Futurum Group Foresees Cybersecurity Futures || Tech Field Day News Rundown: May 14, 2025

Building in AI kill switches, Broadcom's, checking up on some of their customers. Nutanix adds some support for Nvidia Cisco's quantum aspirations. There's some uncertainty in the IT market, pseudo or PS don't.

That is the question, and we're gonna take a look at the latest futureum group security research in this episode of the Tech Field Day rundown. Hello everyone, and welcome to the Tech Field Day rundown for Wednesday, may the 14th. My name is Tom Hollingsworth, and I'm very happy to be back with you and, uh, I hope you're hungry because it is national buttermilk biscuit day and there's, that's really the most superior form of biscuit unless you're one of those weirdos that lives in England and you think that a biscuit is a sweet little cookie.

Um, but speaking of w weirdos that don't live in England anymore, I'd like to introduce my cohost, Mr. Alistair Cook. Al, welcome to the show.

It is a pleasure to be here. And the thing that you call a biscuit, we call aco and, uh, yeah, buttermilk scones would be the English course scone. I'm in New Zealand with a scorn here, mate.

Uh, buttermilk Scon sounds good. And of course, it's also national decency day, so I hope everybody is wearing trousers today. Well, the only decent way to eat a buttermilk biscuit, of course, is covered in gravy, thick, delicious gravy.

Just like the delicious stories that we have coming up for you this week. There's been some AI news and some quantum news and yeah, some security news. So let's just dive right into it.

Export restrictions to keep high-end Nvidia chips from helping China build global domination don't seem to be working. Congressman Bill Foster wants to make Nvidia build in a kill switch that will enable any chip that lands in the wrong place to be automatically disabled, either by some US government authority or pretty much whenever the AI feels like killing it. Now, I don't know about you Al, but I'm trying to figure out what could possibly go wrong with a government mandated kill switch for an advanced technology.

Well, it seems like a, an echo of every other proposal that comes out of government, that some technology needs to be controlled, and so we'll just pass a law and magically things will be controlled. And that was my first feeling on this, was this must just be somebody in, in government with no clue. But it turns out that Congressman Foster actually has a background in, uh, chip design and so has some basis for actually being able to think about whether this is practical or not.

Uh, the idea here is that the, uh, export restrictions aren't working as well as you might hope, and that trans shipping through, uh, mutually friendly locations and, uh, just generally dodgy trading is enabling these Nvidia GPUs to get to places that the US would like them not to get to. Uh, China is the, the most often cited, but it's not the only location that is restricted for these things and is yet, uh, still able to, to, uh, to get these things in. So a couple of elements in here.

One we covered on the rundown before, which is that, uh, Huawei is building their own, uh, GPUs that are aiming to be competitive with the, the latest of the Nvidia chips. And so this restriction may not actually be even beneficial to the US in restricting access to advanced technology. The Chinese scientists are building it.

Uh, but I think the interesting part is this, this restriction being more about being able to disable the chip once it turns up in the wrong place. And, uh, congresswoman foster's ideas are not completely set in stone here. There was some geolocation suggestions about it.

So if it's, if the chip turns up and they presumably GPS identified, uh, location, uh, or possibly that it, it phones home because Nvidia chips are rather inclined to phone home and, uh, check for licensing, or if that phone home comes from a place that it shouldn't, of course that could never be spoofed, could it, uh, then the kill switch could be activated. Seems a little, uh, farfetched still that this kill switch could be used. And then you've also gotta consider whether exporting these chips to friendly nations might also be problematic if there's a kill switch capability.

The United States could disable a currently friendly nation of Denmark or Canada, uh, and disable their advanced AI tools because they had some, uh, designs on, uh, let's leave that alone. What we do see is NVIDIA's not a great fan of this. This would add additional complexity to their chips that involve far more government oversight of the actual silicon that's going on in their ships.

And just the idea of the government, um, being able to turn stuff off at will for whatever reason, some government official be they highly placed or lowly placed, uh, chooses to switch off. It just feels a bit dangerous. And, uh, yeah, not really a free market kind of solution.

Broadcom Broadcom's transitioned all of the new sales upgrades and support renewals of the VMware products to a subscription basis. In order to get that continuing return every month or every, every year, some customers have chosen not to take these upgrades and to sit on their existing perpetually licensed VMware products. And this has led to VMware well, Broadcom, uh, Broadcom legal sending letters to these customers saying, you know, when your support agreement expired, your rights to download and install our patches expired.

The letter suggests either purchasing a sub subscription, clearly the preferred option, or removing the patches that were released after the subscription renewal, uh, expired. Uh, having seen a very large number of, uh, people at the Broadcom legal all hands meeting at a hotel I was at recently, um, just Broadcom really need to get all of these lawyers involved. Tom, That depends on who at Broadcom thinks that they're going to get companies to pay up on this.

Now, in full defense of Broadcom here, if the contract states that you are not allowed to download any patches that are not critical security updates and you install a service pack or any of those other things that increases functionality or features or things like that, yet you did break the terms of the contract. So there's that. I think though, that they might be better off with the carrot approach instead of the stick approach here.

And the reason for that is because the customers may not understand that they don't have the rights to do that. They may think that this is like a Windows patch or something like that, that if it's out there and everybody else is installing it, I should be able to install it too. And again, it comes back to the way that the contract is worded.

If the contract says you can't install anything at all, then that's the way that it's supposed to go. The question that I have for the people at Broadcom is, do you think that this approach is going to make people renew their support contracts? Do you think they're gonna renew their support contracts at the levels that you want them to renew their support contracts at?

Those two things are what you need to answer before you really pursue this. As near as I can tell, this is only affected a few companies so far, and I know that the r story about it really kind of said, you know, they're sending out these cease and desist letters to companies who they're, their things have been out of, uh, support for like six days. Fair enough.

That's, that's probably a little quick. Um, maybe they're thinking that they can use this as a way to kind of, Hey, you really do need to, to re-up your support contract. But I also think that the other solution honestly, would be to put a, basically a, a wall in front of all their patches and say, you're only allowed to update these if you have a valid support contract.

You know, maybe you have the update system check for valid, uh, contract status before you allow them to upload patches. Would it be messy? Yeah, but you wouldn't have to be legally messy.

You would just have to be support contract messy. I don't know that there is a good solution for this right now, unless there is some kind of amnesty set up where it's like, listen, if you guys will agree to sign a support contract for the next six months or year or what have you, that will allow you to do this. Maybe that's a way out for customers that wanna stay on VMware.

I, I worry though that the solution is if we're gonna send you cease and desist letters, that's gonna force people to say, well, if I can't install any security updates or software, then I guess I'm done using this and I'm gonna migrate now. And that ultimately may be a bad turn for the people at Broadcom. Recently at Nutanix.

Next, the company Nutanix announced that they are gonna be adding support for Nvidia, NIM and NEMO frameworks to Nutanix Enterprise ai. Adding AI and NVIDIA to any product will make it 10 times more fashionable and probably 20 times more expensive in the long run. But I guess the question that I have for you, Al, is the alphabet soup and the announcement.

Does that mean there's something substantial behind all of this AI stuff, or is it just kind of marketing? We heard about the Nutanix Enterprise AI solution in AI at, uh, AI infrastructure Field Day a couple of weeks ago. And the aim here is to simplify getting a deployment of an AI solution, particularly on premises, because as the futureum group studies have have shown, there's not a really high success rate to fully DIY building your own AI solution.

Um, about 60% fail to scale beyond the pilot stage. Uh, primarily because the return on investment isn't seen there, A huge amount of effort isn't actually returning a lot of business, uh, value. So reducing the amount of effort that's required to build these solutions and to then operate these solutions is, is important.

And this is where the NVIDIA frameworks step in. They're largely industry standard for building solutions on, uh, Nvidia GPUs for ai. Uh, NIM is the inference micro, uh, services architecture.

So this is the architecture for deploying out, uh, an AI solution both for, uh, a little bit of fine tuning and primarily for inference for the, for the, uh, usually retrieval augmented generation type inference. Um, this is definitely a, an architectural framework that makes it easier to build AI applications. I don't think it's actually easy, but it's, uh, it makes it easier to get through those cycles of building application.

I particularly like the NEMO framework, which includes things like guardrails around deployed models. And one of the common conversations we have both at AI infrastructure and at AI Field Day is around the safety and security of your AI solution. And NEMO is very much aimed at safety part, making sure that you are putting controls around your AI solution, making sure that prompt poisoning and, uh, appropriate responses would only end up with maybe sexist language or, uh, inflammatory language coming outta these tools.

Uh, the use of these frameworks, the use of these, um, collection of tools from, uh, Nvidia definitely is gonna help customers build private AI solutions. And I think that's a, a significant thing. As I see the future of AI solutions, I can see large model training happening on public cloud, uh, but then we're starting to think about where's the data that's gonna actually feed the business application.

If that data is on premises, you may well want to have your AI on premises as well. Uh, of course, if the data's in the cloud, then maybe your AI's gonna run in the cloud. And we saw plenty of options for that.

So I, I think this is good. I think it is useful to have support for the broader set of frameworks that people are using to build their AI applications. And I, I did like that Nutanix is coming back to this idea of making the infrastructure invisible and allowing businesses to focus on, on natural applications that are delivering their value.

Cisco is making the leap into quantum networking, uh, particularly in a big way. Last week they announced new advances in quantum networking, uh, and a new lab set up to develop the technology. The hardware is a new chip developed in combination with uc, Santa Barbara, uh, the chip creates quantum entangled photons, uh, at room temperature, which is a good achievement.

Uh, and then these quantum entangled fo photons can transmit data instantaneously fast on the speed of light. Uh, CHIP can create 200 million, uh, pair of entangled photons per second and uses very minimal power. Uh, the second big announcement is the opening of a new quantum lab in Santa Monica.

Uh, the goal of the new Cisco Quantum Lab is to investigate real world applications for quantum networking, both in quantum computing as well as traditional. It. Uh, Tom, are you now the super position of two possible states that will collapse as soon as we observe you?

Well, I, I hope there's not an entangled version of me out there somewhere in the universe because God helped that poor man. Uh, this is an exciting announcement for a lot of reasons, because it means that we're starting to see more practical applications of these quantum solutions. And I love how Cisco kind of broke this down into we're doing quantum for quantum, but we're doing quantum for traditional two.

Now, uh, in the intro, Al you did a great job of explaining, essentially what you're doing in quantum networking is you're taking two photons and you're getting them to spin the same way that means they're entangled. It means that they are identical copies of each other so that when you separate them, what happens to one happens to the other because they're entangled. So that means that you can do things like transmit data across vast distances instantaneously, that you can create solutions for things that are, like, for example, uh, we had a great, uh, presentation last year from Cisco about quantum security where you can effectively ensure that the key that you transmit to, to decrypt things can't be intercepted because it's, if it's observed before it arrives, then the entire system collapses and the key is invalid.

But I think what was even more important was in the way that they were kind of releasing some of the details behind what they're doing. So this is something that we covered a couple weeks ago on the rundown. Remember when we had that quantum GPS thing where it was like, oh, well we can read the, using a quantum magnetometer to read, you know, minute variations of the magnetic field and it can tell you exactly where it's at.

One of the reasons why that works is because of a hyper precise timestamping. Can you think of another area where hyper precise timestamps can be important? I can Wall Street being able to know exactly when a trade was executed down to, you know, nanosecond technology or sub nanosecond technology could be valuable based on certain, certain market conditions and, and things like that.

But, you know, you look at other things like being able to do quantum research into pharmaceuticals and healthcare. Well, that's a lot of data. Those huge data sets don't transmit easily, but if you have entangled quantum particles, then whatever you find out in the quantum particles in the computer would be instantaneously transmitted to research sites.

Now that's a lot of data and 200 billion entangled photons per second sounds like a lot until you realize how many photons it takes to transmit data. And that means that these things are gonna have to be, you know, developed and increased and things like that. The value, I think, is that one, we're doing it at room temperature, which if you don't know, modern quantum computers require them to be chilled close to absolute zero.

So that's a lot of energy that's spent getting them where they need to be. Second of all, minimal power consumption like milliwatt power consumption, that is big. And, and in case you don't know, a lot of this is done with lasers, which is one of the reasons why this lab was opened in Santa Monica was because there's a massive amount of laser research that's being done down there.

So I am excited for this, but just like kind of everything else, like the quantum GPS story or pretty much anything related to quantum computing, we are still years away from seeing practical applications of this. And I think that, you know, we, we kind of need to take this with a grain of salt. I, I, I'm excited to see where people are gonna go with it, but it's gonna be a while before I have a quantum computer in my pocket, or I'll just be able to think something and Alistair will be able to know what's going on instantaneously on the other side of the planet.

So, you know, stay tuned if, if you've been paying attention in the space. You know, there's been a lot of layoffs over the last few weeks and now we're seeing some news from Compt that the US Bureau of Labor Statistics shows that over 200,000 fewer jobs exist in it as of the end of April. There's a lot of uncertainty around the impact of tariffs in the United States, and that factor among many is leading employers to start to wait before they're looking to fill roles because they never know if they're gonna have to cut their payroll back even further.

And of course, that means that economic uncertainties are really causing people to put their job prospects on hold and companies to, to put their hiring prospects on hold. But I guess the question is, al who's gonna blink first when it comes to who's gonna wait this out? Well, I think that's gonna be a, a real challenge in that this is not a situation where people making, uh, making business decision based on internal business, uh, elements.

Having the, the huge uncertainty from the flip flopping of tariffs has been very problematic. People keep hoping that it will settle and it keeps being, well, we'll get a, things will push out a little further and a little further. Hiring staff is a long-term commitment.

You don't typically get value from a new staff member for weeks, possibly months after they start an, uh, an organization. And so there is a pretty significant commitment time here for people who are, for organizations to, to bring in new IT staff. I mean, the good news is the overall unemployment rate in it is still significantly below the, uh, global, the national, uh, unemployment rate.

5% as the unemployment rate in it, which is below the, uh, the national rate, uh, that is escaping rate right at the moment. Uh, we do also see that there are seasonal trends around employment and unemployment in, uh, in it. It is the bigger feeling that there is fallout from the wider economic situation of it's not exactly a boom time at the moment.

Cost of living is biting people. That's the cost of supplies as biting organizations as well. And so delaying decisions where they're not clearly going to be beneficial is just continuing to be a, a problem.

And of course, it flows on from the end customers, right through all of the businesses that supply those end end customers, um, right, right to the back end of manufacturing of, uh, of IT products as well. So it is gonna be a challenge. It is gonna be a challenge probably for months to come because we don't see this settling very fast.

Hopefully, uh, business organizations will start to commit to the things that are relatively low risk and maybe scaled back deployments, but I think it's gonna be quite a while while things are still tight. And, uh, they may well be a little slow to come back when they do return. So the other factor of course affecting this is the rise of AI as a way of making the, your existing staff more efficient.

And that can be a bit of a challenge. I think it's, it's a little overblown. The, the impact of AI taking away skilled jobs.

Uh, I think there are other market movements that are changing what jobs are valuable, but certainly it's easier to get a job in AI and an AI infrastructure than potentially in, uh, first line support or, um, maybe in in new technologies as well. Linux is no stranger to being updated. Core functions of the operating system are constantly being rewritten, improved, and one of the biggest parts of the software is about to be upgraded in the forthcoming Ubuntu 25 point 10 release.

The SDU command used in order to make sure that your partner will make a sandwich for you is used to temporarily change user privileges and run commands being run. Uh, the pseudo command is used to temporarily change privileges to run commands and is being changed from the core new version to a new rust based version. I'm not sure that I want to sdo and into a rusty sandwich being made by my partner, but that's an old joke.

Why does a bungee think it needs to happen? According to the release, the core reason is security rust has better memory management and memory protection, which means that reduced chances for unplanned privilege escalation. The goal is for pseudo RS to be able to drop in replacement.

That won't break anything. How often do drop in replacements break? Nothing, Tom.

Uh, very rarely, especially if you don't escalate privileges to drop them in. You know, the whole thing of if it ain't broke, don't fix it. Like I get the idea behind this, right?

Like, like if we write it in rust, rust has all these protected memory spaces and, and, and we can drop it in and we'll just rename it and it'll work the same way that it worked as the last one. Most of the time, except for that really weird corner case, wanna know one of the biggest problems they're worried about right now with this switch? The resulting command file's too big to just drop in.

'cause it turns out that when you write something in, uh, was this written in CI don't remember if this was C or if this was in, I don't know, Pascal or man, it could even be assembly language at this point. Um, rust is just bigger. Like, like I get it.

And we've, we fought this like all you gotta do is jump on the kernel mailing list and wait about three months, and then someone who didn't do their homework is gonna jump in and they're gonna ask Linus, Hey, we need to rewrite the kernel and c plus plus because it's better. Hey, we need to rewrite the kernel and rust because it's better. Hey, we need to rewrite the kernel and I don't know JavaScript because it's better.

And every single time the response comes back, no, if he even responds at all, because we've been down this road a number of times. The issue is that just because something is quote unquote better doesn't make it better. If you want proof of this, I want you to go onto your system and I want you to look at the timestamps or the date stamps on some of the commands that are the, especially the, the low level applications that you're using.

They're old, very old. Like think about the, the, uh, command prompt on your system. I don't think CMD EXE has been modified for a very long time other than maybe like changing the date stamp to match all the other files on the system when you upgrade from like Windows 10 to Windows 11.

But the system is feature complete at this point. Why? Like, I, I get it.

Like people want tabs on their terminals and people want this and people want that. And I, there's plenty of drop in replacements. I use one on my Mac instead of the, the, the terminal window.

But that's just it. I am choosing to use a terminal program that's not the terminal because I don't wanna try to replace the terminal. 'cause I did that once on a Windows machine.

I decided that I wanted to replace Notepad with a better version of Notepad, and all I did was drop my better version of Notepad into the system and rename it to Notepad xe. And the amount of pain that I caused myself was way more than I should have. I get it, Ubuntu, you want to, you wanna tweak things and let's be fair, we survived the System D nightmare.

So this should not be this big of a deal, but I need this to be bulletproof before you guys decide that this is done. Because if you screw this up, there's no way to come back from it. Well, I mean, unless you run as root, in which case, more power to you.

We had a story we wanted to take a closer look at and it comes from the future room group because last week saw the release of a big new report around cybersecurity. 6% per year for the rest of the decade, and by then it should be worth nearly $300 billion. In a quote for the linked article, Fernando Montenegro says that cybersecurity is a cornerstone of the boardroom and a key part of enterprise risk management strategies.

And additionally, the report has a lot of great suggestions and information about which areas of security need to be addressed by organizations in the coming years. Now, Al you and I both touch security quite frequently because, well, it's part of everybody's job now, but I was wondering if you had some thoughts about this report and kind of what it might mean for the wider security industry. Well, there were reporters, some nice, uh, in depth analysis of market segregated across different geographic regions, including the APAC region where I live, uh, but also segmented by industry and then by software delivery mechanisms.

So it's a, a pretty comprehensive piece of research and that foundation in, in lots of backwards looking also helps with the forwards looking. I always like predictions that are based on some sort of views of the past as well as some, some insights and other changes. Uh, definitely the idea that the security market is, is going to grow shouldn't be a surprise for anyone.

We know that there is no reduction in risk coming on around us. Uh, the, the internet continues to be a dangerous place and your employees continue to be a significant risk to you as well. So seeing areas of risk management as a, as part of the overall cybersecurity approach, being a, a place that, um, that the report is, is pointing towards that, uh, Fernando and Krista case, uh, pointing towards, uh, risk management and identity management is being important.

I think the identity management one is really crucial because as we're seeing more and more use of hybrid cloud and multi-cloud, uh, the need to identify somebody uniformly across all of those different locations where they might act is really vital. In order to get a consistent security stance, you have to know the actors who's taking what action. So, uh, that focus on, uh, both risk management and identity management is absolutely, uh, strong, uh, in terms of delivery mechanisms.

That's one of the, the interesting findings in here is the growth has been a lot in cloud delivered, um, security and cybersecurity tools. There, there are a number of places where having these tools as a managed service running in a secured cloud location is very beneficial. But also, as I mentioned, that's a hybrid cloud, multi-cloud, it makes sense to deliver is your cybersecurity solutions through those hybrid and multi-cloud locations.

Of course, you have to deliver security in all of the locations you're operating, not just in the cloud and so on-Premises deployment and, uh, software deployment into maybe public cloud instances as absolutely on there as well. Uh, there's some advice in here for both vendors, vendors wanting to get out in front of this and be part of that 10% compounding growth for the rest of the decade, as well as to, uh, end customer organizations who need to consider really planning for an increase in the, the complexity and cost of their security implementations over time. Tom, you are very much deeply into the security of security with your security field day events.

What's your view on this lovely report from rum? So I love the fact that we went out and we found a lot of decision makers who are talking about the things that they're happening. Uh, one of the things in the report that I thought was really impactful was how impactful security events are.

You know, we, we, we think about this like, you know, I've, I've seen news stories about how important it's to have quantum resistant encryption now because the likelihood of RSA being completely broken in the next few years is higher than it needs to be and, and this and this and this. And then I go look at the report and, uh, things that people are really worried about, things that have had the most impact, uh, cloud security problems. You know, like, uh, buckets not being, uh, secured properly or someone losing their credentials.

Uh, good old fashioned ransomware. Hey, we, we do ransomware stories all the time around here, right? Uh, something that you probably think about but you don't really think about.

Social engineering and phishing. Like, look, I got an email today that came in with the right email address, but just something felt off about it, right? It's like, go over here and click here and obviously you did this and you're like, huh, something doesn't feel right.

And so I start doing the research and I even went over, you know, I did whatever security research you're supposed to do, I typed in the address of the service that I was using as opposed to clicking on any of the links in the email. Sure enough, it was bogus, but it goes, it stands to reason that most of the people who are not super security savvy in your organization might have fallen for that. But how do you teach that?

Like how do you teach people to secure us three buckets properly? Uh, how do you prevent them from exfiltrating data like these as, as sophisticated as the attacks get? Like believe me, when we do the re the, the research for these, uh, rundowns, like there was one the other day, it was like, well, we can pick out your password based on the acoustics of the keyboard that you're typing on.

Okay, that's like some James Bond level stuff. And I always go back to the XKCD of like, you know, using polymorphic encryption to guess the guy's password and then the next panel it's like, let's just beat him with a wrench until he gives us the password. Like that's what we're dealing with here.

It's low tech, it's effective. And these are the kinds of controls that need to be put in place. And that's one of the things that the report goes into detail about, as you've mentioned, is the interaction of things like IAM with cloud security and data protection and, you know, endpoint protection.

'cause we are long, long past the days of antivirus. Uh, if you think that people still run antivirus on their machines, um, you need to read this report because it's really EDR endpoint detection and response. And even that is kind of an old way of looking at it because we moved into new areas where the endpoints work with the rest of the infrastructure to prevent these things from happening.

I mean, if I'd have had the ability to prevent code execution on my laptop, you know, 15 years ago when I was still doing this for a living, my life would've been a whole lot easier. And so I think it's valuable for people who are kind of curious about where the direction of the market is headed. Should definitely take a look at this report.

Um, I know Krista and Fernando and the rest of the team over at the future and research division have put a lot of effort into this and you're gonna be seeing a lot about it over the next few weeks. So take, do yourself a favor, go register for the report. The link will be in the show notes.

Um, take a look through it and I promise that you won't be disappointed with any of the information that you find. Another thing that will not disappoint you is all of the great events that we have running here at Tech Field Day. Uh, there's actually one going on this week.

Uh, we are gonna be doing the Tech Field Day experience at Click Connect 2025 in, uh, sunny Orlando, Florida. Uh, Steven is down there right now with the folks from click. Uh, there'll be some great videos coming out very soon with, uh, some of the discussions that they've been having.

And then we, we'll be back, uh, at the end of the month, May 29th and 30th for Security Field Day. We have a wonderful lineup of presenters who are ready to bring you some critical information that you need to know about the state of enterprise security. And we're gonna have some great conversations and some fun discussions going on there.

The next week, Al's back in the States because he has something coming up. Al what are you gonna be working on? I'll be back in the States and I'll be back at San Francisco Airport and really adjacent to there when we have Cloud Field Day returning with a whole bunch of interesting vendors.

Uh, there seems to be a, a bunch of storage components and maybe some networking to be had in your hybrid multi-cloud. Check it out on the fourth and 5th of June And be sure that you tune in for the rundown every Wednesday. We love recording the tech field day rundown for you.

Uh, whether you catch us, uh, right when we get published on YouTube, whether you're listening to us in your podcast feed while you're mowing the yard or maybe uh, possibly even typing out an email on your porch and getting some sun, uh, you know, we love that. it to check out the show notes. Uh, we have links to all the articles that we pulled and sourced things.

And don't forget that the link to sign up for that future and research report is there as well. com. We have a lot of great, uh, resources there that you wanna check out.

We'll also be back next Wednesday with more great rundown news. We hope that you'll be able to tune in then. Until then, for myself, Tom Hollingsworth, for Alistair Cook, Steven Foskett, and everybody who works really hard to make the tech Field day rundown happen.

Enjoy the rest of your Wednesday and we'll see you soon.