IoT GUIDE

This is an episode of the "Ventures in Industrial IoT" series brought to you by GE Ventures. In the series we explore success factors and challenges in Industrial IoT markets with CEOs, investors and experts. 

IoT Spotlight Podcast introduction:

The IIoT Spotlight Podcast shines a light on Industrial IoT solutions that are impacting businesses today. Every episode we interview an expert about a specific IoT use case. Our goal is to provide insight into the planning and implementation of IIoT systems, from new business models to technology architecture selection to data ownership and security. The IIoT Spotlight is produced by IoT ONE, an information platform that provides market insight, partner development, and go-to-market support for technology providers, end users, and investors. Don't forget to follow us on twitter. You can contact me directly at erik.walenza@iotone.com.

Summary: 

What does adopting an IoT perspective mean to you? With a concept as broad and with as many ecosystem players as the IoT, this is a question that is likely to get you a range of answers depending on whether you talk to a technologist, an executive staff or an academic.   

In the ninth installment of the "Ventures in Industrial IoT series" brought to you by GE Ventures, be prepared for a highly dynamic discussion with Tim Chou, Executive Chairman of Lecida and author of the book, "Precision: Principles, Practices and Solutions for the Internet of Things" to share with us more on what he thinks adopting an IoT perspective really means to a firm and how it can impact the way it does business. 

Learn more about Tim Chou and his books: http://www.precisionstory.com/

Learn more about Lecida: http://lecida.com/  

Learn more about GE Ventures: https://www.geventures.com/

Links:​

Libsyn:  http://directory.libsyn.com/episode/index/id/6082468

iTunes:  https://itunes.apple.com/us/podcast/industrial-iot-spotlight/id1228185407?mt=2​ 

Speaker Bio: 

Tim Chou is a man with an illustrious career spanning academia, startups and large corporations such as Oracle where he held the title of President. Currently, he is the Executive Chairman of Lecida a firm that uses AI and Machine Learning to optimize the performance of industrial machines.  

He started teaching in Stanford in 1982 and currently teaches a course on cloud computing with the module code CS309A. Aside from that Tim is also the author of the books "Precision: Principles, Practices and Solutions for the Internet of Things" and “The End of Software”.

Company Overview: ​

From scissor lifts to wind turbines and crop sprayers to granulators, industrial machines are producing huge volumes of sensor data. Lecida is a firm that uses AI and Machine Learning to learn from this data and power the next generation of intelligent industrial machines. The firm was founded in 2015, by a team of engineers from Stanford, Berkeley and Oxford with deep expertise in machine learning, distributed systems and cloud computing.  

Blurbs:

"In terms of IoT adoption, we're in year one or two of a 25 year cycle." 

"Things can be programmed. People can't."

"Your machines are increasingly software-defined. The 2016 Porsche Panamera had 2 million lines of code in it; the 2017 100 million lines" 

"The margins on the product itself are 3,4 or 5 percent but the margins on the Product-as-a-Service are 30,40 percent"

"What the digital transformation that we love to talk about fundamentally is, is a transformation of business model."

Show Transcription:​

Welcome back to the industrial IoT spotlight. I'm joined today by Tim Chou. Tim is wearing quite a few hats. He's the executive chairman of Lecida. He's a board member of Teradata, Blackbaud and Cloudbook chairman of IoT at Alchemist Accelerator. A lecturer at Stanford University and Tim recently published the book "Precision: Pprinciples, Practices and Solutions for the Internet of Things." Tim before we launch into the topic you've got quite a varied and impressive background. What would you like to add in order to just kind of frame where you're coming from. 

Well let me frame it by explaining why. About now 12 years ago I was working for Larry Ellison running what was the beginning of Oracle's cloud business. And as Oracle was shifting morphing into acquiring a number of software companies I kind of went wow. You know I didn't really want to be worked on a lot of operational stuff which is generally what happens in the acquisition phase.So I thought well I have long enough to figure out what I'm going to do the rest of my life which I was thinking I wasn't going to go sit on a beach or go fishing. So I took the opportunity to become a polygamist is how I described this. 

And for quite some time I had two questions in my mind. One was what was next.For enterprise software. Because we had built ERP and CRM cover had generation one and generation two. We put it all in the cloud but really functionally there was no. Real innovation was just the same stuff delivered in a very different economic way and a higher quality all all good things but no new function. And the second part is I have been in the cloud space for quite some time so the guys at Amazon had given me three thousand dollars worth of computer time. For my Tsinghua class which I started about six,seven years ago. And I walked into class and said hey well here's three thousand dollars they'll buy you a server Northern California Virginia or Ireland for three and a half years. Much longer now than it was three and a half years or it'll buy you 10000 computers for 30 minutes.

And which is what I really wanted to think about which is what would you do with that.

And I had a similar question in my hands. So those two questions are in my head for a while because I started this class at Stanford.

Called if anybody's interest it's called CS309a. And you can see a little brief about it that's the cs309a.stanford.edu. Had had at Paul Maritz in my class. Paul was then the CEO of VMware and so about 12 months later I had a call from one of his former employees who had gone over to GE and they called up and said Would you like to have Bill Roode come and do your class. And I said Well yeah I'd heard you guys because I live in Silicon Valley I heard you guys are doing something over there in San Ramon but I have no clue. So I showed up and I expected a marketing pitch. And I. Soon learned a lot more and I went. In crap. This is this is big. We as technologists as software people we pretty much have only solve problems and you know the three big industries financial services retail and telco. But when you go over to a mining construction transportation health care etc. You know we've done very little in World Economic Forum just released a report a couple of years ago that said the Internet of Things will. You know affect two thirds of the global GDP. Because most of the GDP is all this infrastructure stuff. So I about three years ago pivoted into this area and decided you know I'd push all my chips in. Hence as you already can surmise I'm interested in investing in the area small companies interested in large companies roles in this and interested in educating the market. I figured that since I didn't know anything the best way to figure it out was to write a book so I ended up writing Precision which we published a little bit over a year ago and launched on the River Thames in London and this year was a big year to push it into Asia. So we actually translated and published in Vietnam. And then India. We didn't have to translate we published with McGraw-Hill. And then I just got back from China where we launched. The book published by Tsinghua University Press. Anyway that's kind of why all that all fits together let's say it that way.

Yeah. It's interesting you noted there that on the one hand we have we're still largely figuring out how to make these technologies relevant to traditional industry. On the other hand we have these research institutes and. Intergovernmental agencies that are publishing very ambitious reports about how these technologies are going to completely transform the economy. We seem to have a bit of a bit of a gap between expectation anticipation and and where we are today. I think we want to get into your perspective on how we can move towards that future where where companies really are comfortable adopting technology know how to prioritize and so forth but maybe before we go there if you can just give a quick perspective on where are we today in terms of companies actually being able to make rational analysis of what's on the market finding solutions that are relevant prioritised and actually bringing this into their businesses.

Yeah. To answer question I think we're very early. I mean to me this is at least a 25 year cycle. It probably looks a lot. You're too young to know this but it looks like a lot like the move to client server. I mean it's big it's pervasive so but we're early. I don't know year one or two out of a 25 year cycle. And I think largely because. As technology people meaning and I'm mostly talking about software. It's not an area that anybody's been particularly focused on per my comment. Most of us in enterprise software were focused ever in financial services and retail. We never talked to anybody in mining. You know part of it is a pivot of the technologists themselves and a fundamental realization that most of the technology we have built has been for what I like to call the Internet of people in a way that's an e-commerce app or a CRM app. We really believe at the end of the day it's in service of a human. And I go. When I teach my Stanford kids this which is here's a tremendous piece of insight. Things are not people right. Things are not people. I'll give you five reasons why they're not people one way more things connect to the Internet than there are people. I mean if you listen to the John Chambers speeches 500 billion things will be connected well that 100 times a global population rough. Right. I talked to a team at extension Health the largest health care providers in the U.S. and he goes yeah already happened here in the hospital there way more things connected and people tell you things can be where people are not things can be in your stomach things can be a mile underground in a coal mine things can be out in the middle of the Gobi desert. You know this is not like people there things have more to say. I go Look all we can do is type a little bit move a mouse around. Modern day wind turbines have five hundred sensors on them. They have way more to say than any human ever could. Fourth they can say it more frequently in the mining industry. There's a machine called a Longwall shearer it basically digs through a coal seam and as it digs through it forms an artificial roof. Well one of the problems that happens is every once while the roof collapses and they have to go dig out there 100 million dollar machine. So they want to be able to predict roof collapse. They put a vibration sensor on top that runs that 10000 cycles per second. I mean that's way faster than any human can time. And then the last fight and we can debate it is you know things can be programmed. People can't. And so if you fundamentally come to a realization that things are not people then why would technology build for the Internet. People worked for the Internet of Things. And that's why I think we're in the next whole generation of software technologies because of fundamentally that reason.

OK. Interesting perspective when you say we're in year one or two or two so out of a 25 year cycle I think what we've seen in China here is in 2015 16 let's say if we if we take kind of the conference keynotes as an indicator they were very much on. What is industry 4.0? What is the IoT? And then these kind of graphs that we've all seen about you know stage 1 2 3 4. No. Now we see companies getting into the what have you done? How did you do it? What problems did you have? So I think. Already we have a good transition at least here in terms of the mindset and you could say if you look at who has the who has the bandwidth traditionally to look forward. Yeah banking, retail, government also at times has been one of those right, for the U.S. government pushed for the space initiative and so forth. I think the Chinese government as an example is very. Clear minded that they want. To push the development of industrial IoT that they see this as a potential so at least here that it's accelerating and I think the German government's already done a good job in the U.S. we seem to be from my perspective a little bit less less focused unfortunately. In terms of what you see so you're serving on the board you're advising quite a few corporates in different facets. What types of conversations are you having with folks around this transition from OK we've been developing a strategy now we have some concept that this is this is critical we have to start you know gaining experience developing pilots scaling pilots. What conversations are you having with folks at that level as they try to figure out how this can can really as you say double you know doubling revenue quadruple your margins.

Yeah I think let's just let's divide the conversation because I think it's useful.

So there's a conversation about what I'd like to call people who build machines. So you know Radco builds tractors combine harvesters or the builder of gene sequencers and differentiate that from the people who use machines. So you know hospitals use machines right. Farms use agricultural equipment. So.

We'll come back to that later but let's just set with the conversation about people who make machines. So here's my observation. I think. You know a lot some percentage of those guys because we all. Talk a lot in the Internet. They have heard about this IoT thing. They kind of like oh yeah maybe we should do something. But and I point out that you know Cisco released a report earlier this year that said that 75 percent of these projects have failed. Or are stuck in POC. And I think the fundamental reason for that is that you know today. The buyer. The buy that the buyer meaning the guy who's running the agricultural equipment company and the guys running the construction equipment company you know transportation equipment. They haven't they typically I'm not saying this is everybody because Tesla is a counterexample. They typically have not. Software is not an important part of their business. The important part is horsepower and torque. And so if you go into their companies and try to find anybody with software expertise. Pretty much the only guy you find is a guy who's been in charge of the SAP implementation working for the CIO.

 Over on the product side meaning the. The the the make the machine side or service the machine side. You just don't see that much right. And so I think the traditional I'm going to tell you really cool technology because you can consume cool technology. Which is the way we've done it in the IoP world. Doesn't work really well because there's nobody over there listening or asking etc. and if they are okay great they found some cool package or cool thing and they you know built a connection to their machine or they did something and then they go into rollout and I have heard this story multiple times. And the business goes well why would we want to do that. That's a cool science fair project but. Why do I care right?. Shouldn't we invest more and you know building more horsepower and more torque is traditionally has been done which by the way the guy who's running the product group there as well schooled these mechanical engineer well schooled in how to do that. And. And here she says you're going to spend a million dollars. Give it to me. Right. And so I think without a shift in understanding. You know a new business model. This is just going to be us you know talking to the few guys in the few companies who think technology is cool. So you know my my thing over the past couple of months actually because I've come to this realization having a talk and spoken to a lot of these guys. Am I. OK well let's start first of all premise 1 Your machines are increasingly software-defined the example I use right now is the 2016 Porsche Panamera. Had 2 million lines of code in. The 2017 has 100 million lines of software right. So. If you appreciate that Your machines are increasingly software defined then maybe you ought to take a lesson from I'll call it my business which is a software business and I go. What are ours. Business models what are the software business models. Because the machines are increasingly pieces of software. You know I did a talk to all the ninety CTOs of the Tata group about a year ago and I got up there and I said. You know the machine of the future is a bunch of sensors and actuators a big computer in the middle with a lot of software. And later in the day the CTO of Jaguar Land Rover got up and he said Well you know. Tim said it's a bunch of sensors, a bunch of actuators big computer in the middle of my software. He says I really wish it were that way. I have 18 different computer systems in my car and none of them talk to each other right. 

So if you eat you make this mental leap that it's all about software then okay what are the software business models and I will tell you they're fundamentally three models model one is. Sell the product and a disconnected service. So if you're a student of Oracle Corporation frankly of any of the traditional software companies you would know that the business model is largely. Sell your product and sell you a service contract on that product. Which is roughly. 2 percent a month. Of the original purchase price of the product. Recurring every year. Or every month. Right. Oracle in the year before it bought Sun. So it was purely a software company did 15 billion dollars in revenue 3 billion was selling you a new product database application middleware etc. 12 billion was service contracts on the product that had been sold all the years before.

And if you go be a student of that you will come to the realization that it is not break fix it's that's not what is being sold as not what's being bought. It's not break fix what it is is access to information about how to maintain or optimize the performance availability or security of the product. And that that makes it purely an information transfer business which is why the margins on that on that business are north of 90 percent.So a north of 90 percent recurring revenue business. It's something we should all want right enormously high margin recurring revenue gap product plus disconnected service.

 Model two.  If I can connect to your machine connect your software I can provide you more relevant more personal information about your product. I can say you should put the security patch in and make it more secure. You should change this. The disk configuration to make it run faster. Right.nd I can tell you all those things and actually at Oracle other people have done this than in the high tech industry we've done this. We've connected the machines and we will charge you incremental amount of money to provide what I refer to as assisted flash connected services. Right.

Okay. Last Model 3. What is that. Well if I can tell you how to manage the security availability and performance of I can optimize the security availability performance of the product well then shit I can do it for you. And so what we've also called SaaS Software-as-a-Service. Is merely the fact that the guy who made the product is now servicing the product.

And by the way at that point in time can change its business model to charge you you know per seat, per payment per transaction whatever. Right. And the whole renaissance in enterprise software over the past 10 15 years. All these new startups whether that was NetSuite, Salesforce, Workday et cetera are all fundamentally developed in that model. Right. So and by the way just to make a point of it that's exactly the same thing it's happening. In the computing machine business. Once upon a time we bought the computer and bought service on the computer. It has evolved into. I buy the compute as a service from guys like Amazon or Ali cloud or Azure or whatever. Right. That business model 3 the Product-as-a-Service is already happening in the hardware called the hardware product space.

Okay so take that and go. Okay let's talk about machines. First of all right machine plus a product your you know your semiconductor manufacturing equipment your wind turbine etc product plus disconnected service. Well most people don't realize this but General Electric. Why are they so interested in this Internet industrial internet thing. All you have to do is be a student of their financials and people are. Today being much more students of their financials. But what you'll see is last year they did 110 billion in revenue. Fifty five billion was sold in product meaning a jet engine and a MRI scanner right. And 55 billion was service on those products. And since they signed multi-year contracts for that. They end up showing you a backlog of 250 billion dollars worth of service business. Now realize they're operating in a 50/50 model. What if they could move to an 80 20 meaning 80 percent was service and 20 percent was product. I mean right away. I mean we are looking at a 30 percent larger company at massively better margins. I actually spent time with a European company who operates in a 50/50 model. In 2008 he said Oh yeah our revenues went down but our margins went up. And I go yeah I know that's exactly true because the margins on the product business are like you know 3 4 or 5 percent the models and the margins in their service business are 30 40 percent. So model 1 I deliver your product plus a disconnected service model to I connect to it. I get all this sensor data right that everybody likes to talk about. And based on the sensor data I can provide you even better assistance on how to maintain or optimize the performance or availability of your wind turbine your roll gate or your blood analyzer et cetera. 

And if you just want to model this and I tell the guys who make machines just do this simple model take. Make your service price your service product be 1 percent a month. Of the selling price of the product. And look at your installed base. And make the assumption that you can sell every one of your install base a more advanced service. I'll tell you it won't take you very long to realize you could double your revenues and quadruple your margins. And by the way you're only going after customers you already have. We'd set you up for model 3. What does model 3. Well if I can tell you what to do I can do it for you. It is what I like to call product service or machine as a service. 

There are early examples of this out there but the one I like to point out to people because most people like you know know about consumer products is the automobile business right. OK. We sell people automobiles and service contracts on the. You know guys like Tesla connect your automobile and provide some additional services. But if you want to think about what is automobile or Car-as-a-Service. Well just go out and. Bring up your DiDi app or your Uber app and you're looking at it right. That is a per ride business model for an automobile. And furthermore. If you're a student of this you will realize that. In my world of software and hardware. The number one cost for delivering a Product-as-a-Service, a Software-as-a-Service or a Computer-as-a-Service is human labor to manage. The software to manage the hardware. Well what do we do in software as a service and hardware service. We automate the hell out of it. We take the human out of the loop because we improve quality and we crunch cost down. We'll guess why everybody's interested in automation. If I could take the human out of the loop right. Why would anybody ever own a car other than. I mean I love cars. I mean to be fun it's fun to own them. But there'd be no reason all cars would be delivered as a service. Right. Which is why it's so disruptive. I mean people may have noticed Ford just changed out their CEO this year. Right. In an industry which has been thought to be very stay, conservative not changing right. You're seeing a massive shift. And I tell everybody who's builds machines I mean maybe you can hang back and wait. But at the end of the day if your competitor does what I just said you know builds a product plus a disconnected plus a connected and then moves into machine as a service is not an overnight transition to get there. And you're going to be you know left behind. So I see it either as a Carrot problem. Double your revenues quadruple your margins or see it as a stick problem and you're going to get put out of business because the next guy your competitor is going to do it in health care machines agriculture machines construction machines transportation machines etc.. So I know that was a long answer but. That's the fundamental thing and I think we have to get is get. The executives of companies that make machines. To begin to grock that what digital transformation which we love to talk about really means is what I just said. It's a fundamental transformation of business model.

I think that's actually a super clear answer so very very well-structured but I think at the end pointed out one of the reasons that this is not occurring as much as maybe you think it should. Which is that it's a long transition and a lot of the companies that have been successful adopting these models have come out of the VC startup environment right where they are small innovative teams they are given a relatively strong. Let's say runway from VC financing as long as they produce growth they don't necessarily have to produce profit no profit in the near term because there is the expectation that. By by generating this data these economies of scale securing a semi monopolistic role through the providing of services they'll eventually develop a very profitable business as Oracle and Google and these other businesses in the space have. And that I think at least my perspective and I'm interested in hearing how you how you see this. I think that's a challenge for traditional business which is looking at quarterly or annual margins and saying okay you know I was talking with a system integrator last week and they said Yeah we see a great opportunity to provide. Maintenance-as-a-Service or predictive maintenance-as-a-Service. However we don't see it having a strong. Revenue or margin improvement in the near term. We see a lot of other benefit in terms of us learning and feeding data back into our R&D into our production. We see a lot of opportunity to learn and we feel like this could be you know a strong business in the future but somehow they have to make the case if they're going to invest heavily in this. To basically take the dive and act almost a little bit like a VC and say hey we're going to finance this internal organization that's going to build up a new solution launch it to market and expect that it's going to be a highly profitable high margin business maybe in five years. When it matures and that's a very different perspective to how they traditionally look at investments. What do you see this do you see that this is one of the barriers or is that just maybe. My experience in in a couple of instances. 

I mean I think that well let's try it this way. I think the barrier is both knowledge and leadership make what I just said is not the traditional way to see your next generation blood analyzerbusiness or your next generation tractor business right. It's a feature function game. I like to keep using the word horsepower and torque as a proxy for this.  Your VP of engineering sitting at your executive staff. Your product management teams. They're all organized to build more horsepower and more torque. So there's a leadership question which is do I believe that you know I'm going to take some money and into the part that they don't understand another part they don't understand as I actually did talk with my GE friends on this subject is hardware is not software main software is not a game of 5000 people. In fact it's like counter indicated. Right. It's a game of 50 of the right people. So it's not about you know spending billions of dollars. It's about spending millions of the right dollars. But. It will require leadership because you I just talked to the American equipment manufacturing association and a couple of guys came up to me afterwards and I said Yes I know which is you know the V.P. of products is fighting this because he goes You're not giving a budget to put in more horsepower and more torque. And second of all even if you gave me a budget to build the product. But by the way your sales team doesn't know how to sell it. 

I mean I just spoke to a I won't name names but very large semiconductor Manufacturer. Sat down CEOs said how many machines you have in the field. Tens of thousands. Wow. How much service revenue do you generate. Zero. Zero. I went why? Well nobody wants to pay for service Well. I know why it's because you tell them you tell them that service is break-fix. And I go well I just bet 250000 dollars on your machine why should I buy break-fix service. I thought you gave me a reliable machine. You're not selling them that service is information to maintain or optimize of performance availability or security of your semiconductor. You're not even saying that to them so why are they going to buy it so you're. You've got to have a sales organization that's going to sell it. You have to have a marketing organization that's going to market it. Your business operations is not going to be the same meaning you're used to selling them a big tractor. And then paying you 250000 dollars for it. And you walk away. This is this doesn't look that way. This looks like here's a contract which is renewable at five hundred five thousand 50000. A month. So you're going to have to manage a recurring revenue business which well by the way change revenue recognition. So. It's systemic to create this business. It's not only the products technology side but it's also how you sell, market and account for it. And if you if you cannot lead people if you cannot. Make the big decision to go down this path then it's just going to flounder which you know I think right now that's the challenges. US traditional technology people will show up and want to find some guy who wants to listen to why my neural networks better than your neural network. Well that's cool may we'll find that out. At the end of the day. That's not that's not how this is going to be bought. My personal opinion. Right. It's going to be bought because the CEO the executive staff sits down and goes damn it I'm tired of saying which you know go go be a student of Caterpillar and you will say quarter on quarter year on year decline. In revenue. In product revenue and why there is zero service revenue at Caterpillar zero. 

So until you get that leadership religion and then start to ask the questions form the teams hire the people I've told guys and as I said you know you need a chief services officer. The chief services offer needs to sit at executive staff. That's the only way to play this game. You can't bury it somewhere. Which I've seen that one. Oh here's a team that's three layers down inside the CIA organization and they have the quote IoT initiative. Ok cool man. I'm sure there'll be a great POC out of it and you will demo something is really interesting but you know they.Unless in my opinion the business really starts digital transformation really means something and it's not just a moniker for you know I want a cooler Web site. We're not going to get from here to there. 

But by the way once that starts to happen and I keep pointing people the automotive industry you want to see how transition is happening. And I know automotive I don't like to use because it has a consumer angle to it. And I'm not very interested in why you know toasters should talk to coffee makers. But I think automobiles sits at an intersection because it's a fairly complex machine. And you know you're saying it in front of us right now. And the evolution is to the point realize this because it's already happened in the hardware in my business. Once a guy buys the service meaning an instance at Amazon or Azure or Google or Alibaba or whatever. The idea that that's a Lenovo server or a HP server is incidental. I don't even care. And if you didn't realize this those guys do not buy. Because from the traditional hardware suppliers. Because there's no reason to they know exactly what the workload is on these machines they can spec it out and they show up over at Foxconn and they go guys get me a server like this. Well guess what. DiDi can do the same thing or Uber Lyft or whatever. They go. Once you create automation once you know the the model of how this is being done. How it's being used. You just go to Geely and go guys build me a machine that looks like this. And now what does it mean to be Ford or Porsche or Mercedes. Right. You're out of the loop. I mean that's why this is so you know it should be so disconcerting for those guys to realize that that that pattern could happen here as well. And now insert every other kind of agriculture construction transportation etc. machine. Which by the way is going to even be transformed for the guys are building anything that moves. I mean the whole electric battery world is going to transform anything that moves. So. Mean all of those industries either. And you asked an interesting I interpret as a different way of looking at those is is the future I'm talking about a future where. A Tesla of tractors is created. You know a brand new company whose three Fudan university kids go out and build the next generation blood analyzer the next generation. Roll Gater that's software defined uses you know 3D printing technologies you know motors batteries et cetera. Is that what happens or is it. Right the. Traditional companies that morph themselves over here. I don't know the answer to that question. Which one of these is going to happen. But I think the inevitability of it is there. We can only discuss timeframes.

I don't know. I don't know what the time I mean if I need the time frames I probably wouldn't be doing the podcast but in New York. Yeah but this isn't as inevitable and I wrote a book called The End of Software.

Now about 14 years ago basically outlining what would happen in the software business. I identified four young startups. Sales Force,VMWare,Netsuite and open harbor.

And I know you've never heard of. But three others. If you bought the book and bought the stock you would be talking to me. And.

I just I see. To me it just looks like this all over again is just on a much more massive scale that we're discussing this as it talks about every piece of infrastructure out there.

 So let's maybe as a last point which has to end on a success case I just came across Mahindra Mahindra. India has launched a company called Tringo which is basically the Uber for agriculture equipment it's kind of a startup really cool model. I think it really fits India because they have this lack of infrastructure. They need a good way for a small farmer to rent you know a tractor for three hours to get a job then. Maybe as a last point then is there a large company let's say large traditional business that you've seen successfully make this transmission transition and not just launched some POCs but really seriously.

Change how their business is structured in order to evolve. I end up in the world of making machines. Yeah exactly. I think your your your most advanced case study that anybody would actually think about is Tesla. And it is you know as they evolve is my only examples.

Everybody else were still early. I'd go back to World where early you know you see you see little cases here and there I mean I'll give it you know Atco I know in sugarcane manufacturing and there's case in the book. They they start to evolve. You can buy their  per bushel. Right. They certainly have you know invested into. The whole connectivities size etc.. But to say that at completely transformed their business I would never argue that.

Right. Right. You could you could argue that if Tesla hadn't. Taken this very ambitious strategy the rest of the automotive industry wouldn't be where it is today in the midst of a transformation right there. They are to an extent following in their folly because Tesla's now the largest market cap automotive so they have well are being assaulted from multiple directions.

I think there's the Tesla direction which is the machine itself and electricity obviously and then there's the DiDi Uber thing. DiDi Uber so-called machines as a service right.

And the evolution of autonomy. Google doing. I mean Google's doing a Didi. I mean they're all playing autonomy. I could run machine a service on a gasoline engine right. But you take all you take all this together. And that's why life's not going to be easy for you know General Motors pick on somebody else right.

Cool Tim super interesting conversation. Thanks for taking the time with us today.

Well it was fun. How can people learn more about you. Your book your work.

Well first of all the book precision available on Amazon. You know available on Kindle in English and then in Chinese because I know you're in Shanghai. We just launched Tsinghua University Press. I don't have the URL I mean at some point in time I'm sure you could make it available to the listeners that I know it's going to be available at jd.com and at amazon.cn whatever combat's and either relatively certain whatever side recommend that we're getting ready actually to launch a whole class which is based on the book and it'll be in the iTunes store. Precision again kind of a. TED Talk size lectures for people who want to listen and watch more than they want to read. So I you I obviously it's self-serving but I think Precision is a good background for people because we are particularly non-technical people to understand fundamentally the technology. So we talk about the same thing the machine. You talk about how it's connected and how many different ways that might be true. How do you collect data from the machine? What can you do to learn on that data?

Everything from very simple learning all the way to you know using advanced deep learning machine learning technologies as is available from a company that I've cofounded with some of my Stanford students called Lecida.

And then finally you know once you learn what can you do differently.

We already kind of discussed much of what that means for a business model perspective but I think it also can mean something from a software perspective. So I and then we we walk through that technologically both.

What I call principles and then practices meaning let me show you this and practice with a bunch of examples and then the back half of the book is 15 different case studies kind of tell you where you are today so there's an example from Actos combine harvester is a chapter there's a chapter about Nick August who is a farmer in the Cotswolds. And you know how he does this. There's chapters on oil and gas water construction equipment etc. to give you or give the reader I think a good. Understanding of kind of where we are today and what people are doing today. So it's a little bit more rooted. I know you already said it. We do tend to do they. Oh this is going to change the world. I started there right two thirds of global GDP but the book I wanted to give people a little bit of an understanding of kind of where we are. And obviously then what's the gap between where we are and where we could be. So I recommend people pick it up.

Over get we'll get that new show nuts didn't have a great evening. Thanks again for. Taking the time to talk with us today.