 Good afternoon and welcome to the best track there is because here we explore new with all the things and maybe things to sort of think about So with that, the topics. So Dr. Harding asked me to do this talk and the idea was to start out sort of as an umbrella So this talk is a bit of an umbrella as much as the rest of the talks will be more directed With that, you'll see some use cases of some of the stuff I'm currently working on And so far we haven't said the name of the company I work for, which is really good because I'm on vacation right now And so, but I do work for a company and probably some of you flew on airplanes, a couple of me, but a company But I live right down the street from here. I've raised seven kids in this neighborhood that we're in And welcome to Arizona. Probably no one's welcomed you yet here in Arizona So I've raised seven kids here and I live in Seattle. My wife lives here and we commute back and forth each month So this talk is sort of interesting because it reminds me of the state So Arizona has this big ditch called the Grand Canyon And what's interesting about it is that a couple things. One is so close And Arizona is shaped like this. We've got to the north, we've got these big mountains that goes down this valley that we're in here And then there's some mountains in the south like Swordness or this flat desert area But as you go north, you get up to Mt. Alexander, you get about 11,000 feet So it's pretty tall up there. It's about a thousand foot drop in the water between Grand Canyon and the Mexican border So welcome to Arizona. And this talk is about this gap that exists And think of the gap as being on the left and the right between data architecture and information architecture Now, we had some need to get across the canyon there And so it used to be that people would drop down and drop across the top of Hoover Dam Just this down and off to your right a bit And so they built this bridge. It's been, I don't know, three or four years now It seems like maybe it's a little longer. Time passes faster as you get older But this is an interesting story. It's so much as that you had someone sort of architect the view of the bridge might look like conceptually And then someone do the similar structural architecture And then another field into the geotechnical architecture of how that bridge bolts into the wall To make sure it doesn't fall down over time And so there's this bridging of architecture and engineering here Which is interesting and apparently a pretty stable story I think by anyone's argument you could imagine that's probably a pretty stable structure Although it's when you swipe the right across there So this talk is about bridging this gap, if you will, between information architecture and data architecture Many of us have been around the TOGAB trades for a long time Often we'll think of information architecture as a child of data architecture Can you manage such a thing? Maybe you can because that's what we've been saying for a long time But I think it's become just the other And in fact, when I was a kid I used to go to the library And some of you were over the age of 30, not very many of you used to But the idea is that you go into the library and you pull out the card catalog Does anybody remember those things? Wow, there's a few of you And so you pull those out and you look through these index cards to find stuff And the Dewey decimal system tells you how to find the row, rack, bin and shelf And where to find a product Well today, librarians need a little bit more than that They'd like to find out the number of dogs and all the books that are in the library That includes structured data and books Non-structured data and films and tapes and so on And how do you do that? How do you move from this world of structure and cards and library science To information science to link data that's across that fabric And how does IT support that? Well if you can imagine, that's a grand canyon And that's what we're here to sort of run through So to start with a little bit of vocabulary Since I've used these words, especially one and two quite a bit The notion of ontology, this listing and naming some pieces of information So the notion is that we classify data You see it according to some of the vocabulary that's been named beforehand And then we tag it And then the idea is that we'll call that ontology So for instance, I can go into a red piece of machine learning or machine code And what it might be is that it might listen to some, I don't know, speaking industry agriculture And it could be some tables, if you will, of agriculture, maybe crop harvest in Iowa And I can stream that data through my code And the idea is that I'm going to look for nouns So I can classify those nouns as having something to do with architecture or agriculture And when I find them, I'll tag them with a name, I'll be corn or tomatoes or whatever And so that's the notion of being able to build a vocabulary is directed And we call that an ontology I'll say further that most in the world that I personally work with these days I have handlers or bosses that would like us to have tamiocontologies These are managed ontologies that we try to make sure that everything is in the subset If another word shows up like bucket Bucket is probably not part of an agricultural ontology So it's a variance, it wouldn't be outside that set Taxonomy Well, my school, you probably learned about phylocordata and humans and backbones And non-backbones and bugs and all these sorts of things down to jellyfish And that was a nice story because it linked all that ontology into a bigger framework So taxonomy is sort of a bigger arrangement of things into categories The industrial and consortium, you heard that name a few times yesterday and today Is a subgroup of the OMG, our sister, our competitor, our sister organization I suppose, and the OMG has the industrial and ethnic consortium And one of the things that they do is they do test beds of particular kinds of projects That companies are interested in, I'm with this about 30 of them today And they also do vocabulary And the beauty of that is they have vocabulary with instances of things And that provides a taxonomy of vocabulary And those individual vocabulary of course are ontologies And now sort of when we mix and match and try to prioritize things And look at latency and cost and interoperability That's called the choreography of taxonomic objects So as an enterprise architect, on the left is what I think I do And you might laugh, but this is what I think I do I think enterprise architecture looks at global events of drivers First off, I should take a step back Enterprise architects are sort of like the family counselors of enterprise business That's what we do We listen to people's problems, we try to fix it and get it back on track And so, I don't really do this work, I fix somebody else's So companies look at global events of drivers And from that they try to detect some kind of strategy they can get In order to be profitable or to earn value for those particular problems And with their companies And they do that by employing their capabilities Capabilities being work that they do Investments with humans or machines and systems That they put together in order to do work And they apply information products to those capabilities Also called IP I know it's overloaded, I'm sorry But an IP in the information architecture world's information product And underneath those are data models And data models have essentially data in them And data for context is the value generated that goes up the stack Now, imagine you're enterprise with information and business architects All these things are on the left This is just a little cartoon here Bill, sales going great, but we're getting low on stuff Well, why don't you make another run to the landfill Bring back whatever you get your hands on And that's kind of like one IT, isn't it? You go to IT and there's lots of data But how do you contextly find stuff? You go to the library and you go to the cart catalog How do you find how many dogs or how many buckets Or in the library, what's a difficult thing to do And that's a story that we're going to look at So human systems are machines And why I have that on the left is you'll see in a few minutes But humans are interesting to anybody's problems But ultimately systems are machines If the same conditions come over and over again They're pretty perfect We used to say when I was learning to code That code is the only thing perfect Because given the same length, but you always get the same length You put in 2 plus 2, it does addition And probably is correct 2 plus 2 always equal 4, it's perfect Not unless there's a hardware race condition And blah, blah, blah, you might not get it But the idea is that humans and the variable And systems should be directly or maybe a little more accurate But in the world of information architecture Is this notion of being able to connect people To this content, to this data That's across the vast ecosystem And even multiple ecosystems And at the top you'll see some of the things that we Are increasingly doing in enterprise architecture Working with enterprise architecture Information architecture A clear, clear site in your organizations From IA to data architecture There's a link there for the University of Washington's library of science program Which is essentially no longer a library of science program It's an information architecture program They talk about data architecture They are trying to contain Data architecture and patterns So that's it, let's look at a few pictures With standards because if you're If you're at this conference You care about standards If you're at IA You care about open group standards And WC3 standards probably And so I'll claim that there's a dirts to value If you put these two organizations together Now I'd like to think that open group Has standards for everything But we just don't Apparently we've missed out on Anybody traveling from outside of the United States Two foot, oh my gosh, a lot We plunged, you have I did 60 countries in 10 years with HP I had seven different plugs Apparently the open group messed up You didn't notice out there You should not have one plug So before we get started Does anyone here, if you imagine all those Boxes or applications that are purple to the left And right, that all the problems in the world Be solved with API services and microservices Raise your hand if you think they'll also Solve the world Well pretty much that's What, for the 15 dirt So applications Imagine 1,000, 40,000 applications And someone draw that same map How silly would that be? So that's not the solution All the vendors are going to say It's all about API services But all of us know that's not the case So for the rest of this slide deck Last two slides We're going to cover a story That's sort of a journey The story is about industrial IOT Supporting factory products From the left, shipped to a supplier Using some kind of IP Some kind of semantic information products So you're in the left Perhaps it's China Indie or someplace They produce factory parts Manufacturing to bed And then it goes to some depot By a truck or a train Goes on a ship Now it's seen as an export By the company, by China for instance And then as big as the ones Is a Seattle And then seen as an import And then it goes on a train To Portland And then maybe it goes to another warehouse And then it goes over Intel What they even sell with some things So imagine that sort of a story And there's some comments there You're going to see this picture A little bit later from Ron This is kind of kicking that off And there's not a common vocabulary across there In fact, things will happen along the way That typically we've not known about in the past So I'm on a big ship So I went to sea for almost 20 years And I think I've seen a storm or two At that time I've walked a few containers In the Pacific Ocean Can you imagine that? We didn't stop to make it up either But if you could pick it up It would be a little bit waterlogged And I've been in some storms And seen a few waves And sometimes they're quite large And sometimes you can imagine That the ship heals over 35-40 degrees And it's a pretty fun ride Scraping back and forth And imagine your car is inside one of those containers Slapping around left and right Getting banged down Because the plywood was holding it To the floor Sort of getting ripped off And then we put that container on the dock And you can inspect it And there's all these debts on both sides of the car What are you going to do about it? We can turn it into insurance But you don't know what happened You have no notion But you know today we've got IoT in there And we can actually have five sensors On most containers Because that's an ISO standard And so we can tell When it reached when the clonometer rolled That information is passed It's now associated with the bill of lading But you can imagine How does that information get passed Back to the assembler And back to the factory, the senator Even with the shimmy come here to you Is receiver of that product So there's a vast vocabulary across there And it's not really a good way To be able to talk to the next three In order to do that I'm going to show you how that's done And it's in fact being done Before we do that We're going to look at a couple three standards I said a couple three That's not really where you're going to do it We're going to look at three standards And so those are the open data format The open message interface And the most important Open data element framework And these are all open group standards And foundational to information products Has anyone not heard of these three? Just curious Wow We've got to Make the rest of this talk to you, sir I'm glad you joined us And so we'll typically use ODF Inside of a thing And this is a biotea We'll use OMI, the message interface To send data back and forth And we use ODF to sort of put it all together So we can send messages to heterogeneous objects Or to move it outside the IoT space Across the edge of the business, if you will And so we use this envelope story So OMI is like the envelope It's addressed to and from The ODF is the structure Of the letter that's inside of it And the ODF Allows standard meaning To this envelope Of the content of the letter And so Rod will show you a little later The periodic chart And how we think of the notion of Elements coming together to make compounds Or data becoming information So those three standards Are the basis of a lot of these talks So this is a really simple story Of IoT or industrial IoT And IoT, when I say the word IoT Often think it is the consumer way It doesn't really have to last as long And when I think of IoT It has to have reliability And maintainability over time It usually costs a little bit more as well And it might be a difficult environment Where I don't want humans to go to In some of that nature Where's my IoT watch? I don't really care, I can just find that one easily And so you'll see that I've got ODF, the open data format And you can see on the right Of the OMI, the message interfaces Between the two of them And then on this one on the left You'll see the whole stack of ODF With OMI is in there And you can see all of those objects And normally when you draw this picture If you were just to go online And see this picture You'd see an IoT gateway And a cloud goes off to some apps But I've included the semantic rules engine Because in my line of work I'm increasingly doing that And what I'm doing is I'm classifying That data with machine learning So as I get that ODF, ODF content What I'm doing is running a piece of machine code I'm classifying that data And how it compares to other known Vocabularies And that variance becomes an event And I do something about it Again, I look for known vocabularies By vertical That's what the industrial and consortium is very good at But there's some other names out there For your various industry verticals That you're at But I look for those words and now Once I can classify and I can tag them I can say this is a tomato This is a bucket This is a pound or bitch Or long hair or short hair Or whatever it is And all those are kinds of dogs And I have things like lasara Multiple thesauruses In order to compare that too And once I do that It means that everything I put in the IoT cloud I know about How many times a day do you put stuff in the cloud And you don't know what it is Is that dumb? Can you imagine putting stuff in your basement It's a mess with you What a bad idea Do you know where things in your basement are No, don't you lie to me Can you imagine putting stuff in the cloud You don't know what it is People do that, everyone does it Raise your hand if you don't Of course we've got to stop We've got to stop, we should know what we've put in there So the IoT gateway A lot of us use machine learning semantic rules Through the beauty of ODEF In order to know what we've put in the cloud And we can mask it and we can encrypt it But we know what's there And now I can move that data Onto my mobile apps and my business applications And demonstrate business value And if I do business value Now I can measure the Relationship between these sensors And the metrology or metrics it produces Compared to my strategy And that's around your ticket And now I've proven Business intent And that's what enterprise architecture Is supposed to do Is manage business intent across the enterprise Okay, so isn't this a mess You saw this earlier, it wasn't so messy I hope this doesn't offend anyone But I'm going to go with it Now I've done some mean things to this beautiful stand I've extended the little white box Isn't this beautiful what you can do with graphics And I've added these little dots here Which you can see from my key Is the industrial IoT And so the gentleman this morning Talked about how you could have sensors down there And that's just what this is And you notice I've penciled in ODEF So now these sensors can talk to each other They don't have to go up the stack They can exchange messages between each other Because they have the same data model Either homogeneous or heterogeneous But in any case we can use ODEF to do that Or we can send our ODEF packets Up to the real-time bus And we can go into a relational database Or we'll see from the next slide We can do some even more special things But once I'm in a relational database I can take those encapsulated messages I can draw them into an ODEF graph Which is just an edge and node Sort of picture of a thing And then I can go into some W3C standards Like owl or sparkle And from there I can I'm creating information products Now I semantically understand that data And I can connect that To business capability Up to my strategy Now traditionally this whole picture Is what we've done in the factory Or an OT And what we're doing up here Is traditionally been the world of IT And so we put them together We have OT and IT And this is what every manufacturing shop I think on the planet is doing OT and IT are coming together To perform one factor Why? It's because of you And you, you millennials You're wrecking our broken world And you're making it more efficient Because when you're executive You really want to have OT guys and IT guys You want to have one bunch I think you want to have one bunch Because one bunch is a lot easier to manage than two bunches Especially if they're different cultures And so by doing this We now can turn this into a OT, IT party Or a singular equation Where software is used all over the organization In order to demonstrate value Now what I have will be an open platform 3 Because in that world The open platform allows us to have catalogs And I typically put a pencil in it Three catalogs, pencil in it And we do what the company work at Is we have data catalogs Product catalogs, we sell products to each other Within the organization of manufacturing We have service catalogs And you would think that all three of those should talk together Well the open platform 3 Supports this notion of multiple catalogs And now I can have a complex business Ecosystems of multiple catalogs Working within this space I now can start to see that I got All the different locations all tied together Because of the open platform So that's OP3 With WC3 With OPAP, the open process automation Form With ODEP So this is a simple picture Of the picture that you just saw That I just redrew an hour ago So we have your factory up here Sensor data and we go through Kafka And you can wait in Kafka Essentially With Kafka I can subscribe to Subjects From subjects I can run through spark streaming And now this is going to turn it into RDF And now I can do some various things from Structure data with spark sql Not to be confused with sparkle And graph x And now I can start to have graphs And I can have essentially bar graphs And now I can get information from that And use that information to go back And influence the factory And guess what, humans may or may not And so this notion of semantic Relevance Between resources based on ontological Representation That took me a while to write that down If you can figure out I'm sorry, that's about as simple as I could say it But what I'm saying here is through that Tagging stuff I was talking about If we can take our resources and tag it Now I can get semantic relevance Because of ODEP Because ODEP, which you'll soon learn a lot About movement, allows me to take Lots of words and give them integers And to wrap them into some extra stuff So this is a nice simple picture And it sort of replaces that complex one That we saw earlier Take a picture of it, this is a couple of graphs In this graph We had some graph dd in the last picture I've got some sensors here It's got some temperature and some Location and some values and so on The beauty of an ODEP Is it allows me to take that data And then semantically organize it To some words That allow me to use English language To sort of deal things with it So the graph can move to your classes And it's in a semantic Character in so much as I can Make sentences and ask questions about Which I'll see in a second And this context allows me to build These information products Now it'll be real clear in a second The notion here is you want to think about Is the notion of taking ODEP And moving it into a way That I can use things like American English And let's just take a look at that So we've got some questions here The last one was that data that you just saw Was control data What controls represent quality limits The basis of the semantic web Is what? We should all yell this out RDF The basis of the semantic web is RDF It's a beautiful thing as an adopter As I do because it's beautiful Because it's three things Subject predicate object Look at that subject predicate object Isn't that beautiful? So the control is my subject The verb form is that it's represented It's a condition of what is And the quality limits are The predicate objects And so now I can take that data That I just saw and I can start to Have humans write business Questions about the data And so this is really the story Of tracing of taking data Architecture and joining it to information architecture So it's traceable to business value Now this seems I mean we've covered Two years worth of Graduate work, 15 minutes And so I'm going to let that sit for a second Because now I'm going to really Hopefully deploy our error Because this Topic of semantic applications Has been fascinating Because how much What percentage Of your application development Time is involved in UX user experience In the front end? 30%, 60%, 25% 25% More than two digits I mean more than a lot of time A lot of cost We're trying to figure out what humans want Well I claim the following The semantic application does require A little query is made Isn't that fascinating? So you know that when we When you have a business strategy And you have a capability You're going to assign to it You assign human systems to machines And if you give it to a human They've got a role And roles answer questions And the idea is that You can go to apqc.org And you haven't seen that one But most of this is all seen in 536 questions that they have for CIO They have I forgot how many for an aerospace engineer But they list all the questions Well the idea is that we think We have to build the application Based on their questions But I claim that's wrong I claim with the semantic application You ask the question And the front end is to sign on And approve it in a second But there's a really lousy definition At the WC3 site Is already at the Webital And ontological language And other metadata standards I said this That doesn't require front end The semantic application used by our system Or machine doesn't require front end Is consumed as data Now you want to see a while I'll show you in a second So this notion is Information products are not being exploited yet How many times Has anyone seen the notion of semantic application I'm thinking chat box right now Well that's good I'll stick your hands on that That's getting really close So watch this a minute But it's going to happen a lot I think the next couple of years Is going to just blow up So finally This bridge is really about Data information We've got the screwy Rup Goldberg thing going on Once you have information You can make decisions Human systems machines have made And put that back in the bucket of data And just make this And continue to improve and make it better But this is the piece that I came to It's weekend standards You notice in my previous pictures There were lots of three letter acronyms and so on Because those are the things that come from Patching these other groups And the great thing is that It's nice, it's working But we need standards around it I can't just say things like RDF The better ways of doing that The standards are a way of doing that And so that would be one more thing You have your questions that you'll do And then I'm just going to Open up the browser Because I can And this is just A browser in a command line So I'm going to type in some stuff So let's see A bottom bucket here Anyone know bottom bucket here Now look how Google's Formatted this Bottom bucket head was my question That I wanted to know about And I didn't know when I exercised Bottom bucket head what it was going to give me It gave me back my application And look, there's an article about Wikipedia There's some pictures of Bob There's some videos of Bob There's a whole way of interfacing With Bob What else is on here And then there's the normal stuff That you might use For instance in time If I just So if I type You know who the funny how you get old You don't know it's you And so like it comes up It's a little bit about me Some pictures, some apparently videos They say I was in A bunch of pictures And so on But this whole little application has been made For me based on data that I've entered So Before But So apparently this kid right here Is related to me He's a civil engineer I'll look at his Facebook But his application looks different than mine does And if you type on Somebody that's on a roll To the thing I'm not political But let's see His application based on The data that Don has put in Has This application That shows all this stuff This is the beginnings Of this notion Of a semantic web Of linked So the notion is this is linked data So there's a thing called DBP There's Wikipedia The thing behind it is DBP Which is series of collections of your eyes The beauty of this from a librarian's perspective Or someone in information sciences I'd like to have a catalog of your eyes That's called what? A catalogue Can you imagine a catalogue that's for Don For Rob or Chris This is designed for them When they press the button They go to link data And can you imagine that there's only one kind of Data that we're all linking to that For instance, if you go to Wikipedia How many articles are there on Joseph Stalin To be the guy from formerly a Russian Does anyone have the answer to this? No We know one Because that's linked data We'll all go to that same piece of data And that's the beauty of linked data So now we can have a semantic web Of linked data objects Using URIs Back to our service data product catalogs That use those URIs Down the stack To our infrastructures We can tie all together into a big mesh Now, I'm one minute over So we're going to take some questions So the idea behind this is we're setting Sort of this foundation This picture to over the rest of the afternoon But the idea is that our OP3 Whatever we decide to name it to We change the name because it's moving faster When we came up with open platform I don't know how many years it was But it was a very new concept But now what's happening is things are Starting to change faster We're starting to deal with autonomous vehicles Autonomous, all kinds of autonomous things We recently put out a paper about The maturity levels of autonomous things And now we're starting to think about Well, how mature should our NL be Where it was live Because what if you're running ML For finding library books And it's 5-9 is compliant and 5-9 is accurate Well, that means that essentially 5.34 times a year You're not confined by a library book What if it's humans? That means 6 humans Will die because of your data Because of your NL That's probably not tolerable This is kind of our story So the idea is that ODEF Encapsulating whole line Whole line of ODEF Inside the OP3 framework Using some of the consulates like OPAF And some of the other ideas we've shown Is the basis of building our semantic web And I do have to make an apology About 4 years ago 5 years ago we were in London That church place that we had The CIO We've been there a few times I think it's the guy It's Reagan Ryan The guy from the CIO guy He made fun of me He said you really should read more He was talking about the definitive paper From 2009 on the semantic web And so I'm publicly apologizing to him But I've read it a whole bunch of times The last few years That really is the definitive paper On this notion of moving away From the historical web That looks like a bunch of library books That's connected to a semantic web That connects all these All these linked data with URIs So we have a less costly Less slow Incense of an information space So things don't cost as much And they're interoperable So cost latency and interoperability The three pieces that we want to think about So that said We've got time for a few questions Before we jump into the rest of our afternoon Thank you The information age And things like Fake news that are out there today Where do you think the standard Is going to go for data As this Semantic advances over time Are they going to add For instance a higher quality Validation to data collection As part of the overall standard To so we can avoid Some of the false information We're tracking metrics on False positives That may as well come back That's such a fascinating question That we could do a week long conference And we'd just go on for the rest of the year Because the notion of truth So in Aristotle's poetics He wrote quite a bit It seemed like every army officer Was required at some point in their life Or sometimes to understand truth And what is And we used to think that truth Was this sort of permanent thing That was fixed in concrete That was immutable and universality That all truth would appeal to all peoples It seems like that's not even true It seems like we're making an idea To do more at Gallifars The idea is that In enterprise architecture We know about this notion of view And viewpoints We recognize that Given your viewpoint You have that number of views on the thing And how it turns out The truth is, or that particular view Is interesting Because there's not a universality to it There's not a common cultural On Earth The answers to your question I can I want to answer In a way that's relevant to Us here in the standards But I jump in that I try to imagine when I think about Just what you've asked On Earth And to me from Earth It looks like a bunch of brown people Because most of the world is of color And so When I'm on Earth and I look around And I see how governments and so on Are working these days I see in the news it's populated In a different way I see a fairness model That may be not popular in a way That I would if I was standing On Mars looking at Earth I mean there's not a meaningful truth And how do you register What version of truth you want to see So I think it's sort of beyond This discussion but no I think you and I were on Skype recently And I had seen this Someone had developed a piece of program That allowed a Former president of the United States To speak In a way that had nothing to do With what you would know what to say To be filmed Of all those mouth parts moving To actually have him say Things that you wouldn't have said So we're to the point now That we can manipulate the truth And how do you legislate against Those kinds of things? Do you build a world Of warm laws? Or do you Teach people more so that They perhaps don't want to do those kinds of things So I can't really answer Any questions I suppose because The answer is so complex And I keep thinking it's Like if we stick with things like We have free education and everyone in the world That's educated and things get better But what do you teach people? I mean the U.S. Army taught me And I don't want anyone to learn those lessons So that's maybe not a good pattern So I want To give you an answer but the only thing I think I can rationally say Is that We all need To stay alive We all need Life and pursuit of happiness We all need health and education And what we're seeing in this Machine tool space I personally and this is what I do for a living I know that it changes the Dynamics of jobs in big companies Because I'm there and I'm watching it And one couldn't think that It would make companies kinder And so I don't know the answer to these Because I know a reality Is not appropriate to this Conference but I don't know the answers to those Great question I was just thinking where would you put it In the standard? Would it be on the Ingestion of data? I can imagine you have problems With concerns about Censorship and First Amendment rights Other things And computers actually controlling that Is on the ingestion part Before it became part of the data model And so from the standards perspective How would we improve quality data? And it may not just be Information You have information hiding So remember we had robot Text so you would set that in a tree So a robot would sort your tree And so you put it everywhere And then you have All those boxes Honey boxes all over the place To try to get people to go to those Instead of this one So we have ways of hiding data And we have ways of adding a calculator We can influence the calculator The reason that the calculator Stakes sense from past factories You could essentially have one calculator With the one that I deal with quite often Is say I'm making an object We'll say whatever this headset thing is here So I make this thing But a box of them is supposed to be Shipped out tonight But I'm missing a part So I don't get a comparable part Does it get the same part ID as the other parts? I add a letter X to it Let's say And then that stays with it And so that calculator is now changed Then it happens with another part next week And consequently my bill of materials Is now grown This is a real story Imagination So say a 737 That has two main parts on it And 1200 suppliers You're constantly dealing with the notion That I finish the plane I don't have to pay taxes on it And then what happens to it And then what happens to it Is that I get I have to do part replacements Because I didn't have something for my supply chain So my supply chain mathematics Turns in from linear algebra No longer can I do supply chain Anymore incidentally Just being able to forecast What's happening in an extractive way I have to do complex calculus Because now I've got multiple supply chains And globally snow, bad weather, windstorms People running out of gas, somebody's dog in their homework All that's happening And affecting my parts getting to supply chain And I have to take that into account And so truth becomes Sort of this funny story Because all information Is increasingly available How I put it together In a way that's honest And not trying to hide my contract So that I don't get my place In the assembly line kicked out So I was late for my part You're breaking my day You know We're in this room We're supposed to have these answers Our executives come to us And they ask us those questions And so that's the question If you have a question Just raise your hand and I'll bring the mic to you Pressure around Just don't give you that question Anybody have any other thoughts about this one? Me When is ultimately your best When we're regulated least But it's been proven That every once in a while I end up in another wound in my body Because of that Apparently it's not quite true But I won't believe that it's true So when it comes to developers I don't want to regulate developers I want to think to compile And then to fulfill this specification When does the machine be compiled From the legal perspective That's a good question Well let's run through it Corporations are people in the US So if you're A corporation apparently alone I've got one for you because this just happened to you This last year, some of you know the story What happens with I'm the CEO of a founder of a company In Slovenia And I hire two assistants And they happen to be machines And then I build my org chart And I'm first in line And I hire a company And they take over So now machine A is the CEO Machine B is the operating officer Are they responsible for the company? Well I think they are What if it's something that's outside Their domain and control of curfew What if they give you the wrong answer Well I think they are So now machines have responsibility And they can help the camel And if I sue them and they pay up What if they don't pay What do they do? These are areas of humanity that we've not Figured out But it's real That actually happened by the way It's a test case for you to go read about it