Predictably Good - Conversation with Patrick Quade

Patrick Quade is the Founder and CEO of DineSafe and creator of Iwaspoisoned.com, a crowdsourced platform for reporting and detecting foodborne illness outbreaks in real time. Originally from Australia, he has a background in finance and technology, with roles at Morgan Stanley and J.P. Morgan, and founded DineSafe after experiencing food poisoning himself. Based in New York, he leads efforts to use consumer data and analytics to improve global food safety, with his work featured in major media and research publications. 

---

Where AI, Data, and Predictive Intelligence Shape Better Decisions

From public health to financial markets, from food safety to nuclear systems, the world runs on decisions made under uncertainty.

Predictably Good explores how AI, machine learning, and advanced analytics are transforming how we assess risk, model outcomes, and make better decisions across industries.

Each episode brings together leaders from science, industry, finance, and policy to unpack how predictive intelligence is being applied in the real world—whether that’s safeguarding food systems, managing financial risk, designing resilient infrastructure, or navigating global uncertainty.

We go beyond theory into application: early-warning systems, probabilistic modelling, scenario analysis, and decision-making at scale.

At its core, this podcast is about one thing: how to make better decisions when the stakes are high and the future is uncertain.

Join Predictably Good to understand how predictive thinking is shaping safer, smarter, and more resilient systems across the world.

Hosted by Cronan McNamara, Founder & CEO of Creme Global

Transcript

SPEAKER_02 0:03

Welcome back to Predictably Good, where we explore how data AI and analytics are shaping the future of food safety and decision makings. One of the biggest challenges in food safety outbreaks has been detecting food safety outbreaks early. How quickly we can identify risks and respond. Traditionally, that has relied on structured reporting systems, which sometimes can take time to get the data we need. So today I'm really pleased to be joined by Patrick Quaid, founder and CEO of DyneSafe and creator of the iwaspoisoned.com platform that uses real-time consumer reports to identify signals and potential outbreaks and illnesses earlier. The company then works with industry regulators and public health organizations to help them turn those signals and insights into action. Patrick began his career in finance and technology before a personal experience with food-borne illness led him to apply his thinking and background in data and systems to this space. Since then, iWasPoisoned.com has grown into a globally recognized early warning system supporting many organizations to identify outbreaks and improve response times. So today I'm delighted to explore this idea of crowdsourced data, AI, and real-time reporting that are reshaping the future of food safety, I think, and what that means for the industry. As always, we're open for questions. So please do type them in the QA box and we'll respond towards the end. So Patrick, it's great to have you with us today. I believe you're in Austin.

SPEAKER_00 1:52

Yeah, uh yeah, here in Austin, Texas. And yeah, thanks for that. So thanks so much for having me. Great to be here, and I really appreciate it.

SPEAKER_02 1:59

You're welcome. Let's dive in. Um, so your journey journey into food safety is very interesting and actually parallels my own. I I did um technology under undergraduate and master's, and I worked in finance as well for a little while until uh uh a serendipitous uh food safety risk project in Trinity College brought me into this industry and I've been working for it ever since. So can you tell us a bit more about how you know the creation of iwaspoison.com came about and how it vol how it has evolved to today?

SPEAKER_00 2:29

Yeah, it it was like you said it at the top, it was a personal experience. I got sick, I um called the place that I I thought I I got sick from to let them know, not yelling or screaming, and I got the hung up the phone as soon as as soon as I said uh you know, I I think I got sick from your foods. I was like, wow, that's that's you know, there should be a system. I was like, what if people are kind of getting harmed all throughout, you know, throughout the area, and and they simply at least at the at the store, they they had no interest in in um in hearing about it, you know, there should be a system for that. And um I I wasn't like launching a company, I just thought that there should be something there that could capture that. And I built it and that was it. It was there was no fanfare, there was no um kind of like launching a company, and I didn't think I I didn't know if it would sort of take it was more of a I just m thought I should do something, and I that was my my way of doing something. Yeah, it just organically um grew legs, um, and it um it became clear that there was a very big gap there, and um and that there was an opportunity to really do some good work there. So that was that was what drew me into it.

SPEAKER_02 3:42

Excellent. And um, so from the time you launched it until you decided, oh, this is actually something we could build a business around, how long did that take? Uh was it weeks or months?

SPEAKER_00 3:52

Oh, I I I would say years. Yeah. I mean it literally, yeah, it it launched, it launched, and I think like I didn't even know if anyone would find it. I actually didn't know anything about how people found things on the I mean, I was like entirely clueless. And um, and uh someone reported like, oh my god, wow, like that that was it. I was like, you know what, that that gave someone a way to speak up, like awesome, you know, and and I didn't really think my and then another, then another, and it literally just snowballed until there were really powerful signals that I was sitting on, and I was like, oh my god, this is there's like very important information in here. And it was kind of that point that I was like, okay, this this this is a thing, you know, and and there's there's uh you know, it it needed my full attention to to to take it to the next level.

SPEAKER_02 4:39

Yeah, that's a super like organic growth kind of story, you know, just providing that really useful tool and then seeing it take off must have been exciting. Um, so that was the problem you wanted to solve was an as a system that could join join the dots there between a report and and then on a venue, perhaps that needed to know about these things. So, how does the um platform actually work? Um, like it how much data is in there, what what does it do and how do you make signals from all of that?

SPEAKER_00 5:11

Yeah, so so and that's the heart of it, Cronin. It's like, you know, one of the criticisms and their criticisms that you know that comes up when talking about crowdsource data and consumer reporting is that's well, you know, people um always think it was the last thing they ate, which which is um not always true, frankly, but but it it can be true. Um and also these are you know it's primarily self-self-diagnosis. So, you know, there are inherent weaknesses in when you consider both of those things. So the way that, you know, and it it became clear as we were going along that, you know, obviously, clearly a single report um you know has value from a consumer sentiment perspective and a customer care and customer recovery, even if they were wrong, there's still an opportunity for the the brands to kind of to do something that are interested in that. But um, and if it's a physical contamination event with photos and videos, or it's a packaging event, some of these things are like much more of a stronger signal singularly. But the the real value was when you get you know the 40th single report from independent households citing a sing a singular location with a 48 hour time. That was when it's it it's hyper powerful because you know when you do the math on that, the likelihood of of something like that happening by chance is you know, you know, it's extremely strong signal. Um, and then when you start to layer additional things on, which we maybe we'll talk about later in terms of benchmarking, and you're looking when you have a large amount of data and you can look at a brand or a product relative to peers, normalizing it, whether it's for you know market share or for store count or whatever it might be, and you see something that's running 10, 20 times higher or suddenly spike up out of nowhere. So that's the way that we've found it to be um strongest, you know, and and um and have the most value. Like there's value in the single reports, but there's there's an immense amount of value when you start to to add this stuff up and tally it using those techniques.

SPEAKER_02 7:15

Yeah, absolutely. That's uh yeah, exactly. It's coincidence it's beyond coincidence when you see that number of reports, um, and bearing in mind that people are a bit fuzzy about where they might have picked up, but if you see those trends arising, then it you you know there's something going on. Very, very valuable information for industries and businesses. So then um on the other side, then do you approach companies and organizations with this early insight signal, whether it's around, as you say, sentiment or packaging or food safety, do you actually go out there and try to talk promote this as an offering then to those businesses?

SPEAKER_00 7:52

Yeah, that's exactly what we do. So there's there's two parts to it. You know, well, there's there's really three parts. There's the consumer side, which is uh people who want to speak up, which is the beginning, that was the the genesis of the this um uh you know of this platform. And then there's the regulators and industry, um, and and associations to some extent, um, and you know, meta-organizations, but um public health, we have very large penetration for public health adoption. Um I were in um over 500 agencies worldwide, and um we operate natively in seven languages, I think it's seven languages. So um we um and we we translate, you know, uh to to make sure that irrespective of where it originated or how, it can be um understood by the recipient. And we um um also have done, you know, do outreach to brands and and you know uh and and try and let brands in is to know, like, hey, you know, we're here to, you know, I think there's some perception that like we're we're out to get brands or you know, or something like that, which is absolutely not true, um like you know, further thing from the the truth, um, but that we you know we have offerings to help them, you know, see these things early, you know, see the signals through the benchmarking, um, and and basically be plugged into this data to supplement it's it's a replacement for for anything that exists, either on the public health side or on the industry side, but it's it's a really powerful way to supplement you know other things they might be doing and seeing that that that we think and we try and communicate that and um and and let brands know. And you know, if there are individual incidents and we're not in contact with a band, we'll you know, we do the best we can to try and reach out to those brands. But um, you know, the the the thing that's um uh that's tricky is getting through to the organization or the organization within a brand that that is most um interested in it. And sometimes it can span you know the the QA, um, it can span um you know customer care, it can span um, you know, the social, you know, the social side that there's a variety of um areas and it's you know corralling that together to make sure that we're in time in contact with the right people inside inside the brands.

SPEAKER_02 10:04

Yeah, and what kind of reaction do you get? I guess there's a a variety of different organizations that some are more forward-thinking and they want to know, and others probably want to keep their head in the sand. Do you see that spectrum?

SPEAKER_00 10:16

Yeah, I mean, I think you know, the the deli that there was a deli um that uh in uh in um Tribeca that I was the the I've hanging up the phone I as soon as I heard it. So so that persists even at like at a much larger uh level. So um, and then there are other very tech forward brands, you know, that that understand how the world works. You know, this you know, people are on their phones, they speak up, they share that's what happens. That's not going away. If anything, that's gonna get stronger, and like they're ultimately like ignoring that is not gonna work out well, you know, because if you're sitting on the sidelines when something explodes that you could have been in front of, that doesn't end well for anyone. So um, so it's a mixture, but it's the the I think initially more resistance, but over time, you know, as brands kind of as things have evolved, brands are like have have kind of overwhelmingly sort of embraced the idea that this is we need to be looking at this, you know. Is every report right? No. What are the things that matter? Do they matter? They really matter.

SPEAKER_02 11:21

Yeah, yeah, it's better to know and have a look into it. Um so yeah, I suppose if you're looking at anything else on the market, like more traditional outbreak detection, like the things you might see the CDC doing and other like after the event um analysis can take time, you know, weeks and more to find out, you know, what happened or identify a pattern. Whereas yours could be um a signal before anything even, you know, anything like that occurs. Um, so the speed and dynamics um is is changing um in in the way yours can work. So, how does this real-time reporting like how how much does it change that speed compared to traditional methods?

SPEAKER_00 12:04

I mean, it's it's it's near instant. So now now there are things that are um that that that you know we naturally do less well at. Um and list theory is an obvious example, you know, with the with you know the the onset uh onset time and the the likelihood that um that we will triangulate something with that long an incubation is low. You know, just you know, so so it really depends on what the thing is. Um on the flip side of that, if you have someone that um had a um you know a diagnosis, because we when people report to us, they're reporting, you know, it's not just a commentary, it might also be, you know, it's uh it's photos, it's packaged, they'll send in um you know diagnostics. Um that's all the idea de-identified, but they'll share a great deal of information. So um so some of that can be um you know, can can be helpful as well.

SPEAKER_02 13:00

And I understand you've a lot more data than is actually what's visible on your website, you know. So a lot of people report and like without going public in a sense, like making a public um pro post on the site, right?

SPEAKER_00 13:13

Yeah, that's a great point. I'm glad you mentioned that because that's something that a lot of um you know, we we make it clear on the website. I mean, it's right there, you know, on the on the form. So consumers can see that they have an option to do that. Um, and we try and make it clear to um, you know, to other potential participants that hey, you can't just sort of look at the website and think we got this. We're you know, we're scraping the website. I it's around a third of the data is actually shown on the website. So um you will miss signals if you're simply scraping the website. Um, and you know, that that's important to know. Um, yeah, and uh and about that, it's it's consumer choice. We we literally just you know give consumers the option. Um the most typical is they'll opt into um helping detect an outbreak, sharing data with regulators, sharing data with brands who are interested, and um, but not necessarily putting on the web putting it on the website. And um yeah, that's that's a really important point. So thanks thanks for the mention.

SPEAKER_02 14:10

Yeah, so that uh it kind of comes to the motivation of the consumers, they're not trying to publicly, you know, shame or anything a brand, they're trying to literally report, and even if it's in private, that they know that you can use that for good and on the back end, which is is is a good thing to know about how the system works. So when that when then you do engage with a health authority or or a company, um how are they using those insights? Have you seen any good uh examples or case studies where they've used that to achieve a uh a good outcome in in one of those organizations?

SPEAKER_00 14:45

Oh yeah, absolutely. You know, whether it's on the public health side, um literally detecting outbreaks that go on to be um you know, go on to be confirmed, uh, you know, uh, you know, some agencies will fire all of their ad hoc inspections based on our reports, um, which I think is is you know, we where you know, if it's ad hoc, you know, inspection cycles anyway, then I we you know, we think that's like a totally decent way to do that. Um and um and certainly when there are clusters. And then on the the industry side, whether it's looking at um again, looking at clusters and doing a deep dive and remit try trying to remediate things or remediating things in advance and and basically preventing that headline story, you know, that that can come out because I mean the other piece of the puzzle here is there's I said there were three participants, there's actually a fourth participant, um, which is the news media, you well media, um, and I'll say news media and social media that um can can see this data and can amplify, you know, can amplify things. So um and that's if something never gets big enough, then it it typically doesn't, you know, it won't get amplified across you know across the broader set. So helping brands really, you know, um avert that by being early on it.

SPEAKER_02 16:04

Yeah, exactly. So these um so the agent the health authorities are agencies that can target their ad hoc inspections, as you say, based on the signals you're providing and then detecting outbreaks. So how what's the time difference, you know, between your prediction and them detecting?

SPEAKER_00 16:21

Is it the would it be a couple of weeks later or how oh no no, so like um, so you know, the the the health agencies will get a daily report for their jurisdiction, everything's geofenced. Um and they get um uh we'll notify them of clusters. So it's like, okay, well, it depends on what their threshold is. Like, you know, this is the fourth of the meta independent household reporting about this um specific location or product in an X you know time frame, how whatever their thresholds are. So those they you know, they um have the opportunity to investigate right away, and that's within a day or two, you know, 48 hours. Um if they're doing ad hoc inspections, inspections based on singular reports, that's the next day. So you know that can be the next day. So they get the reports and then they do. Um, so um uh and um yeah, for um, you know, we we also have a real-time um feature for industry so they can actually get reports like immediately, like as soon as they're moderated. And and the other thing, uh I'll do I I said the word moderated, so I'll just explain that what that is a little bit. Every report that we get goes through a multi-step process to um to eliminate authentic um and malicious um uh and um uh false reporting, you know. So whether that's spam, um whether that's duplicate, which often you know that that it can be as simple as just you know deduplicating and um taking spam out, um, but um also other signals where where you know where it's we have a lot of a lot of tools on the back end um to help us identify where something so we think of extremely clean data feed, um, yeah from that perspective. And um so once something's been through that that multi-step process, which is using technology and humans, um, then it can become available for the brands immediately.

SPEAKER_02 18:11

Yeah, so that then and the authorities are the brands can respond within 24 or 48 hours, whereas they wouldn't have any idea um before that there was something to look into. Um, are you using any of the the newer um AI tools in your in your system yet? Or have you looked at that?

SPEAKER_00 18:29

Yes, uh we do. We've been, you know, we're a technology first and um always, and uh yeah, we've we've heavily deployed um AI. Now AI is not it isn't a magic wand, it's like, oh, point AI at the system and it will tell you everything, which which is you know it's not the case, as as as you well know. So and some and then I think some people hear, oh, they're using AI, and then they saw AI hallucinate something while it's like this is all, you know, it's it's very it's very process driven. Um, it's very it's it's much more boring than it sounds. Um but it's critical in terms of like the speed with which we can operate, you know, whether it's um, you know, the labeling, you know, the optical, you know, even some older AI type stuff like optical character recognition that even predates the LLMs, but then the LLMs in um you know in a whole range of places that's allowed us, you know, to do a lot more, especially as we scale, whether it's like identifying um physical contamination where where a consumer might not call physical, they might call physical contamination by any number of ways, or then they might not even um kind of you know, it might be um from a picture, or it but able to like reason through and surface like physical contamination with high confidence and accurately, like in at scale. Like that's just one small example of like where the LLMs really shine lean into that thing, yeah, yeah, yeah. Categorizing things, yeah, um stuff like that.

SPEAKER_02 20:01

That's great. Yeah, no, the the the latest, I suppose, iterations of these large language models have become quite impressive and excellent at the kind of thing you need, which is text um analysis and understanding. So I'm sure you can they can help you make your detections even more accurate or faster to process um in your system.

SPEAKER_00 20:22

Absolutely. Yeah, no, they're my it's incredible it's it's incredible. It's I'm shocked every day, like the the things that yeah that they can help us with.

SPEAKER_02 20:32

Yeah, yeah, he's like set up a little agent that has a job to do uh with each incoming post. Yeah, but as you say, they're not quite there yet, where you just say have a look at my system and tell me what's going on. It's not would not be uh kind of the use case, just yet not yet, yeah, exactly. You still have a job to do yourself there. So that's good. Um yeah, so I suppose you know, looking ahead, have you plans or thoughts on the future of what this platform can do? Um are you going to expand it into any other dimensions of products or um how do you see this crowdsourced intelligence, real-time data streams, you know, helping the global food safety system uh in the future?

SPEAKER_00 21:17

Yeah, I I mean I think it it's a case of um industries uh you know might have initially been have a you know a negative reaction to this. And I think you know it's it's been an evolution to where like no, this is a part of the puzzle, really critical part of the puzzle, aligned with other tools that that they can use, you know, whether it's creme global or whether it's um you know the like um harvesting data that they actually collect themselves and kind of getting their own data in better shape and then bringing all this stuff together. So I think it's a natural evolution for um uh that that will be supporting and kind of helping iterate on with brands and bringing more brands into the folder. Who want to get involved with this kind of thing. So and I think in the end, looking at this data will be um simply be a best practice. Yeah. Yeah. Like ignoring it will not be a defensible position for you know for the C-suite or for, you know, or for um you know for the heads of QA or whoever the case might be. And like if you're sitting as a brand, you're, you know, or a a practitioner, and if you're sitting atop uh, you know, uh a very negative outcome, which may not be your fault, you know, but but if it's something where there were additional tools, additional affordable tools you could have used to help spot earlier, not just protecting your the consumers, but also protecting the brand, and you you opted out, you chose not to look at it, that's that's not a position that you that you want to that you want to be in, you know. And then yeah. So uh and so it's supporting that and then you know evolving with AI. I mean, we think the benchmarking, the benchmarking is massive. Um, the ability to see um, you know, uh and I'll you know, I'll I'll give I'll give one example, like a specific example that um um uh um if there's time for it, um relating to um Lucky Charms Um that actually got you know a lot of headlines. Um and um it turned out there was no um the FDA didn't didn't figure out exactly what had gone on, which was um uh which which is not unexpected, um because if it's beyond microbial, like you know, and it could be it's very, very difficult to figure out what had happened, but in that case, we had um you know, you have a very long baseline of reporting for Lucky Chums, like you know, over eight years of reporting. And if you uh you know out of nowhere have a spike that's over 40 times higher than historical, yeah. And this was pre this was pre-the-news coverage. So this pre-the news coverage, and uh you you uh you know, you ask yourself as a brand, like, do I care about that? We think the answer is yes. I that's something to look at, you know. Yeah now, especially that was an interesting one because um it was at the time, um it was you know, sort of in the middle of COVID, and actually on an earnings call, they you know Lucky Charms had said, you know, and I'm quoting the biggest issue we're seeing is around material selection, ingredients coming to our plants around our products, and some products we've reformulated over 20 times a year to date. Okay, so so as a company, you know you've re you've got a problem with materials, you've got you've reformulated over 20 times a year to date, you sell it publicly, and you've got a massive spike pre pre-news coverage. So I think some people maybe who are aware of this issue, like, oh well, this got in the news, and then the reporting went up. It was a 44x over baseline and peers before a single new line item of news. So we say that's worth investigating. Was it, you know, the the odds of that being a coincidence, you know, the the math on that, you know, better than anyone croner, like that that's um, you know, that's very clear. Um and you know, it it's you know, from where we're sitting, it's something you want to look at, you know. It's yeah, now the fact that we had 8,000 people come forward, over 8,000 people come forward after um, you know, after it it got traction and people started piecing it together in their heads. Now, some of those would be false positives, um, you know, confirmation bias, but we've had um we've done this many, many times over. We've seen other things where there was absolutely a can a contaminant, um, and it would it was um also got massive media coverage. We've seen 200 people come forward. So when you you you there's all kinds of benchmarking say, hey, this is the average reaction we get after something hits the news. This is the reaction that happened here.

SPEAKER_02 26:03

Again, like does it make it a truth?

SPEAKER_00 26:06

No, but it's certainly something you really want to tear into. And it's certainly having that lucky charm signal over four months ahead of the news. Uh that's a lot of time to to do to get ahead of that and to do something. So um that's you know, and it's putting these tools into the into the right hands of the brands that are interested in it is um is where and and customizing them, you know, to make sure that it fits for what they do.

SPEAKER_02 26:30

So do you you went with this information that you were seeing this? Um they may or may not have been receptive to that, but they never found out. Well, the the FDA or the authorities never actually found out any real root cause or what had happened, but there's certainly some kind of incident, and then it settled down again, or is that what happened?

SPEAKER_00 26:51

Uh so well, if you're that company, we'd and we'd like and you'd like to see where your brand stands, then we'd be like to hear from you a week. Um, but but uh there's well, I won't get into the details on that, but but certainly that um I mean and this gets back to the AI side of it a little bit is you know, when you're doing OCR on the um the product codes and you can trace it back to a month and a plant. Oh yeah. Yeah, very, very, very powerful. And so um that batch appeared to, you know, um, if it was, you know, that's what it it sure as heck looks like. Um yeah, that it that's certainly abated, but you know, in terms of where they sit now, that yeah, we'd we'd but I mean frankly, we're sitting on a watch list of like uh manufactured goods that we can see right now that are that are Lucky Charms esque and we're kind of we would love to have more of the brands that you know uh that are out there like working with us so they could actually look at it alongside us or take this data and do their own thing because there are brands that are sitting on that type of thing, you know, um today. Um so um yeah it's making sure that we're connected with those folks and they they're able to act on this and kind of you know uh work with with the data um to help their situation.

SPEAKER_02 28:14

Have you any um data or thoughts of expanding into cosmetic products or pharmaceuticals even?

SPEAKER_00 28:21

Or is it um we we have had we get reports across we actually have um toyed around trying to expand on that, um but we haven't for now. Um we we actually um uh accidentally expanded into fraud um by by on a different website um because um scams and online scams and fraud are are simply enormous. So that's like a tangent. So and that ex that kind of has exploded, unfortunately. Um uh and so we we definitely could do that. It's just a matter of like where we're the right place, you know, the the time and resources in terms of where to put our energy, but it's it's definitely something we've considered.

SPEAKER_02 29:06

Interesting. Okay, well, we don't have any questions coming up from the audience as far as I can see. So let me check and know. Um do you have any upcoming events or information you'd like to share with the with the audience before we wrap up?

SPEAKER_00 29:21

Yeah, um, I we got a few things coming up. There's um uh we're doing a data and ai safety workshop. Um, we're we're gonna be a speaker on that for AFTO, um, the Association of Food and Drug Officials in June. Very good. Um uh speaking alongside an esteemed company, Cram Global. I think you'll heard that one, is that right? I was gonna say, I think Brandon will be there. Yeah, I believe he will. Okay, yeah. Um, and uh yeah, we have uh a poster at the upcoming IAFP if you want to check that out. Um it's um on crowdsource data and benchmarking, and um we've got a white paper coming out um in concert with NC State. So uh that's another thing we're excited about for later this year. And um another AI consumer data talk with the manufactured foods um regulatory community that's coming out um uh later in the year. But uh if anyone wants to, I'm on LinkedIn, if anyone wants to follow and I can send out updates on that stuff if uh you know, love to love to love to see you there or um watch my posts.

SPEAKER_02 30:24

Excellent. Yes, you can follow Patrick or um Dyne Safe, I believe, on LinkedIn.

SPEAKER_00 30:30

Um yeah, it's actually just on my name, is where I post most of my stuff. Um we're a little uh not as we're not posting on the company page as much. But if you just follow me, then yeah, you you'll definitely have the latest and greatest. Yeah.

SPEAKER_02 30:42

Yes, indeed. And uh look forward to IFP. We'll have a few sessions there ourselves. Um, we'll try and organize our traditional drinks reception at one of the events. So if you're there yourself, we'll hope to be great to catch up again. Yeah, yeah.

SPEAKER_00 30:56

Yeah, I'm not sure I'll make it in person, but if I am, I'll definitely do that.

SPEAKER_02 30:59

Now you should uh be tempted to go now. Yeah, I yeah, well, I'll relook at the calendar, yeah. The Krem Global event there. Uh so yeah, it's become a bit of an institution. Um, well, thanks everyone for joining us. It's been a really fascinating talk um about a new, you know, a new area of uh um using data and and this crowdsourced insight that's generating signals, and Patrick's sitting on a lot of interesting signals there. So um keep in touch and reach out to him on LinkedIn, and I'm sure um we'll we'll be able to learn more at some of these posters and white papers and things you're gonna publish in the next few weeks. So thanks, Patrick. Pleasure. Thanks so much for having me, Cronen.

SPEAKER_00 31:37

Really appreciate it.

SPEAKER_02 31:38

All right, thanks. Thanks, everyone. Bye.