around and wait for your data engineers and their esoteric platform.

They created that has like 50 000 dependencies that now you need an ide to like install them all resolve all the conflicts just to write. One transform right. You can literally log in through a web browser, write the transform in your web browser issue a pull request, and now you've got your data in the format you needed it so that bringing that barrier to entry down to the point where it's equivalent to a bi Tool you use like tableau to write a sql like join or something that's going to unlock people like you and it's all governed under one security model. That's the power of like first power, first party integration right so they're they're gon na, and i believe that that is probably one of the biggest value props is in order to move at the speed of the data stream.

The org has to remove the engineering bottleneck because that one you touched on that's one of like a hundred. You know it's like yeah, there's there's like yeah, so like that's the other thing that i'm psyched on the platform and really like trying to evangelize to the engineers out there. So could you say that it could become a platform where so i'm trying to think i'm trying to i'm trying to like run this down like say three to five years from now? Right, so is this something that we could say could potentially become a platform where you even remove the analyst layer and allow just decision makers to tap a platform to make their decisions without having to say have anyone develop anything? Is this the sort of like? That's? That's the goal right yeah, it's like yeah, it's uh, we're in the early days. The the reason being is like there aren't enough models of data that exists right now to make that totally possible right, like hyperauto, would be the closest thing i think they've gotten to where you could say you don't need an analyst anymore.

You know, because what it is is hyperauto you, you kind of explored in one of your videos. They have this thing: called software defined data integrations and there isn't a lot of documentation out there about like how does that system work, but they created artificial intelligence and some uh heuristics to look at data inside known systems, and one of the questions you had is Like how do they do that right and interpret what the data actually is? Well, like one of the ways you do, that is with a heuristic like a regex right, you write a regular expression. You can tell that that data is a time you know, or that data is an address or that data is, you know a person's name, but you can also build models to do that and what they did is because they were doing this stuff for so long. Dealing with these systems like over and over and over and over again is in order to speed themselves up.

They came up with these tools to actually look at the metadata and the data itself using models, ai and basically interpret what you're actually modeling. What is it? Is it a car? Is it a car part? Is it what car part is it and then, from that what they do is they create your entire pipeline, like all of your data, transformations, your ontology layer and your app layer for you automatically like, and that happens within hours of the data being ingested, then you Can sit down and you're, starting from a standpoint that an actual person is responsible for inventory in a plant, can now start rebalancing the response to alerts that the system is telling them about. So it's like it's pretty crazy that they're getting there that fast, but that's only one like it's. I don't know how deep that integration goes number one.

I don't know how many types of data it can interpret, but as they scale and that's part of what i love about their go to market strategy. Is that the way you scale, that is by generating massive network effects with huge companies that do business with hundreds of other suppliers? You know hundreds of subs, basically right and on that level it makes a lot of sense. If i want to get to is a world in which i understand as much of the data and can model it effectively kind of in a universal way, using uh software-defined data integration going out and finding the biggest companies on earth and generating massive network effects and Understanding everyone's data is the way to do it so yeah, and so i think, they're gon na. I think we will get there, but i think there will always be a job for an analyst, because no system is ever perfect and right.

You also need human, like machines are good at certain things. People are better at others, you know, so it's there's, there's a role in there and i don't think we're getting to general artificial intelligence anytime soon so, like, i think, there's a role in there for the analyst, because there's not every problem is going to be a Machine problem, you know, and also like in a lot of these scenarios, maybe you're using some kind of explainable ai to to come up with a set of signal data. But an analyst still needs to look at that and review how the ai came to that decision. Typically, especially in something where it might be critical like you want to make a purchase or you want to transfer money or whatever, it is, maybe you're taking internal action of some kind.

But it's high stakes you're, always going to want an analyst in there to kind of review what the what the system's coming up with and responding to that signal, data and foundry's made to do that for sure. Got it and guys. I want to jump in here with a quick question from the audience, since i really liked it again sitting here is the layman just enjoying this conversation, i feel like i'm watching mj and the larry bird playing hoops, so this is pretty cool um. This is way above my pay grade, but there was an interesting question in the chat.

So as there was a gentleman i forgot his name who asked a coach trap. Do you know of anybody who's, a user of foundry that is critical of the system and did not find it to be of value to him or her? And if that's the case, why that happen? Yeah i mean even at my org right like we evaluated foundry and the decision was to not use it. You know um. I think that a large part is due to the price tag number one: it's not cheap um.

The other is no one knows how this thing works right. It's like a giant black box. Looking at the platform, i could spend several years training on this platform. It's not something you learn overnight and so one of the biggest criticisms that people have is like dude.

It's like drinking from a fire hose man, you know, like i don't understand this thing and um and and the other is that, even when they're evaluating the features themselves, it not everything seems like it to a software engineer, you're telling them you're going to use an Online ide, for example, like most software engineers, will say. No, i want to work locally, but the local development experience isn't as good as the online experience and and that can lead to people feeling frustrated in a way. So maybe they need better integration into ides. For example, would be one thing that i think is important.

I think they need a command line. Interface like they don't have one right now, like that's another thing that the developers are critical of the other in machine learning is like they don't think it's performant enough to do real-time inference when you need a feature store, you know, but that that's theoretical and the Reason it's theoretical is like we didn't get a chance to scale it and test it like internally at our org, but the feeling from the engineers was like hey. This thing isn't going to work for for feature store side, but, like that's a super like what i told them too, is like that's an edge case. Man like most of the time.

We don't need a feature store to do real-time. Embedding of these features and to request inference request, that is one thing they beat them up on. The other is like it's not obvious when they present a lot of those ml and ai capabilities like what how they would work and like what libraries are all supported. Even though, like everything they need is there, so it's i just get a feeling from like a lot of the engineers that they see it as a really different way of sort of working.

The terminology palantir uses is really foreign to them. Um, it doesn't feel like it's aligned with open source tech, even though it is right so they're using a lot of open source tech under the hood same tools, you would typically use so yeah. I think it's just a very intimidating platform. I would say that the online ide is not epic, you know like it's not the best online editor.

I would much prefer like stack blitz. I, like i, would prefer to work locally, although the local integration is fairly good. So you can definitely do high performance, local workloads, but if they can improve their online ide with a command line interface and a lot more generators and make it a little more performant. In the autocomplete features, i think it would go a long way to like improve the experience for engineers and then the other side to it is like really cleaning up how they talk about the ml side and proving out that they can handle that low latency high Concurrency requirement for model serving, i think, would go a long way but um, i think that's more presentation than it is like substance.

You know yeah so yeah and the quick shout out to buck mcenterson. I'm presuming that's a real name. If it's not that's the one. That's a killer killer, name uh, he's saying thanks to tom, for hosting this and for zarin code for healthy debate still trying to decipher volunteer - and this is awesome.

Ps metallica rules could not agree more. Let's go metallica. Will you play something at the end, for that? Are you set up, for course, bro come on now shout out to them guitars here and not play them. Let's go.

Let's go. Let's go okay, so 710 people here listening. I believe you had a question before i uh interrupted you. So no people hate when i do go ahead: you're good bro, you're good, so i actually had a couple, but let me start with the with sort of the aip.

So the one thing you mentioned correct me if i'm wrong here. So the one thing you mentioned is that uh palantir is taking steps uh to try and essentially auto generate, and i'm gon na try to use layman terms here so auto, generate um, uh, kpis or or columns or some sort of data points for users to to Hit or exceptions that come up and say yo something's happening in your business. You need to jump on this by utilizing. Uh data sets from uh previous companies or previous applications to try and do that did i did.

I explain that correctly. Was that something that they're looking yeah with hyperauto they're? Doing that? And it's not it's not kpis it's. They generate a whole app like a skywise like app, that's meant for inventory management, and that includes the data pipelines and they do that directly from the erp systems. You're using so like, i believe, it's sap and a couple, others they go and they have this mod.

These models they built along with heuristics, to interpret the data, that's in those systems and then they'll completely auto generate your entire pipeline for the to the ontology layer and then the app that sits on the ontology layer. That does the supply chain management and all that kind of stuff so, and that includes alerting, there's, there's features that are built in there and then from there you can extend it. You can start building all your own features on top of it got it so that that's one of the things that's most attractive to me, because i think, when it ultimately comes to from the world that i lived in exception, monitoring was really how we drove the Business, you know we just had alerts that came up that said yo, this thing's broken go fix it and we try to be as proactive as humanly possible right. That's so, when i think about data, that's how i think about data.

It's like how does it help me uh, be as proactive as humanly possible when something is about to or is breaking right um. So so one of the things with the aips, though so one of my understanding is that palantir doesn't actually own the data. It's it's. The user that owns the data, so how would it be able to leverage the data sets um um to do that unless it actually knows what data you're using or is that just part of the deal that you strike? That says, you still own the data, but we have the right to kind of look through it and see what kind of data set you're using, so that we can generate these mls and sort of ai tools to try and predict these things for other businesses.

Walk me through that a little bit, because that that point is confusing to me yeah so like as far as in ip agreements they have with their customers. I have no idea like right. That would be it's totally speculative, but like every company, i've worked for. We struck ipa agreements on you know what we own.

What's our core tech, what are we going to be able to take with us? What are you getting you know? What's your your ip so like when you're doing these projects, i don't need like i can develop a base model. Take that base model. I don't need your data from that then on and continue to improve that base model down the road on additional data. So my guess is when they were building software-defined data integrations as a feature is that they came up with some kind of ip agreement for that as they were analyzing.

All these data sets yeah, but they absolutely don't own the data, it's a processing layer. So, like yeah, i always tell people like how many fisa requests do you think palantirs had that number is zero because they don't own the data dude. You know it's just a processing layer. I think a lot of people don't know that about volunteering like it's just a processing layer, yeah russ gerber, does not know that apparently bro he was very upset.

He was very upset bro. I agree with everything he said before. That or after yeah, you know what's hilarious is like people just conveniently overlook your phone, i'm like dude the companies that own your phone own, you, okay, that that's where the government goes when they want information on you, they go to google and they go to apple. They don't go to balance here so like come on yeah, so that that's that's helpful to understand and it makes sense right.

So it's it's they're, looking at the at the processing layer, they're, not really! Actually, looking at the data you can, what i'm saying is you can develop? You can develop our you know ai and models on just processing the data, provided you make that agreement with people. You know right right and of course some companies would want to to benefit if that were that was a capability at some point and i'm assuming that hey, you will allow you to take advantage of these feature sets if you agree right, i'm going to sit here. I'm going to have a similar agreement with the us government and there wouldn't be a foundry sure. You know.

No, that makes sense yeah. That makes a lot of sense. Would you say um so would you say palantir is more geared towards companies that have a tremendous amount of data and needs from a data perspective versus say a large majority of companies. So i'm trying to like figure out as i think, about palantir's long-term uh customer base right so far, so based on some of the explanations you've given.

Is that it's a it's a it's? It's a it's a tool that has a lot of capabilities. It sort of reinvents how traditional engineers would interface with a product like this um, but it sounds like the feature sets it has and the capabilities it have. It has are profoundly uh, great okay, but is this reserved for say your facebooks and your googles and your and your advisors like you're, really really large companies that are gon na, have a lot of data? Or is this something that is also being geared towards say, say a humongous majority of the businesses that exist today that have data, but they don't know how to use it like any any sort of small business right you can. You can pick and choose who whatever business you're talking about.

If these guys want to take advantage of the data they have to hire some some. You know third-party dude, that's in a different time zone that is not doing what they're asking right. So how is paluntir viewing this market, or is it just not geared towards that, because i think i think if we really think about long-term potential of bi tools or any sort of tools that leverage data? I think that is the market that they have to crack. So i would love to hear a little bit more about it.

Yeah i mean we all know that, like i mean we're, not talk, what's funny is like there's a book called palantir foundry by use case this guy kai wrote it it's dope. If you want to check it out, it's a good way to kind of learn more about the platform but like in there. The book opens with like a use case of a woodworker who makes like i forget what it was. Maybe it was like russian nesting dolls or something it was a small business, though maybe like 15 or 20 people, but explained how they were able to improve their business using foundry right so like.

But but i think the broader point is like bi and data analytics is valuable, no matter what business you're in i mean tom, probably uses bi and data analytics to understand his youtube audience. You know it's like you need a one-person shop and take advantage of data. I don't, i don't think the size of the org dictates that um so, but i think the broader question the one to unpack in there is like how is palantir seeing this problem. You know and like i, i think it's obvious from the go to market strategy.

They're not going after the smb space yeah like it's, it's straight up, full-on enterprise, you know yeah, that's it right. There kai's dope. He has a great relationship with um the people at powell reviewed four and a half stars there you go so i have that book. Uh there's a second thing: another one reviews bro, it's all code strapping, but i heard a lot about this book.

I heard the book is legit. Actually it's legit and he has another one out too um that talks about using it in the intelligence kind of space. Um, but so like i think, palantir is absolutely going after enterprise. You know they're not going to be going after the same customers as snowflake or salesforce.

Right now you know, but i think that speaks to maybe where they're, at with the tech you know and like it's just not ready to scale to you, know 25 000 customers they're looking at it as like. We want to get the global 2000. We want to get the enterprises where we can generate massive network effects and we'll scale our technology and invest in making or make investments to scale our technology alongside those initiatives as more and more people on board with this system. Now that's a really risky strategy.

If that is the i'm speaking for myself, this is an intuitive understanding of where i see it, but that is an incredibly risky strategy. No one, i know, has ever gone to market that way. Every single product i've been associated with from day zeros. Infinitely like we have a scaling model built in you, can sign up on the website.

Customers are onboarded into a multi-tenant environment automatically. There are no scaling issues so like like it's weird that they wouldn't be. You know walking and chewing gum at the same time, going after both markets. I think that that points to some inherent issues related to scalability of the product - i don't know - i don't know that speaking for myself, but that that's my feeling, but i don't know if it matters in the long run, if they're right, if they're right about how They how this product can generate network effects for their customers and their strategy works.

They win if they're wrong, they lose and they lose big time. You know and they've lost in other spaces they lost in the data warehouse space. Why are they losing the data warehouse space because they didn't go to market seven years before everyone else with a data warehouse product? You know so that that inability to read the market and get good product market fit could be a weakness for them yeah. It's something that i worry about a lot when i, when i see their strategy playing out, but at the same time very i love the product.

I can see how it's going to be a huge win for those orgs that are adopting it. I wish my org would have adopted, but, like i think that um i see a lot of a lot of potential upside in what they're doing it's just you got to be right, you know yeah. I think i think that, for me is like from like the perspective of an investor. If i were to put my and then maybe tom, you can sort of add your two cents in this too.

It's like um, like i have no doubt that the product is is really good like from the very beginning. I i knew the product has to be the change in lingo since their yeah, but to have somebody, no somebody who i'm sure that it's a good product yeah, because we have somebody who obviously has experiences, is somebody that that's it's able to to verbally, say in A technical matter like two technical people we're sitting there understanding why it's a valuable product, but i'm still the the the the one thing that i'm still skeptic about is um from from an investment perspective, say how? How are we sure that this is going to be a company? That's going to have a runway from here to say five years from now, if the scalability and its go to market strategy could potentially hamper it right because of because of how adamant they are. That's the risk man, that's why the stock's not out a hundred right now, it's like, in my opinion and by the way like just full disclaimer, do not invest. Based on what i'm saying, i do have a relationship with volunteer.

I do have access to a free, foundry stack. I have to give some disclaimers, but, like i don't want, i can't really speak to the stock, but um i've been calling this out prohibited from speaking yeah. I can't even speak but like that's the execution risk right. That's what i've been calling out yeah, you can talk about the strategy, the market don't say anything about the stock.

I don't want you to get into trouble because i know the the rules and regulations. Unfortunately, yeah you can talk about the strategy yeah and you can. The strategy is inherently risky because of what you're saying you know like, like that's an inherently risky strategy and - and i always say like what happens down market like with the snowflakes and the data bricks. That's going to have an effect on you, you know, because that's who gets certified right like if all the engineers are certified in snowflake and not foundry, that now we can now we're having hiring full discussions right so like yeah.

These are things where the competition gets. A vote, the market gets a vote and if, if they decide to break a different way because your platform wasn't accessible or affordable or whatever like these are all inherent risks and go to market strategy. That says: hey we'll worry about the smb space later, let's, let's focus on the enterprises, because that's where the network effects are going to be generated, like you know, that's risky, i mean i do think. I do think the fact that there's a product that exists out there, that is almost like, i think of so based on your explanations, i think of volunteer, like you said, like an os like a do everything thing that utilizes data to like just help you be As powerful as powerful as humanly possible, utilizing data like that's that's how i am sort of perceiving palantir and if they have a platform.

That's that's unique in this way and is able to make it as easy and as efficient as humanly possible for the users to really capitalize on those things, if they're really trying to utilize the entire thing and all and businesses exist that are looking to leverage that That's phenomenal, but, like my how many businesses actually exist out there that are that even know that they need to do that right that they even know every business knows they need to do that so, like well, who's been talking about digital transformation. Well, it would the whole industry's been talking about digital transformation for five years. Where is it it's nowhere? Because it's hard, you know so, like everyone wants to get there, i don't i don't see anyone saying we don't want to get there like everyone wants to get there. I think that, at the end of the day, like more people on youtube know who talenteer is than engineers, you know and like that's one major problem, you know is that like, if you're thinking about how do you approach digital transformation and you're, the cto working with The ceo trying to figure that out if palantir is not top of mind and none of your engineers have pallents you're top of mind you're not going to be part of those discussions.

So i think that one is just an educational kind of outreach thing: people they need conferences, they need certification, they need open access, they need a developer community like asap, you know to help ramp that thing, but at the same time you know you were talking about Premiums as well yeah freemium, i mean that's a requirement right like if i'm an engineer, and i want to actually try the platform, i'm not going to pay a million dollars to like you know, try the platform and you need to be able to evaluate something before, Like you're, going to make a multi like a multi-year seven-figure deal with a company, you know, like that's yeah, unless you're willing to jump off the cliff with them, and some and a lot of organizations have because they see the product. Is that valuable and like i see it when i'm you know when i, when i use foundry, i'm like dude this thing's just gon na take over the world one day like that, all it's all dependent on you've got to ramp that dev community people need to Know who you are, it can't just be youtube investors that know what palantir does so. I want to ask a question code, so my friend works in the industry and his take on volunteer was as such. What he told me is that look um the problem with the with the whole enterprise software category, slash industry in the us is essentially it's uh way more complicated, convoluted and political than the government, which is kind of moronic.

But it is that's, and it's true yeah. It's definitely more competitive than europe. It's not even a comparison, and it's a it's a it's a bloodbath of competitive uh hardships and what he told me is like look if we, when we think of a new um new competitor like this, like they have to compete, not with the comparables, because there Are none but they have to compete with the frankenstein of a solution we built over here, which cost like gazillions of dollars, and somebody said it's the best things in sliced bread. Now they have to go to the ceo and the board and explain that it's not and there's no way in hell, they're going to do that, even though the volunteers they're going to have to do it anyway.

Bro uh 85 of those initiatives fail right and they're. Failing right now, if anyone here has worked in this space, they'll tell you right now, like the vast vast vast majority of those platforms are failing or have failed and they're having those conversations anyway right. So it's like that part. I just set aside that.

That's just math and statistics. By definition, the vast majority of the internally built systems fail and the number one reason they fail is because distributed systems. Problems are hard, you know, and you can't build a system of 100 plus. You know, services and products all glued together with duct tape and chewing gum and expect it to work.

It's not going to work, so i think that at the end of the day that that part is a little less trouble worrisome. But the part, that's that kind of concerns me is like. I think that they you're right the politics matter, especially like when you're doing these, like multi-year seven-figure kind of contract deals, the politics really do matter and i think, on some level, um there's there's pr problems. You know like pr issues they're gon na have to work through um.

I don't know if you guys caught palmer, lucky's speech at the all in pod when he was talking about his company, that's in the defense space, but it's worth checking out, but i think that there needs to be more people in tech. Talking like that and work past these pr issues, because a lot of people aren't going to do business volunteer due to just misperceptions about what they do, their spy company they own. You know they're using everyone's data to track what they're doing like there's all this out. There and then it's not cool to work with people who work with the government in tech.

You know a lot of people just just by that. Don't want to do like if you told your org, oh we're adopting foundry or talent. Here you might have a certain contingent in your org. That's just going to get pissed, you know and they're going to have to work through that stuff.

So i think the politics deeply matter um and but i think, as more and more orgs, this becomes normalized and more and more orgs are getting a benefit from it and not failing at an 85 rate but succeeding at like a 90 rate. That's going to turn the tide pretty quickly, especially at the c-suite level. I think there will be politics, internal politics, especially related to like volunteers, a spy company, and do we want to be working with this company here, i'm going to quit because you're working with them yeah yeah, the the analogy i can draw just based on the research I've done is that the like i'm going to use tesla and palantir as an example. So so and thomas you - and i talked about this a little bit when you and i sat down so so.

The the users of palantir appear as evangelical about the product as tesla drivers and tesla owners right. But i think where tesla benefits is that they have a leader who's willing to go out there and really communicate the the mission and then there's users and people that actually understand when they, when they interface with the thing they understand. What it is in the case of palantir, because it's technical in nature, unless you have someone out there, that's like trying to dumb it down to the general user base or just to companies that are trying to somehow uh really leverage their data. If they don't really invest time in that, because it's technical by nature, like i feel like like, where do you go from there, you know and sort of i don't know if it's and you guys maybe helped me understand, is it a lack of? Is it too much trust in the product that people will eventually get it or is it just not an understanding or a lack of of sort of investment and saying that you know what we just don't think that's the right strategy, or is it that they, if They do that they're going to grow too fast, like i'm, i'm trying to get my head wrapped around that because, if you're, if you're a company, that's in the space - and you know you're technical - you have to invest time in explaining what you do and like you Should squash anything, this is the problem.

I'm working on i've been working on this problem for over a year now like trying to get them to. This is what a developer community does right. You create developer advocates of your platform. You need technical users who become evangelists who go around training people what they do, but in order to do that, you got to have free access to the platform.

The documentation's online now, which is great like i helped you, know, push that initiative forward and you know it's. I want to get to more places where every engineer can start working in the system and start to learn it but like until you're there.