11 Burst results for "Nick Pilkington"

"nick pilkington" Discussed on Artificial Intelligence in Industry

Artificial Intelligence in Industry

02:05 min | 3 months ago

"nick pilkington" Discussed on Artificial Intelligence in Industry

"Be act, and they're gonNA be vehicles on job sites taking action automatically based on the results of that analysis i. think that's where you get this really exciting closed loop system where they're gonNA be pots of industry in construction and Agriculture and energy that can be entirely. Automated where you have vehicles on every job site that on a schedule takeoff collective visual assessment of the site analyzed that imagery decide which actions need to be taken and sand that to other vehicles that take action automatically I think the clearest example of that is potentially in construction and agriculture where you could have drones flying over fields, assessing what inputs needs to be made into the field, and then sending that information to an applicator which takes off and does m crop spraying or if an orchard is ready to harvest, vehicle garden harvests off the oranges. Closing, that loop on some of these industries which they can really start to what to make themselves. I think that's the next step change in the industry where we actually stopped to take these actions automatically with unmanned vehicles as opposed to a foaming the inspections interesting. Interesting. Well, I. Like that as a I got all this Sifi stuff dancing in my head. Now about where that can where that can head off to but it'll be neat to see a couple things number one cow that comes to manifest in the future how that actual action loop. As you'd mentioned, kind of course, Lew Costello? We're where that leads secondly where companies like yours will fit into the mix right? Are you guys gonNA remain the drone intelligence that? Kicks off those actions or are you also going to have an ecosystem of the the hardware that actually goes in takes those actions you know who the heck knows but but what a what a wild space, it'll be ten years from now I guess now it's going to be super exciting I mean it all plays into all single vision of the future is. That, there is going to be a drone on every job sites and the capabilities. The abilities those vehicles is just going to increase over time. It's GonNa be exciting a drone on every job site. Will you heard it from the man himself Nicholas thank you so much joining us for another episode of the I.

Lew Costello Nicholas
"nick pilkington" Discussed on Artificial Intelligence in Industry

Artificial Intelligence in Industry

05:36 min | 3 months ago

"nick pilkington" Discussed on Artificial Intelligence in Industry

"Inspections manually but what's happening behind the scenes is we're actually seeing how many of the suggestions of machine learning algorithms of making off being accepted versus rejected by the Person, doing the inspection, and those are the metrics that we're looking at to see how effective the machine learning algorithms actually being at doing the same thing that the menu inspect to a do. So it's just by virtue of using the tool to do their inspections that these algorithms are getting better. I'm inspections of becoming more efficient is no sort of extra channel of communication on like how well as it doing it simply usage. So so that there isn't they like a an update cadence. So to speak as like, hey client, how's your experience been with the product? All right. Cool. Hey, look here's what we're seeing on our end. You know it seems like people are labeling successfully because I can see this going up like that's not part of that ongoing Combo. We can report on that, but typically you can see that from the usage of the products if you imagine like trying to perform one of these inspections where you're looking at a bunch of images for issues. But immediately that you look at those issues, they're already of suggested areas where the issues are on the image and you simply clicking accept or reject them rather than having to grow your own locations. That's a really really strong metric of how well the machine learning algorithms are doing automating this tasks and Sydney will check in with these customers on their their perception of the results but the rate. At which they accept in rejected suggestions of machine learning algorithm, that's already strong indicator for automated it is. Got It. So last question throb here, nick is around the present and future of AI products. You folks have built a product that various companies use. You mentioned the the industries you operate in. Now I sort of think about AI products in a couple different ways either there's going to be products that really do a great job and I really mean great as in I..

AI Sydney nick
"nick pilkington" Discussed on Artificial Intelligence in Industry

Artificial Intelligence in Industry

05:36 min | 3 months ago

"nick pilkington" Discussed on Artificial Intelligence in Industry

"A lot of people think that it's it's a one off process whereas it's an ongoing investments to build and maintain improve machine learning solution and I think that's where having let drone imagery in one place making it accessible to these customers. Machine learning models plugged into the workflow from the beginning. Makes that a lot more compelling a lot easier to handle. Can you walk us through? You know you can pick whatever industry comes to mind Nicholas Walk us through that consistent iterative building development, etc that you just articulating. What does that look like in practice you can pick any business process than just kinda give us a mental picture here I'd love to imagine what you're seeing in your head. Awesome. Let's imagine we're inspecting oil tanks on a site in these tanks need to be inspected once a month and we're looking for rust on any of the hatches on these tanks says typically a manual process that now every month adrain flies over that sites collect bunch of imagery in a pissing looks to that imagery instead of having to be on the sites Then initially, we may train a model just on the results of one of those inspections. So we'll take all of those images will label them, which to them must showing evidence of rest, and we'll train a model that stands up alongside the human inspection. So the second month when we do that inspection, those images have already guts, the problems highlighted in them, but a person still goes through checks and confirms and every time presidencies. One of those issues that the machine learning algorithm is tagged correctly, it mocks that. Antibodies and every time it sees the machine learning algorithms made a mistake. It just deletes that annotation. So now you still performing the inspection on the person's looking through those images manually but that process has gone through a lot quicker and in the background that machine learning algorithm has not lent a bunch more data it's lent where it's made a mistake in its. Length was correct and then in months three when that process takes place, the result is even more accurate and kind of was ongoing correction cycle that makes these algorithms more and more efficient more and more robust and ultimately better for the business and I i. think that's the dream is that it's better in month three obviously, it's not always right but but There's oodles and oodles of examples of the opposite happening..

"nick pilkington" Discussed on Artificial Intelligence in Industry

Artificial Intelligence in Industry

05:08 min | 3 months ago

"nick pilkington" Discussed on Artificial Intelligence in Industry

"So Nicholas glad to have you back on a making the business case episode. . We're GONNA talk about kind of ease of deployment at finding ways to integrate into existing enterprise workflows. . Before we do that, , you know your work, , a drone deploy vary some clients. . It's a much more robust kind of integration of the client's talent in your your talent of the clients tack and your tech others are plugging in playing it as a software. . Can you address that gradient for folks that aren't familiar with your? ? Business. . Absolutely I think when it comes to machine lending, , you've got this kind of spectrum of sophistication among our customer base. . There are some customers that have their own in house machine lending teams in training their own models they're got these models in hand and they're coming to join the boy saying, , Hey, , we need to plug this into our existing workflow. . And in those cases, , our role is a company is to make sure that we're providing the flexibility withdrawn deploy software that we can plug into accustomed existing workflow without disrupting, , everything and bring that kind of machine learning intelligence and efficiency gains to their whole workflow. . On the other end of the spectrum, , there are customers that less familiar with machine learning they still want to benefit from those those machining products and there are role is is actually take on the machine learning. . Task and to provide new machine learning based tools within drone deploy that just surfaced to those customers so where they would ordinarily just be performing vigilant SPEC with drums. . Now, , vision inspections are machine learning powered and they're much more efficient than they're highlighting issues automatically. . So we WANNA service on both ends of the spectrum there and I think the reality is still in twenty twenty production sizing machine learning tools is still a challenge is software that kind makes it easier and making. . Things more commodities, , but it's not just about training the model. . It's not this kind of fire and forget process. . We set something up once and then it just works forever. . It's really this kind of its ration- of getting something working, , adjusting retraining, , and it's all of that kind of integration work and the support for that whole model running in production that we want to provide to make it easy for our customers and I wanna go into what those elements are of making. . That production model work because that's the nitty gritty I think the business crowd who didn't go to Carnegie. . Mellon. . For Data Science they might have gone to you know working for business could really use some detail on but just just to flesh this out and put this in context drone deploy obviously you folks use drones is it primarily for Infrastructure Inspection Nacre? ? Should we frame your company in sort of a little bit of a wider sense here? ? So people know what you? ? Yeah, , we actually operate across industries. I . think Lodges Three Industries Are Construction. . Agriculture energy but there's this huge long tail of use cases in drums. . There is vehicle infrastructure this session rescue operations mining in aggregates just I think at this point, , we're finding that almost any industry in some way benefit from aerial imagery. . Got It in that context of training algorithms and deploying models in that world, , you know we we could talk about you know inspecting some kind of oil and gas equipment. We . could talk about inspecting crops. . Imagine you folks work the whole gambit there what are those real considerations for deploying models? ? Again, , we want to make this as easy as we can. . Especially, , you know you do <unk>, , you're selling the stuff. . So you WANNA be able to get this up in working for clients as quick as possible. . What are the factors that really have to fit into play that are sometimes quite challenging but you gotTa. . Work through them. . To actually make deployment work. . Absolutely I think the I is the is typically the machine learning problem that you're looking to solve how bespoke and how customers is there a bunch of very common patents in machine learning. . If you're looking for objects in an image, , you're looking to classify an image as good bad or exhibiting a problem or not exhibiting a problem. . There are kind of cookie cutter approaches to those sort of tasks in machine learning I. . think that's withdrawn. . Deploy can do an excellent job off the shelf by providing a bunch of machine learning technology that appliance can just plug into straightaway and reap the benefits of. . On. . The Mole bespoke side if you're digging deeper into these different verticals and more specific problems in more nuanced machine lending. . Solutions then I think it's much more of an iterative process and that's where. . John apply will typically deferred those customers to really understand that the main that trying to solve this problem and how to bring machine lending solution into production, , which usually involves getting a model up and running running on some type of infrastructure which drawn deploy can take care of, , and then working to improve what that typically means is letting it run assessing the outputs, , making corrections and retraining it and I think that's one of the pots of deploying machine lending solutions in production as typically overlooked. . A lot of people think that it's it's a one off process whereas it's an ongoing investments to build and maintain improve machine learning solution and I think that's where having let drone imagery in one place making it accessible to these customers. . Machine learning models plugged into the workflow from the beginning. . Makes that a lot more compelling a lot

Nick Pilkington UNITA CTO consultant e. m. founder
Building AI Products That Clients Actually Use - with Dr. Nick Pilkington

Artificial Intelligence in Industry

05:08 min | 3 months ago

Building AI Products That Clients Actually Use - with Dr. Nick Pilkington

"So Nicholas glad to have you back on a making the business case episode. We're GONNA talk about kind of ease of deployment at finding ways to integrate into existing enterprise workflows. Before we do that, you know your work, a drone deploy vary some clients. It's a much more robust kind of integration of the client's talent in your your talent of the clients tack and your tech others are plugging in playing it as a software. Can you address that gradient for folks that aren't familiar with your? Business. Absolutely I think when it comes to machine lending, you've got this kind of spectrum of sophistication among our customer base. There are some customers that have their own in house machine lending teams in training their own models they're got these models in hand and they're coming to join the boy saying, Hey, we need to plug this into our existing workflow. And in those cases, our role is a company is to make sure that we're providing the flexibility withdrawn deploy software that we can plug into accustomed existing workflow without disrupting, everything and bring that kind of machine learning intelligence and efficiency gains to their whole workflow. On the other end of the spectrum, there are customers that less familiar with machine learning they still want to benefit from those those machining products and there are role is is actually take on the machine learning. Task and to provide new machine learning based tools within drone deploy that just surfaced to those customers so where they would ordinarily just be performing vigilant SPEC with drums. Now, vision inspections are machine learning powered and they're much more efficient than they're highlighting issues automatically. So we WANNA service on both ends of the spectrum there and I think the reality is still in twenty twenty production sizing machine learning tools is still a challenge is software that kind makes it easier and making. Things more commodities, but it's not just about training the model. It's not this kind of fire and forget process. We set something up once and then it just works forever. It's really this kind of its ration- of getting something working, adjusting retraining, and it's all of that kind of integration work and the support for that whole model running in production that we want to provide to make it easy for our customers and I wanna go into what those elements are of making. That production model work because that's the nitty gritty I think the business crowd who didn't go to Carnegie. Mellon. For Data Science they might have gone to you know working for business could really use some detail on but just just to flesh this out and put this in context drone deploy obviously you folks use drones is it primarily for Infrastructure Inspection Nacre? Should we frame your company in sort of a little bit of a wider sense here? So people know what you? Yeah, we actually operate across industries. I think Lodges Three Industries Are Construction. Agriculture energy but there's this huge long tail of use cases in drums. There is vehicle infrastructure this session rescue operations mining in aggregates just I think at this point, we're finding that almost any industry in some way benefit from aerial imagery. Got It in that context of training algorithms and deploying models in that world, you know we we could talk about you know inspecting some kind of oil and gas equipment. We could talk about inspecting crops. Imagine you folks work the whole gambit there what are those real considerations for deploying models? Again, we want to make this as easy as we can. Especially, you know you do you're selling the stuff. So you WANNA be able to get this up in working for clients as quick as possible. What are the factors that really have to fit into play that are sometimes quite challenging but you gotTa. Work through them. To actually make deployment work. Absolutely I think the I is the is typically the machine learning problem that you're looking to solve how bespoke and how customers is there a bunch of very common patents in machine learning. If you're looking for objects in an image, you're looking to classify an image as good bad or exhibiting a problem or not exhibiting a problem. There are kind of cookie cutter approaches to those sort of tasks in machine learning I. think that's withdrawn. Deploy can do an excellent job off the shelf by providing a bunch of machine learning technology that appliance can just plug into straightaway and reap the benefits of. On. The Mole bespoke side if you're digging deeper into these different verticals and more specific problems in more nuanced machine lending. Solutions then I think it's much more of an iterative process and that's where. John apply will typically deferred those customers to really understand that the main that trying to solve this problem and how to bring machine lending solution into production, which usually involves getting a model up and running running on some type of infrastructure which drawn deploy can take care of, and then working to improve what that typically means is letting it run assessing the outputs, making corrections and retraining it and I think that's one of the pots of deploying machine lending solutions in production as typically overlooked. A lot of people think that it's it's a one off process whereas it's an ongoing investments to build and maintain improve machine learning solution and I think that's where having let drone imagery in one place making it accessible to these customers. Machine learning models plugged into the workflow from the beginning. Makes that a lot more compelling a lot

Lodges Three Industries Are Co Nicholas Mellon John
"nick pilkington" Discussed on Artificial Intelligence in Industry

Artificial Intelligence in Industry

04:47 min | 3 months ago

"nick pilkington" Discussed on Artificial Intelligence in Industry

"So Nicholas glad to have you back on a making the business case episode. We're GONNA talk about kind of ease of deployment at finding ways to integrate into existing enterprise workflows. Before we do that, you know your work, a drone deploy vary some clients. It's a much more robust kind of integration of the client's talent in your your talent of the clients tack and your tech others are plugging in playing it as a software. Can you address that gradient for folks that aren't familiar with your? Business. Absolutely I think when it comes to machine lending, you've got this kind of spectrum of sophistication among our customer base. There are some customers that have their own in house machine lending teams in training their own models they're got these models in hand and they're coming to join the boy saying, Hey, we need to plug this into our existing workflow. And in those cases, our role is a company is to make sure that we're providing the flexibility withdrawn deploy software that we can plug into accustomed existing workflow without disrupting, everything and bring that kind of machine learning intelligence and efficiency gains to their whole workflow. On the other end of the spectrum, there are customers that less familiar with machine learning they still want to benefit from those those machining products and there are role is is actually take on the machine learning. Task and to provide new machine learning based tools within drone deploy that just surfaced to those customers so where they would ordinarily just be performing vigilant SPEC with drums. Now, vision inspections are machine learning powered and they're much more efficient than they're highlighting issues automatically. So we WANNA service on both ends of the spectrum there and I think the reality is still in twenty twenty production sizing machine learning tools is still a challenge is software that kind makes it easier and making. Things more commodities, but it's not just about training the model. It's not this kind of fire and forget process. We set something up once and then it just works forever. It's really this kind of its ration- of getting something working, adjusting retraining, and it's all of that kind of integration work and the support for that whole model running in production that we want to provide to make it easy for our customers and I wanna go into what those elements are of making. That production model work because that's the nitty gritty I think the business crowd who didn't go to Carnegie. Mellon. For Data Science they might have gone to you know working for business could really use some detail on but just just to flesh this out and put this in context drone deploy obviously you folks use drones is it primarily for Infrastructure Inspection Nacre? Should we frame your company in sort of a little bit of a wider sense.

Nicholas Mellon
"nick pilkington" Discussed on Artificial Intelligence in Industry

Artificial Intelligence in Industry

01:54 min | 3 months ago

"nick pilkington" Discussed on Artificial Intelligence in Industry

"Call making the business case. This is where we don't talk about use cases. That's what we cover on Tuesdays on Thursday. We talk about buying selling ai and measuring the Roi of artificial intelligence and the bigger picture of a and business, and today we speak with Nick Pilkington, who is the? CTO and founder of drone deployed deploy has raised an awful lot of money apply drone technologies and computer vision to the maintenance of heavy machinery and other expensive equipment, and we speak with Nick about an important topic that we've been writing more about here at emerge, which is really around building AI products that clients will actually use. How do we build ai solutions that really keep the barrier to point low? That keep the the barriers to the kind of AI maturity. The kind of a understanding that the client needs very low so that even if novice level with regards this technology, they can still make the most of the solution that we built. Some of you listening in are already customers of our roadmap for building an AI product, which can be found under e. m., e., R., J. dot com slash reports. Pretty easy to find all of our reports there on that one page. But this is also a topic. We've written a more and more about an emerged plus when covid nineteen I hit my presumption was that companies we're going to focus more and more on pre trained models and more and more on finding minimal data sources that they could garner maximum value from putting less of an emphasis on transforming the. Data infrastructure of their clients and more emphasis on quick accessibility I think drone employees is a great example of a company that's really succeeded with that approach as opposed to a very bespoke hands-on white glove approach to UNITA client data, and it's something really important for those of you who are thinking about building a I products or building out a suite of services for a new company. So whether you're. A consultant or you run a company or thinking about spinning out a company I think this episode will be exceedingly helpful. We've.

Nick Pilkington UNITA CTO consultant e. m. founder
"nick pilkington" Discussed on Artificial Intelligence in Industry

Artificial Intelligence in Industry

02:38 min | 6 months ago

"nick pilkington" Discussed on Artificial Intelligence in Industry

"Let them innovate by making an easy to bring those machine learning algorithms to the platform, and have them incorporated in flow, so they can focus on that specific niche business case like looking for corrosion on a gas stack instead of that whole end to end of wears model gonNA run. WHO's going to maintain it? Who's going to retrain yet? Just the tip of the spear. When we're going to have another episode around the making the business case stuff in the integration considerations, but just to clarify where you're going number one. I think that that's a needed compelling vision I can see where you're headed. There is the aim in part for you folks to probably startups for for the most part I mean up until a certain point will sort of you know. Grab business if somebody raises their hand along as they have, the competency to deliver is the aim to sort of knuckled down more and more. More. You're saying individual client having enough data about a certain kind of infrastructure items so that they can keep iterating improving. Would it not also make sense for you folks to take the data from so many different instances when somebody else plugs it in for inspecting, let's say cell. Towers for telecom or flare stacks roiling gas that the you already have those models trained, or is the data proprietary per customer, and you really do have to kind of not extrapolate it into a broader product offering for you to take to the market. Yeah, we definitely need to kind of respect. Customers data in that sense like about customers are collecting data for their sites than it remains their data Goddamn Use that to build a better products and I think that's a key consideration of like. If you're using data from one customer train a model, let's giving an advantage to one of their competitors than that something you've got to be yeah loves where possible week keep those things Brits, but I think what's surprising is held little data and examples. You actually need to get something that incrementally drives value, even if you're training on a hundred images in, you're able to drop the number of images that needs to be inspected by human by fifty percent. That's still a huge saving you don't. Don't necessarily need a ton of data to drive value in these industries big time, okay cool so interesting clarification on the business model. I know there's some folks that are extrapolating from project to project clearly for you again. There's data considerations there. I always like to just get a sense of hey. Where's as company actually going and I appreciate you looking under the hood a little bit, so we've got a great picture of the before and after here neck I. Know That's all we have for time on this episode, but thanks so much for joining us here on the business podcast. Thanks for having me down..

"nick pilkington" Discussed on Artificial Intelligence in Industry

Artificial Intelligence in Industry

04:42 min | 6 months ago

"nick pilkington" Discussed on Artificial Intelligence in Industry

"We want to look at machine leading as this. This extra source of information that drives efficiency, and in this case it could be prioritizing which images need to be looked at could be verifying human results, and ultimately reducing the time to data, so the time from that asset being inspected to the results, actually being delivered where actions need to take place one person looking to thousand images. This is a machine learning algorithm going through those in two minutes is another big our Y. drive of their time to Beta introducing. Introducing them big time and the analogy I would use for the listeners in Nick. If you disagree or if you want to double down on something here, let me know again. It's really important for these episodes. The use case episodes is making our business audience have this new superpower where they can kind of see an example of a problem and understand where and how I could fit into it, not because AI is should be solving all problems, but to know which. which ones it could be snug fit for, and I think a good analogy. Here would be in radiology for example we scan a bunch of chest x rays. Are we have a system that is not gonNA prescribe somebody a cancer medication..

Nick cancer AI
"nick pilkington" Discussed on Artificial Intelligence in Industry

Artificial Intelligence in Industry

05:41 min | 6 months ago

"nick pilkington" Discussed on Artificial Intelligence in Industry

"You've probably seen one either in the cartoons or the news or maybe in real life so okay, so we're shutting one of those down means stopping the processes that lead up to it, which kind of means pushing pause on the money making engine here. Absolutely on these facilities only make money when they're operating. You turn these things off. Dollars getting lost. Yes, so okay inspection is solving. Inspection is a is a big problem. Okay, understood on end energy has enough problems as it stands naked this time of this recording right so inspections should should not be one of them so now we're talking about evolving process you folks are in the drone business here. What does it look like to limit that financial hit and bring about maybe a more efficient process for investigating this infrastructure? I think the two key parts to that. The first is drums, and the second is machine learning on the inside. The first thing is to make that inspection contactless to not have a person on that infrastructure, which means that you don't need scaffolding around the the asset on awesome. You don't need to shut it down at all. You can continue to perform these inspections on the same kind of maintenance cadence, so even frequently without having to shut down a toll, so that immediately is a massive kind of our y gain. These facilities continue to operate, but the inspections can also take place without having to cope people in harm's way. Got It. Maybe! We could just talk about the drone part I Nikki when it comes to using drones, I think somebody tuned in might just say okay. Well, you know we. We find a set of routes up and down. Whatever this piece of equipment is, that will hit on the same catalogue set of points that we needed a human to hit on. Maybe it's GonNa. Take pictures along the. The way is that roughly what we're talking about? We're programming this thing. A remote controlling this thing to essentially just snapshot Xyz you are s across all the different angles and positions that they need to in order for us to have the same set of pictures..

Nikki
"nick pilkington" Discussed on Artificial Intelligence in Industry

Artificial Intelligence in Industry

04:48 min | 6 months ago

"nick pilkington" Discussed on Artificial Intelligence in Industry

"So, Nicholas we're GONNA be talking about infrastructure inspection. It's been a while since we've actually talked about this theme on the show I think we had neurology on back in the day at and T. touching on it briefly you folks who work in a number of sectors and oil and gas and energy is a really big one. Can you talk to us about what infrastructure inspection is in that sector and also how? It's done today of the pre Ai Picture? Absolutely damn I think within energy there a ton of different verticals if you think of these different verticals is lots of different infrastructure in each of them. If we take an example from oil and gas, you can imagine something like a stack. That needs to be inspected and their number of different reasons why might need to be inspected for erosion corrosion for any type of damage or any sort of maintenance inspection? Inspection the way that's currently done is you have to shut down that facility to do it safely once it's been shut down. You start to erect scaffolding tax. She gives you access to the assets, and that costs a bunch of money, typically hundreds of thousands of dollars and can take about six weeks after that. You're actually in a position to put people on the asset to actually do manual visual inspection. Got It so it currently we're talking about cutting the thing off an enormously speaking for this particular unit first and foremost. Is there some context on how many of these things there are? I'm imagining that maybe I'm shell imagining and maybe I'm BP imagine maybe I'm a smaller firm I happen to not work in that sector specifically, but just. How prevalent is this particular kind of equipment? Yeah there's definitely a range depending on the company sized. Companies can have thousands of these assets across the United States. Okay so just in the US so lots and lots scaffolding people set it up. And then what are folks doing? I imagine we have some kind of a checklist. These are experts who understand what does broken component AAC opponent B components. See look like and what does. Clean and working component, a component components look like they've been trained enough. They have their manual. Whatever the case may be declined, they go through a checklist. They come down and then there's like an assessment and we get to decide. What do we WANNA go back up and fix is this. Is this roughly what it looks like? Would you add some some detail to that example? I think you're on the money they're. They're typically looking for a bunch of known problems I think there are these kind of edge cases where something comes up that they would have liked to inspected at the time, but they just didn't really have it on the the topless while at infrastructure shut down. So you see that in looking for something specific, and there'll be a catalogue of issues that trying to detect. Also qualified the amount of damage, and if there needs to be some response to that got it okay, so and then if they come down in there, there's some consensus on. Hey, I'm using a very simple set terms here nick for the audience, but we have a one through five right one meaning. There's absolutely nothing we need to do this. Basically, no need to look into it a five being by Golly this is about to you. You know be a horrendous mess. If we have a number of items that are maybe in the three or four range, you know there may be a consensus decision that okay. We need to go up and we need to take X. Action, or you know have a couple of things in the three range, but they don't seem to be super pivotal. We might say well. We're doing another inspection in twelve months or six months or whenever. And, so we actually think that it makes the most financial and safety sense to actually just checking on it at that later time..

United States Nicholas Golly nick