>> Announcer: Live from New York City. It's theCUBE. Covering CyberConnect 2017. Brought to you by Centrify, and the Institute for Critical Infrastructure Technology. >> Hey, welcome back everyone. Live here with Cube coverage in New York City, our favorite place to be when we've got all the action going on. CyberConnect 2017 is an inaugural event where industry, government comes together to solve the crisis of our generation. That's cybersecurity. I'm John Furrier, co-host theCube My partner Dave Vellante here. Our next guest is Chris Novak, VTRAC Global Director, Threat Research Advisory Center at Verizon. Welcome to theCube, great to have you. >> Thanks, pleasure to be here. >> So you do all the homework. You've got the forensic data. You're the one looks at the threats. You're the burning bush of cyber intelligence. What's happening? Tell us what's the threats? >> Everything. So, it's interesting because I always find what I do to be wildly exciting just because it's always changing, right? Everything we see. It's kind of' like being a cop. Ultimately you're investigating unknowns all the time, trying to figure out how they happen, why they happen, who they happen to, but more importantly than that, how do you get ahead of it to prevent being the next one, or prevent it happening to others? And that's really the thrust of what we're out to do. >> Talk about the challenges 'cause General Keith Alexander was on stage talking about how he compared it to an airline crashing, where they come in looking for the black box, and it's worse because you don't even know what happened, who was involved. >> Chris: That's right. >> The notion of anonymous, public domain software is causing all kinds of democratization, good and bad, bad being actors that we don't even know attacking us. What is the landscape of how you identify what's going on? >> Yeah, and it gets even more challenging than that because I like that analogy, and I'd say I'd almost take it one step further and say the analogy of the airline and looking for the black box. In many cases when we go in to do an investigation, we're just hoping that there was a black box to look at to begin with. In many cases, we get there and there was no information, and we're trying to take all the pieces and put it together of what's left. And ultimately what we see is, it keeps evolving, right? It keeps getting harder, and the threat actors keep getting better. What I always tell folks is, while many of us all have to play by a set of rules, or regulations, or compliance obligations, the threat actors don't have to do any of that. They're free to do whatever works for them, and repeat it over and over again, and, for them, it's a business. >> So Dave and I were talking earlier. I want to get your reaction to this. About the importance of Stuxnet. Ars Technica has a report coming out that certificate authorities were compromised well before Stuxnet. But Stuxnet is the Pearl Harbor, cyber Pearl Harbor, as a point in time. So much has happened since then. So from that kind of Pearl Harbor moment of the wakening of, oh my God, to today, what's the landscape look like? How important was the Stuxnet to that point in time now, and how has it evolved? What's changed? >> Sure, and I think a couple of key things that come out of that. One is, you start to see more and more attribution to government-related attacks. Some are actively sponsored and known. Some are, we're just diggin' through the details and the weeds to try and figure out who's actually behind it and attribution may never actually take place. >> Or it could not be real 'cause they want to blame their enemy so that they get attacked. >> Well, and that's the either beauty or downside of cyber is that you can conduct it in a vacuum, in an anonymous fashion. So, in many respects, you can conduct an attack remotely and try to give it all the hallmarks of someone else, making it further difficult to attribute it. >> And the tools are now available too, so like, I hear reports that states are sponsoring, or releasing in the public domain, awesome hacks, like Stuxnet of the future, which some say was released and then got out of control by accident. >> And that's always something you have to be concerned about is the fact that once this stuff gets out there, even if you only intended to use this malware or attack vector once. Once you use it on that victim, there is a potential that that spreads. >> But you guys have been doing this study for the last decade. >> Correct. >> So you've seen the shift from sort of hacktivist to nation-sponsored malware. What has the research shown you over the last decade as that shift has occurred? >> Yeah, it's interesting because you look at it and a lot of what we still see today are financially-motivated and interestingly enough, opportunistic, low-hanging fruit kind of attacks. About 70 to 80% fall in that category, and about 20 to 25, depending on the year, are nation state, but that keeps growing each year. And, I think a lot of it is. >> John: What the nation state piece? >> The nation state piece. But it's still the smaller piece of the pie or the graph, whatever you're looking at, because, at the end of the day >> It's cash. >> It's cash. >> They want the cash. >> And so much of what we find when you look back at the old days of breaches where the majority of them were, they weren't even really breaches of theft of data, it was someone. >> Confetti, graffiti. >> I should have actually asked that question differently because it's really went from hacktivist to criminals. >> Chris: Correct. >> To nation states and you're saying the dominant now is criminal activity. >> That's correct. Yeah, we find the large piece of it about more than half is organized crime. It comes down to, look, you can steal money in a variety of different ways. This is a way to do it safely from thousand miles away >> And no one knows who you are. >> on the other end of a keyboard. >> So it's annoyance. >> And by the way, no consequence. Who's going to? >> Virtually, yeah. >> What court do you go to? >> So its annoyance is the hacktivist. Okay, we can kind of' live with that. It's cash and it's threats to critical infrastructure. >> And we see kind of a graduation there where you see the activists realize, I can this and make a point, but a point doesn't necessarily make me money, or I can do this for an organized crime group and make millions of dollars. Hmmmm. >> And, by the way, to your point which we were just teasing out, Dave. There is zero downside, because if you get caught, what happens? >> Yeah. >> If you get caught. >> If you get caught, yeah. And then what happens if you get caught? >> There's no jurisdiction. >> You don't make money. >> No, no, there's no courts. >> It's very hard to prosecute. >> There's actually no process for that. >> So, we heard this morning that WannaCry and other examples of malware really weren't about malware. I mean, sorry, they really weren't about ransomware, they were about sending a message, or politics. So, you're obviously seeing more of that in your research. >> Chris: Exactly right. >> Fake news, and I wonder if you could comment. >> Absolutely, yeah. So, in fact, it was interesting because some of those had continued to come out. Everyone kept thinking that it was all ransomware, and then as we studied it further we found some of these, they never had the intention of collecting a ransom, or giving the data back. It was all about making a political point, and you now have this kind of injection of politics into something that was really, traditionally, just organized crime, smash and grab, make cash. Now politics is feeding into that, going, wait, we can affect and influence and all sorts of things in ways people have never imagined and people don't even know it's going on. >> So you must be seeing a dramatic improvement in the quality, hate to say this, but the quality of malware, over the last decade. Less bugs, less errors, >> More sophisticated. >> More insidious, sophisticated. >> That's exactly right >> Vectors. >> We do see that continuing to improve and for them, like I always tell folks, they operate it like a business. You'll have some of these groups where they'll have different divisions or departments. People will have clearly-defined roles and responsibilities of what they're supposed to be doing in generating that malware, troubleshooting it, and they'll even reward people for how well it works. >> Chris, I'd like to get your personal opinion. If you could put your Verizon hat on too, I will take any opinions that you have. How do we solve this? 'Cause this event here. We like this inaugural event because it's the first industry event that talks about the big picture, the holistic view, the 20-mile stare, if you want to' say it that way. Not the Black Hat, which has its own conference, and there should be more of that. This is industry coming together. Governments now intersecting here. What's your opinion on how this gets solved. We heard community, shared data, that's been going around. What do you think? >> So, that's probably the hardest question I get asked, and, honestly, I think it's because there's not really a simple answer to it, right? It's like saying, how do we stop crime? We don't. It's not going to be possible. It's a matter of, how do we put up better defenses? And also, important, how do we put up better detection, so that we can see things and, potentially, stop them sooner before they blow up into these big, multi-hundred-million record, or billion record breaches? So, one of the biggest things that I advocate is awareness. We also have to do things like pro-active threat hunting, right? If you're not out there. It's kind of like having security guards, right? You go through any office and you've got security guards walking the halls, sitting in the lobby, looking for things that are unusual. If we're not out there in the cyber realm looking for unusual things, you can't expect that you're going to see them until they've reached a certain blow-up point. >> Or are they cloaked? Completely cloaked. You can't see 'em. >> That's also true. >> Security guards are looking for someone they can't see. >> That's true. >> Chris, thanks so much for coming here and sharing the opinion. Follow the research. And your report's public, or? >> Yes, the reports are all available on the VerizonEnterprise.com website. >> Okay, VerizonEnterprise.com. Check it out. These reports are a treasure trove of information. Always getting it out. Thanks for your perspective. Lookin' for more trends. Chris Novak here inside theCube here in New York City's live coverage of CyberConnect 2017. I'm John with Dave Vallente. We're back with more coverage after this short break. (techno music)

>>Hello everybody. I am Toshihiko Nishimura from Stanford. University is there to TTT out here, super aging, global OMIM global transportation group about infections, uh, or major point of concerns. In addition, this year, we have the COVID-19 pandemic. As you can see here, while the why the new COVID-19 patients are still increasing, meanwhile, case count per day in the United state, uh, beginning to decrease this pandemic has changed our daily life to digital transformation. Even today, the micro segmentation is being conducted online and doctor and the nurse care, uh, now increase to telemedicine. Likewise, the drug development process is in need of major change paradigm shift, especially in vaccine in drug development for COVID-19 is, should be safe, effective, and faster >>In the >>Anastasia department, which is the biggest department in school of medicine. We have Stanford, a love for drug device development, regulatory science. So cold. Say the DDT RDS chairman is Ron Paul and this love leaderships are long mysel and stable shaper. In the drug development. We have three major pains, one exceedingly long duration that just 20 years huge budget, very low success rate general overview in the drug development. There are Discoverly but clinical clinical stage, as you see here, Tang. Yes. In clinical stage where we sit, say, what are the programs in D D D R S in each stages or mix program? Single cell programs, big data machine learning, deep learning, AI mathematics, statistics programs, humanized animal, the program SNS program engineering program. And we have annual symposium. Today's the, my talk, I do like to explain limitation of my science significance of humanized. My science out of separate out a program. I focused on humanized program. I believe this program is potent game changer for drug development mouse. When we think of animal experiment, many people think of immediately mouse. We have more than 30 kinds of inbred while the type such as chief 57, black KK yarrow, barber C white and so on using QA QC defined. Why did the type mice 18 of them gave him only one intervention using mouse, genomics analyzed, computational genetics. And then we succeeded to pick up fish one single gene in a week. >>We have another category of gene manipulated, mice transgenic, no clout, no Kamal's group. So far registered 40,000 kind as over today. Pretty critical requirement. Wrong FDA PMDA negative three sites are based on arteries. Two kinds of animal models, showing safety efficacy, combination of two animals and motel our mouse and the swine mouse and non-human primate. And so on mouse. Oh, Barry popular. Why? Because mouse are small enough, easy to handle big database we had and cost effective. However, it calls that low success rate. Why >>It, this issue speculation, low success rate came from a gap between preclinical the POC and the POC couldn't stay. Father divided into phase one. Phase two has the city FDA unsolved to our question. Speculation in nature biology using 7,372 new submissions, they found a 68 significant cradle out crazy too, to study approved by the process. And in total 90 per cent Radia in the clinical stages. What we can surmise from this study, FDA confirmed is that the big discrepancy between POC and clinical POC in another ward, any amount of data well, Ms. Representative for human, this nature bio report impacted our work significantly. >>What is a solution for this discrepancy? FDA standards require the people data from two species. One species is usually mice, but if the reported 90% in a preclinical data, then huge discrepancy between pretty critical POC in clinical POC. Our interpretation is data from mice, sometime representative, actually mice, and the humor of different especially immune system and the diva mice liver enzyme are missing, which human Liba has. This is one huge issue to be taught to overcome this problem. We started humanized mice program. What kind of human animals? We created one humanized, immune mice. The other is human eyes, DBA, mice. What is the definition of a humanized mice? They should have human gene or human cells or human tissues or human organs. Well, let me share one preclinical stages. Example of a humanized mouse that is polio receptor mice. This problem led by who was my mentor? Polio virus. Well, polio virus vaccine usually required no human primate to test in 13 years, collaboration with the FDA w H O polio eradication program. Finally FDA well as w H O R Purdue due to the place no human primate test to transgenic PVL. This is three. Our principle led by loss around the botch >>To move before this humanized mouse program, we need two other bonds donut outside your science, as well as the CPN mouse science >>human hormone, like GM CSF, Whoah, GCSF producing or human cytokine. those producing emoji mice are required in the long run. Two maintain human cells in their body under generation here, South the generation here, Dr. already created more than 100 kinds based on Z. The 100 kinds of Noe mice, we succeeded to create the human immune mice led the blood. The cell quite about the cell platelets are beautifully constituted in an mice, human and rebar MAs also succeeded to create using deparent human base. We have AGN diva, humanized mouse, American African human nine-thirty by mice co-case kitchen, humanized mice. These are Hennessy humanized, the immune and rebar model. On the other hand, we created disease rebar human either must to one example, congenital Liba disease, our guidance Schindel on patient model. >>The other model, we have infectious DDS and Waddell council Modell and GVH Modell. And so on creature stage or phase can a human itemize apply. Our objective is any stage. Any phase would be to, to propose. We propose experiment, pose a compound, which showed a huge discrepancy between. If Y you show the huge discrepancy, if Y is lucrative analog and the potent anti hepatitis B candidate in that predict clinical stage, it didn't show any toxicity in mice got dark and no human primate. On the other hand, weighing into clinical stage and crazy to October 15, salvage, five of people died and other 10 the show to very severe condition. >>Is that the reason why Nicole traditional the mice model is that throughout this, another mice Modell did not predict this severe side outcome. Why Zack humanized mouse, the Debar Modell demonstrate itself? Yes. Within few days that chemistry data and the puzzle physiology data phase two and phase the city requires huge number of a human subject. For example, COVID-19 vaccine development by Pfizer, AstraZeneca Moderna today, they are sample size are Southeast thousand vaccine development for COVID-19. She Novak UConn in China books for the us Erica Jones on the Johnson in unite United Kingdom. Well, there are now no box us Osaka Osaka, university hundred Japan. They are already in phase two industry discovery and predict clinical and regulatory stage foster in-app. However, clinical stage is a studious role because that phases required hugely number or the human subject 9,000 to 30,000. Even my conclusion, a humanized mouse model shortens the duration of drug development humanize, and most Isabel, uh, can be increase the success rate of drug development. Thank you for Ron Paul and to Steven YALI pelt at Stanford and and his team and or other colleagues. Thank you for listening.

>>long from Las Vegas. It's the Q covering a ws re invent 2019. Brought to you by Amazon Web service is and in along with its ecosystem partners. >>Hey, welcome back to the Cube. Lisa Martin at AWS Reinvent 19. This is Day three of the Cubes coverage. We have two sets here. Lots of cute content are joined by Justin Warren, the founder and chief analyst at Pivot nine. Justin. How's it going? Great, right? You still have a voice? Three days? >>Just barely. I've been I've been trying to take care of it. >>Impressed. And you probably have talked to at least half of the 65,000 attendees. >>I'm trying to talk to as many as I can. >>Well, we're gonna talk to another guy here. Joining us from data ICU is well, Novak, the solutions architect will be the Cube. >>Thanks for having me. >>You have a good voice too. After a three day is that you >>have been doing the best I can. >>Yeah, he's good. So did ICU. Interesting name. Let's start off by sharing with our audience. Who did a coup is and what you guys do in technology. >>Yes. So the Entomology of date ICU. It's like hi cooze for data. So we say we take your data and, you know, we make poetry out of it. Make your data so beautiful. Wow, Now, But for those who are unaware Day like it was an enterprise data science platform. Eso we provide a collaborative environment for we say coders and clickers kind of business analyst and native data scientists to make use of organizations, data bill reports and Bill productive machine learning base models and deploy them. >>I'm only the guy's been around around for eight years. Eight years. Okay, >>so start up. Still >>mourning the cloud, the opportunity there That data is no longer a liability. It's an asset or should be. >>So we've been server based from the start, which is one of our differentiators. And so by that we see ourselves as a collaborative platform. Users access it through a Web browser, log into a shared space and share code, can share visual recipes, as we call them to prepare data. >>Okay, so what customers using the platform to do with machine learning is pretty hot at the moment. I think it might be nearing the peak of the life cycle pretty hot. Yeah, what a customer is actually actually doing on the platform, >>you know, So we really focus on enabling the enterprise. So, for example, G has been a customer for some time now, and Sergey is a great prototypical example on that. They have many disparate use cases, like simple things like doing customer segmentation for, you know, marketing campaigns but also stuff like Coyote predicted maintenance. So use cases kind of run the gamut, and so did ICU. Based on open source, we're enabling all of G's users to come into a centralized platform, access their data manipulated for whatever purposes. Maybe >>nobody talked about marketing campaigns for a second. I'm wondering. Are, is their integration with serum technologies? Or how would a customer like wanting to understand customer segmentation or had a segment it for marketing campaign? How would they work in conjunction with a serum and data ICU, for example? >>It's a great question. So again, us being a platform way sit on a single server, something like an Amazon ec2 instance, and then we make connections into an organization's data sources. So if using something like Salesforce weaken seamlessly, pull in data from Salesforce Yuka manipulated in date ICU, but the same time. Maybe also have some excel file someone you know me. I can bring that into my data to work environment. And I also have a red shift data table. All those things would come into the same environment. I can visualize. I can analyze, and I can prepare the data. I see. >>So you tell you it's based on open source? I'm a longtime fan of over. It's always been involved in it for longer than I care to remember. Actually, that's an interesting way t base your product on that. So maybe talk us through how you how you came to found the company based on basic an open source. What? What led to that choice? What? What was that decision based on? >>Yeah, for sure. So you talked about how you know the hype cycle? A. I saw how hot is a I and so I think again, our founders astutely recognize that this is a very fast moving place to be. And so I'm kind of betting on one particular technology can be risky. So instead, by being a platform, we say, like sequel has been the data transformation language do jour for many days now. So, of course, that you can easily write Sequel and a lot of our visual data Transformations are based on the sequel language, but also something like Python again. It's like the language de jour for machine law machine learning model building right now, so you can easily code in python. Maintain your python libraries in date, ICU And so by leveraging open source, we figured we're making our clients more future proof as long as they're staying in date ICU. But using data ICU to leverage the best in breed and open source, they'll always be kind of where they want to be in the technological landscape by supposed to locked into some tech that is now out of date. >>What's been the appetite for making data beautiful for a legacy enterprise, like a G E that's been around for a very long time versus a more modern either. Born in the Cloud er's our CEO says, reborn in the cloud. What are some of the differences but also similarities that you see in terms of we have to be able to use emerging tech. Otherwise someone's gonna come in behind us and replace us. >>Yeah, I mean, I think it's complicated in that there's still a lot of value to be had in someone says, like a bar chart you can rely on right, So it's maybe not sexy. But having good reporting and analytics is something that both you know, 200 year old enterprise organizations and data native organizations startups needs. At the same time, building predicted machine learning models and deploying those is rest a p i n points that developers can use in your organization to provide a data driven product for your consumers. Like that's amore advanced use case that everyone kind of wants to be a part of again data. Who's a nice tool, which says Maybe you don't have developers who are very fluent in turning out flashed applications. We could give you a place to build a predictive model and deploy that predictive model, saving you time to write all that code on the back end. >>One of the themes of the show has been transformation, so it sounds like data ICU would be It's something that you can dip your toes in and start to get used to using. Even if you're not particularly familiar with Time machine learning model a model building. >>Yeah, that's exactly right. So a big part of our product and encourage watchers to go try it out themselves and go to our website. Download a free version pretrial, but is enablement. So if you're the most sophisticated applied math PhD there is, like, Who's a great environment for you to Code and Bill predictive models. If you never built the machine learning model before you can use data ICU to run visual machine learning recipes, we call them, and also we give you documentation, which is, Hey, this is a random forest model. What is a random forest model? We'll tell you a little bit about it. And that's another thing that some of these enterprises have really appreciated about date I could. It is helping up skill there user base >>in terms of that transformation theme that Justin just mention which we're hearing a lot about, not visit this show. It's a big thing, but we hear it all the time, right? But in terms of customers transformation, journey, whatever you wanna call it, cloud is gonna be an essential enabler of being able to really love it value from a I. So I'm just wondering from a strategic positioning standpoint. Is did ICU positioned as a facilitator or as fuel for a cloud transformation that on enterprise would undergo >>again? Yes, great point. So for us, I can't take the credit. This credit goes to our founders, but we've thought from the start the clouds and exciting proposition Not everyone is. They're still in 2019. Most people, if not all of them, want to get there. Also, people want too many of our clients want the multi cloud on a day. Like who says, If you want to be on prim, if you want to be in a single cloud subscription. If you want to be multi cloud again as a platform, we're just gonna give you connection to your underlying infrastructure. You could use the infrastructure that you like and just use our front end to help your analyst get value. They can. I >>think I think a lot of vendors across the entire ecosystem around to say the customer choice is really important, and the customers, particularly enterprise customers, want to be able to have lots of different options, and not all of them will be ready to go completely. All in on cloud today. They made it may take them years, possibly decades, to get there. So having that choice is like it's something that it would work with you today and we'll work with you tomorrow, depending on what choices you make. >>It's exactly right. Another thing we've seen a lot of to that day, like who helps with and whether it's like you or other tools. Like, of course, you want best in breed, but you also want particularly for a large enterprise. You don't want people operating kind of in a wild West, particularly in like the ML data science space. So you know we integrate with Jupiter notebooks, but some of our clients come to us initially. Just have I won't say rogues that has a negative connotation. But maybe I will say Road road data Scientists are just tapping into some day the store. They're using Jupiter notebooks to build a predictive model, but then to actually production allies that to get sustainable value out of it like it's to one off and so having a centralized platform like date ICU, where you can say this is where we're going to use our central model depository, that something where businesses like they can sleep easier at night because they know where is my ML development happening? It's happening in one ecosystem. What tools that happening with, well, best in breed of open source. So again, you kind of get best of both worlds like they like you. >>It sounds like it's more about the operations of machine learning. It is really, really important rather than just. It's the pure technology. Yes, that's important as well, and you need to have the data Sinus to build it, but having something that allows you to operationalize it so that you can just bake it into what we do every day as a business. >>Yeah, I think in a conference like this all about tech, it's easy to forget what we firmly believe, which is a I and maybe tech. More broadly, it's still human problems at the core, right? Once you get the tech right, the code runs corrected. The code is written correctly. Therefore, like human interactions, project management model deployment in an organization. These are really hard, human centered problems, but so having tech that enables that human centric collaboration helps with that, we find >>Let's talk about some of the things that we can't ever go to an event and not talk about. Nut is respected data quality, reliability and security. Understood? I could facilitate those three cornerstones. >>Yeah, sure. So, again, viewers, I would encourage you to check out the date. ICU has some nice visual indications of data quality. So an analyst or data scientists and come in very easily understand, you know, is this quality to conform to the standards that my organization has set and what I mean by standards that could be configured. Right? So does this column have the appropriate schema? Does it have the appropriate carnality? These are things that an individual might decide to use on then for security. So Data has its own security mechanisms. However, we also to this point about incorporating best Retek. We'll work with whatever underlying security mechanisms organizations organizations have in place. So, for instance, if you're using a W s, you have, I am rolls to manage your security. Did ICU comport those that apply those to the date ICU environment or using something like on prime miss, uh, duke waken you something like Kerberos has the technology to again manage access to resources. So we're taking the best in breed that this organization already has invested time, energy and resources into and saying We're not trying to compete with them but rather were trying to enable organizations to use these technologies efficiently. >>Yeah, I like that consistency of customer choice. We spoke about that just before. I'm seeing that here with their choices around. Well, if you're on this particular platform will integrate with whatever the tools are there. People underestimate how important that is for enterprises, that it has to be ahead. Virginia's environment, playing well with others is actually quite important. >>Yeah, I don't know that point. Like the combination of heterogeneity but also uniformity. It's a hard balance to strike, and I think it's really important, giving someone a unified environment but still choice. At the same time. A good restaurant or something like you won't be able to pick your dish, but you want to know that the entire quality is high. And so having that consistent ecosystem, I think, really helps >>what are, in your opinion, some of the next industries that you see there really right to start Really leveraging machine learning to transfer You mentioned g e a very old legacy business. If we think of you know what happened with the ride hailing industry uber, for example, or fitness with Saletan or pinchers with visible Serge, what do you think is the next industry? That's like you guys taking advantage of machine learning will completely transform this and our lives. >>I mean, the easy answer that I'll give because it's easy to say it's gonna transform. But hard to operationalize is health care, right? So there is structured data, but the data quality is so desperate and had a row genius s, I think you know, if organizations in a lot of this again it's a human centered problem. If people could decide on data standards and also data privacy is, of course, a huge issue. We talked about data security internally, but also as a customer. What day to do I want you know, this hospital, this health care provider, to have access to that human issues we have to result but conditional on that being resolved that staring out a way to anonymous eyes data and respect data privacy but have consistent data structure. And we could say, Hey, let's really set these a I M L models loose and figure out things like personalized medicine which were starting to get to. But I feel like there's still a lot of room to go. That >>sounds like it's exciting time to be in machine learning. People should definitely check out products such as Dead Rock you and see what happens. >>Last question for you is so much news has come out in the last three days. It's mind boggling sum of the takeaways, that of some of the things that you've heard from Andy Jassy to border This'll Morning. >>Yeah, I think a big thing for me, which was something for me before this week. But it's always nice to hear an Amazon reassures the concept of white box. Aye, aye. We've been talking about that a date ICU for some time, but everyone wants performance A. I R ml solutions, but increasing. There's a really appetite publicly for interpret ability, and so you have to be responsible. You have to have interpret belay I and so it's nice to hear a leader like Amazon echo that day like you. That's something we've been talking about since our start. >>A little bit validating them for data ICU, for sure, for sure. Well, thank you for joining. Just to be on the kid, the suffering. And we appreciate it. Appreciate it. All right. For my co host, Justin Warren, I'm Lisa Martin and your work to the Cube from Vegas. It's AWS reinvent 19.