Initial Event Feedback Made Easy

Initial event feedback

we are just days away from palantir's much anticipated demo day this is likely the most important day in palantir's life as a public company thus far because of the currently limited knowledge on how intricate palantirs platforms are and what they can truly accomplish today i want to provide you with a high level preview for what you can expect from this upcoming event this video will be broken into four main sections an introduction overview demonstration and conclusion let's get started introduction palantir's first inaugural live demo day will commence on january 26th at 4 30 pm eastern time i'll be live streaming the event on this channel through youtube for those interested palantir asks everyone to join them to see their platforms in action and hear about their product development roadmap the event will kick off with opening dialogue from shyam sankhar who has been the chief operating officer for pound tier for over 14 years then he'll hand it off to bianca and john for an explanation on the current and future applications of the quickly growing palantir foundry bianca is a product group lead and john's role is responsive dynamism next doug andy and martin will explain palantir gotham in detail and offer demos for better understanding they will also discuss the new titan release of gotham the next generation of palantir gotham with new capabilities built on 10 plus years of learnings overall titan's platform upgrade makes gotham more performant open and proactive so that the world's institutions can continue turning data into intelligence the coo says quote titan is a force multiplier doug is a global defense lead andy is a product designer and martin is a product manager and then there's the demonstration of apollo palutear's software as a service product explained and shown off by clark and ally clark and ally are software engineers behind the relatively newer platform finally perhaps we'll get closing thoughts from ceo alex carp outlining the historic day in which they've opened up to share the power and capabilities of their programs overview to give you an idea of how this demo day event will go down i'm going to cut together some clips from govcon 7 featuring a palantir 101 presentation i've pulled together just the clips that are meaningful to the discussion about what pound tier is focused around and after this overview we'll get to the product demonstration my name is saad abdali and i'm a forward deployed engineer here at palantir technologies so we know that humans are really good at chess we know that some computers are really really good at chess but we also know they're good at chess for different reasons they have different strengths and so the computer obviously has a huge tactical advantage a computational advantage a computer can calculate the value of tens of thousands of positions every single second but a human can take a shortcut through those computations sometimes a human has experience human has insight the human can feel out his opponent in a way that's very hard for a computer to do the combination of a human and a computer can actually defeat the greatest human players and can also defeat the greatest chess supercomputers let me take my single strongest human player and put at his disposal the single strongest chess computer make them into a team and they will dominate the tournament you might think that and you'd be wrong because it turned out that the team that won this tournament was not a particularly powerful player he's a modestly skilled player it was also not particularly powerful hardware or software it was commercial grade laptops running freely available open source chess tools and so what happened how did this combination defeat teams that had stronger players and stronger computers well it turned out there's sort of a third a hidden factor here something we didn't consider which is not just the individual strength of the machine and not just the individual strength of the human but the interface the quality of the interface between the two how easy is it for the human player to ask the questions that he need to ask is it possible for him to focus on the things that he's really good at the strategic oversight the guidance and to rely on the computer for those for those tactical components and so this actually aligns very closely with the way that we at palantir think about the problem of using computers using the power of modern computing to assist organizations with the analysis of huge amounts of data everything we do focuses on reducing that friction we call it the friction between the player or the analyst who wants to use a computer to perform a task and the data or the computation that he wants to do with that computer paypal was faced with the exact same problem but they took kind of a different perspective so they started at the same place they said we have a huge amount of decisions to make a huge number of decisions to make over a large amount of data and we have to do it quickly but they took the opposite perspective they said well obviously the answer is to increase the effectiveness of every human analyst to put the computing power at the service of the human who's still making the ultimate call who's still doing the analysis who's still asking the analytical questions and the focus is that everything we do on the computing side everything we do is all about empowering that analyst to explore his data faster to answer questions quickly to discover the anomalies he needs to discover and to make those decisions for the business ebay saw the writing on the wall they had an exit strategy from this battle which was to purchase paypal for a billion and a half dollars and after this happened the the team at paypal became notorious in silicon valley as very very prolific entrepreneurs and some of the members of that team went on to found another company you may have heard of called palantir because it turned out that the same exact problem that the fraud team faced that palantir is applicable across many domains think about the world that they lived in they lived in a world with a huge amount of data coming in at an increasing rate and they lived in a world where very quickly they had to make good decisions based on fast analysis of that data and this is applicable across a range of domains this is true in law enforcement it's true in the intelligence community it's true in defense it's true in public health if you're tracking the outbreak of a disease we live in a world with an increasing amount of data and increasing need to use that data to drive good decisions and that's what palantir is all about palantir is an infrastructure for analysis and i'm using the word infrastructure very very deliberately here because uh palantir is among other things not a visualization tool palantir is not a closed environment and finally palantir is not one database to rule them all so you've been hearing me talk a little bit about what is it you know what it is that palantir is doing under the hood and we actually divide that into four major areas of functionality which i'd like to go through with you in a little bit of detail but they kind of build on each other so we're going to start at the bottom here and work our way up those four areas in order are data integration search and discovery knowledge management and collaboration data integration is absolutely the foundation of everything else that palantir does and by data integration we simply mean the palantir takes all the forms of data that exist across your enterprise everything from the documents and spreadsheets that live on your analyst's desktops all the way up to your large enterprise grade databases and makes all of those accessible from within a single unified environment the second major area we call search and discovery so by search i mean that palantir provides a single point of access literally a single search box that allows you to reach out and touch all of the various data sources in your enterprise and we're not just talking about simple search so we're not only talking about searching for what you know we're also talking about tools to help you discover the things you don't know so by that i mean we enable advanced conceptual searches allowing you to search on the basis of the relationships between things the way a given network might look as opposed to merely the things themselves and you might want to know for example who's actually allowed to see this so we might need to separate our data into different compartments and so palantir is designed from the beginning with this idea of knowledge management in mind every piece of data that enters palantir is tracked with respect to how and when it entered the system who's allowed to see it how the information is changed over time and to what data sources we attribute this information and again all of this is key to doing good analysis knowing not just what you know but how and when you knew it how the information has evolved and in most analytical organizations we find that the most scarce resource is actually something else the scarcest resource is actually analysis it's what your analysts produce it's the application of that human judgment experience and insight against the raw information at your disposal that's the most valuable thing so from the beginning palantir was designed not just to accommodate the act of analysis but also to make it easy to share the results of that analysis across the enterprise the analysis that you produce on top of your data is actually as important or more important than the raw data itself and so this brings us to the idea of collaboration allowing individuals and groups of individuals to benefit from each other's work to work together to build this common picture of reality that analysis implies so finally a few notes to end on palantir is scalable palantir is secure palantir is low risk and finally we're talking about proven technology so this is technology that works technology that's ready now and we're very excited to to get started with you demonstration i hope that was insightful now i'm going to cut through the latter half featuring the best publicly available demonstration of pound tier software that i have yet seen a couple of things to keep in mind the software now shown is old and therefore the interface looks quite outdated especially when compared to palantir's updated design language today regardless keep that in mind as well as how palantir has no doubt made improvements to the core efficiency and features of its systems today all that said enjoy the following well articulated demo day in the life of a palantir analyst and in this case i'm going to be playing the role of a counter-terrorism analyst who's particularly focused on terrorism finance in north africa you'll see that i'm tracking some leads in egypt and i want to show you sort of from start to finish what my day might look like so starting from new information coming in from the field i run that information down and do some investigation and in the end i have some conclusions that i want to share with the rest of my team and with the rest of the agency and and you'll see what sort of that whole life cycle looks like and so i'm going to look at my new hit here and it looks to be a cable from our asset named ct blue who's operating at cairo station and it looks like he met an organization and several members of the organization the al-muja charity and he believes these guys are potentially involved in in the financing of a terrorism operation and particularly talking about attacking an iconic symbol in a u.s city sometime soon so what i want to do is to take these names this information that i've gotten run them down in palantir and see if there's anything else we can find that might corroborate this information or give us something else to go on you'll notice these are the three business cards that he captured and you see some blue highlights all that means is that another analyst has in my shop has already gotten to this document and he started making it more useful for the rest of us by doing what we call tagging it and that simply means making it clear that some of these pieces of text are referring to information we have in the system so this hyperlink actually will lead me to a record a dossier that we have on this guy mike vikry and i can add more information to that so i can say this phone number actually belongs to that guy i simply tell palantir that's a phone number and this is simply a way of of making sense of and making use of data that's inherently unstructured like a written report so what i'd like to do now is grab these three guys bring them over to my graph which is sort of the primary link analysis workspace in palantir and see what else i can find out about them how they're related to each other how they're related to other known entities you notice that one of them already has a a picture a mug shot so we definitely have information about this guy i'm going to double click on him and open him up and i want to remind you that what we're kind of looking at here is a combined view of what information we have about mike vickery from all of our different data sources so that might be raw reporting coming in from the field it might be some of our databases that we have in-house it might be external data things like payments and phone calls as you see over here so what you're looking at is kind of a summary view of this individual we can see several versions of his name a couple of different addresses his name is rendered in both latin and purzo arabic script so we can handle all kinds of of language data a couple of different phone numbers we have different kinds of attachments so you can attach here video feeds audio images and finally we have things that he's connected to so we know that mike vickery is connected to at least 23 phone calls and two different payments that we've we've brought into the system and once we get back to the graph which is a good place for analyzing networks and relationships we're going to take another look at that so i'm going to do that now what we could do is look at each of these three guys in that level of detail but what i'd like to do is to ask a somewhat broader question so instead of looking at just the the dossier profile of each of these three guys i'm going to ask a deeper question about how the three of them are connected not only to one another but also how they're embedded in the broader network of information that we have available in palantir right now and to do that we have a great tool called search around so i'm creating a new search and palantir is now asking me what kinds of matches what kinds of uh connectivity i want to explore so i can i can look at relationships i can say who are these guys related to and who else are those people related to i can see where there's matching information in the system so i can ask palantir is there anywhere else where we've seen this same street address is there anywhere else we've seen this email or this name show up and we can also see who they're connected to through events so who's on the other side of those payments if we know who's on the other side of those phone calls and who are those people connected to so palantir allows me to configure very precisely what kinds of questions i want to ask in this realm and having configured that one time i actually don't have to reconfigure it again so what i'm going to do is use a preset search that i've already that i've already created here and it's going to ask a few different questions it's asking what groups are these guys a part of and who else is a part of those groups it's asking what events have they taken part in and who else is connected to those events and it's doing all this out to four degrees of separation so it's not stopping at asking the question once but it's repeatedly asking the question in order to build me a broader network as a result and you can see that we built out quite a large network here so i'm going to finish and bring that into my investigation and start taking a look at what we got so the first thing that i like to do when i get a large result like this is i like palinder to tell me you know at a high level what am i looking at exactly so to do that i'm going to use a tool and palantir called the histogram and the histogram will basically give me a summary of whatever it is that i have on my screen here and so i'm looking at 40 different people on my graph and it also tells me some of the things they have in common so it shows me for example that four of these guys are actually living at the same address in berkeley california and here are the four that had highlighted for me same thing three of them living at the same address in vancouver three in toronto it's very easy to see see these groups and why they popped up you can do the same thing with other types of information so i can see who's living in san francisco berkeley daly city i can see the most common domain for an email address here is hotmail.com and i can see who's got a hotmail address who's got an aol address matches on last name nationality and so on so what's important to note here is not just that palantir can hold all these different types of information but that the the type of information you want to hold is actually flexible so in the context of counter-terrorism these are the types of things you'll likely be tracking you'll be looking at people with their nationalities and by graphical information and events like phone calls and payments they've been involved in but let's think of a totally different context something like cyber security well in that case you'd configure palantir so that the the objects you're looking at instead of people might be computers and servers and the events you'd be looking at might be communication traffic between those computers right in the healthcare context you might be looking at outbreaks of disease and patients rather than terrorists and and phone calls right so this this framework for looking at information is completely flexible and applicable to many different domains so now that we have a sense of what it is we're looking at i'd like to take kind of a closer look at what we got here the first thing i notice is that out of our three initial uh suspects only one of them really bore any fruit so the other two guys we don't seem to know very much about them yet and so for the purposes of of right now i'm going to get rid of them from my screen haven't deleted them just got them out of the way for the moment and i noticed that mike fikri one of our initial suspects is connected to this interesting cluster of individuals over here so when i take a closer look i see this is labeled a suspected bay area cell so this is a group of individuals that someone else on my team is tracking and that that person has labeled as a potential terrorist cell operating out of the bay area and mike fakery is connected to these guys through just one degree of separation so he's connected to someone who is connected to the group so that automatically raises my interest that that potentially starts to corroborate our suspicion that this guy that our asset met in cairo might be involved in an attack somewhere in the united states now we also saw that he's connected to this very large an interesting looking group of individuals over here so the first thing i want to see is how he's connected so our original individual mike vickery is connected to these guys through this sort of unknown entity named mf and i'm kind of suspicious because they have the same initials so let's let's see what else they might have in common i just open up the connection between these two guys and i can see what it is that they share so they both list california addresses they have the same initials as we already knew they're both iranian and nationality and they both share one of the same phone numbers exact phone number so to me right now this is not bulletproof it could mean a lot of different things but i'd like to test out the hypothesis that these guys are potentially the same person and so there's a very easy way for me to assert that in palantir i can simply take these two guys right click on them and say resolve i want to assert these are actually the same individual and palantir is going to merge those two guys for me we're retaining all the original data from both records retaining all the original information about sourcing where did each piece of information come from when did each piece of information enter the system we still got all of that but now we simply have this combined view that incorporates all the addresses all the phone numbers all the information from each independent record now a couple of things might happen as a result of this i might decide later on that you know we get new information and this original hypothesis was bad it doesn't make a lot of sense and at any time i or another analyst can actually undo what i just did so i can say unresolve and restore the original data the way it looked before but something else might happen i might say get more information that corroborates this hypothesis and i might become confident enough that i want to share this hypothesis with the rest of the team the rest of the organization and so what i can do is i can take whatever changes i've made to mike fikri and i can say i want to publish these changes so so far i've been working i've been making my changes in an environment that's private to me but now what i'm doing is saying that anyone who looks up mike fakery should see the result of this work he should see that mike vickery and mf are actually the same person he should see that combined record the way that i see it and we're done so let's move on we've got mike faker now directly connected to this large and interesting group of people and you see there's a whole lot of different kinds of connectivity going on here so among these this group you see some payments happening and some phone calls you see shared addresses and shared flights to people traveling together and it's kind of hard to get our head around what's happening here you know who's paying whom uh when are these things taking place and so there's two ways that i want to to to narrow down this information a little bit further the first thing that i want to do is take a look at directionality so if there's a payment going on somebody is paying someone else and i want to see that visually and we have a great tool and palantir for doing that called flows so what you're seeing now is a visualization showing you where money is originating and where it's being sent and the little red dots that you see are actually sized according to the size of the payments so we can see that there's some smaller stuff going on in the rest of the graph but really the main action is down here large payments going out from this individual to these three groups of people these three wings so that's very interesting and there's two things for you to understand about flows one is that flows is actually completely extensible so it's not just a tool for analyzing payments but it's a tool that allows you to visualize and understand any kind of directional or transactional information so in this investigation something else we might look at is let's say calls you might look at what your network looks like in terms of phone activity right and you get a different picture of how these guys are connected to each other um but there's a second way in which this is a reflection of the openness of our platform and that is that flows which looks and feels very much like it's an integrated piece of the the software is actually written as a plugin it's written as one of these third-party apps so it just goes to illustrate uh the degree to which the software is customizable and can be made to perform different kinds of tasks as needed all right so we understand the directionality of these payments now we see that the guy mike vickery is connected to seems to be sending money out to these three different groups we also want to take a look at time so we know that money was sent we know these guys traveled what was the order in which those things happened were all the payments at the same time were they right after one after another were they before the flights after the flights and so we have another tool called the timeline which will do i promise exactly what you think it will do show us all the data that we have on our screen in sequence so that we can understand what it is that's happening over time and so i'm just going to make this a little bit easier for us to understand by coloring the different events i'm going to take the flights and make them blue take the payments and make them green i'm a diehard verizon customer so i'm going to make my phone calls red and now we start to get a little bit clearer picture of what's happening so we see that the first payment you have a lot of phone activity and then the first payment takes place on the 20th of this month and you have a phone call quickly followed by a money transfer you see the same pattern repeat itself just one day later with this second payment and the third payment happens again just two days after that again accompanied by a phone call and if we look shortly afterwards what happens we see a lot of flights so these three groups of people who if you recall they're living in vancouver mexico city and toronto they receive large amounts of money from a guy who's connected to our suspect accompanied by phone calls and then the very next day all nine of these guys are getting on flights so where are they going i'm going to go back to my histogram here and find out what i can about the airports so i scroll down you see the three original cities vancouver toronto and mexico city and all nine of these guys are heading on a flight to chicago a u.s city so at this point i'm pretty alarmed i think i have something interesting new intelligence came in indicating that a group of people might be involved in plotting an attack on a u.s city or financing an attack on a u.s city we found that that person one of those three people is connected to someone who's making large payments of money to three different groups of individuals who are living outside the u.s and immediately afterwards all three groups are traveling to the same u.s city so this looks to me like it very well may be a cell activation pattern these guys are getting ready for some kind of operation and so what i can do is quickly share my analysis with members of my team with other agencies with local law enforcement and i can do that very quickly so i'll just take snapshots of a few of the key pieces of evidence here i'll take a snapshot of this timeline which shows you the pattern of activity i'll take a snapshot of my graph and having done that i'll head over to export what i've done into a format that's digestible in this case build a quick powerpoint presentation so palantir will do here is take the path that i took through my investigation so first i looked at this guy i built out a network i explored that network in a little more detail and of course i might want to change some of this right so i might want to explain what exactly this timeline indicates but essentially now i have something that's almost ready to go it's almost ready to be shared after i tweak it just a little bit and so what i just walked you through is sort of the the full life cycle of a day in the life of a palantir analyst starting with getting new information from the field doing a little bit of investigation and analysis on that on that information making some discoveries that might be important and then sharing that information with other people you'll notice what it is that i actually spent my time on so i spent most of my time talking with you about the actual analytical process the kinds of questions i wanted to ask about who these guys are and what we know about them what you didn't see me do was spend a lot of time searching through different databases for these names you didn't see me spending a lot of time fuzzing with different kinds of data formats or arcane queries and you also didn't see me spend a lot of time at the end reproducing all my work in a format that's digestible i didn't have to spend a lot of time creating a presentation or creating a report about what i did what you saw me do was essentially spend time doing the things that as an analyst i'm good at spending my time applying my domain expertise applying my intuition following the leads that i think are interesting and allowing the computer to take care of the things that it's good at searching for data retrieving large amounts of data uh converting data and transmitting it into different formats to make it digestible so all of this again comes down to this idea of reducing the friction between me as an analyst and the information that i'm working with allowing me to quickly answer questions to quickly do an investigation and to quickly share that data share that analysis across the enterprise i hope this has been a fairly useful demonstration a very cursory demonstration i promise you of what it is that palantir does there's a whole lot that you haven't seen yet conclusion i hope everything i pulled together in this high level preview was insightful for you and i think if you made it thus far that you're now well prepared for what is to come from palantir's first inaugural demo day technically i think it's going to be quite exciting and should outline palutear's ambitions as well as their abilities to achieve such lofty goals as being the most important software company in the world i'd really appreciate a like if you found this video useful and subscribe if you want to experience pltr's demo day live with the live chat active until next time