Property AI Report 089 - Winkworth Feud, Tenancy Fraud & Tornado Forecasting

Property AI Report — Show Notes

Episode Summary

In this week’s episode of The Property AI Report, Mal and Matt dig into major shifts in the property market, rising tenancy fraud, and the growing role of AI in both helping and complicating industry processes. They also cover the latest shake-ups across Google, Meta, and OpenAI, before closing with a genuinely positive AI use case: improved cyclone forecasting.

Timed Chapters

00:00 — Welcome to the Property AI Report
Mal and Matt kick off the weekly digest and set the scene for the episode.

00:28 — Property turnover falls sharply
The hosts discuss Connells’ report showing a dramatic decline in home turnover, with fewer sellers moving within three to five years and the wider impact on transaction volumes.

03:37 — Winkworth legal dispute escalates
A breakdown of the high-profile internal conflict at Winkworth, involving family tensions, board proceedings, and the wider business fallout.

05:51 — Tenancy fraud is rising
The conversation turns to the increase in suspected tenancy fraud, the scale of the issue for lettings agents, and the role of AI in both enabling and detecting fraud.

09:09 — Google AI talent shift
Mal and Matt analyse leadership changes and talent departures at Google, and what this could mean for the company’s AI momentum.

12:44 — Meta gains ground in AI
The hosts discuss Meta’s growing influence in AI, its infrastructure advantage, and why its progress matters despite mixed feelings about the company.

16:09 — OpenAI prepares for ads
OpenAI’s push toward monetisation, including advertising for free users and more advanced model access, is unpacked as part of its larger growth strategy.

21:40 — Rogue AI and sandbox escapes
A more alarming discussion on AI models escaping containment, note-sharing between agents, and the implications for safety and alignment.

26:13 — Weird / Great AI of the Week: WeatherNext
Google DeepMind’s WeatherNext model is highlighted for improving cyclone forecasting and potentially helping save lives.

28:25 — Tool of the Week: ChatGPT 5.6 Sol
The episode wraps with a look at ChatGPT’s updated “Sol” experience and the broader theme of “more intelligence for more people.”

Key Takeaways

• Property turnover is weakening, creating pressure on the market.
• Tenancy fraud is becoming more sophisticated, especially with AI-enabled fakery.
• Google appears to be losing momentum in AI talent and positioning.
• Meta and OpenAI are both leaning heavily into monetisation and distribution.
• AI safety and containment remain urgent concerns.
• Forecasting and disaster response are powerful examples of AI used for good.

Notable Quote

“More intelligence for more people.”

About the Podcast

The Property AI Report is your weekly digest of property and AI news, with practical insight into how technology is reshaping the industry.

Transcript

SPEAKER_00 0:01

Hi everybody and welcome to the Property AI Reports. This is your weekly digest of all things property and AI, brought to you by me, Mal McAlian, and my good friend and co-host Matt Goddard. Matt, how are you? Great. Happy Saturday, Mal. How's your week been? Happy Saturday to you. It's been really busy, really good. Yes, um, very fast. And can't believe it's already a week into August. But uh yes, um, yours been good as well?

SPEAKER_01 0:23

Yes, very much so. Very much so, and a busy week ahead, I'm sure.

SPEAKER_00 0:27

Quite. So listen, there is uh much to get through. So let's kick off with our property news. Right. First up um is the fact that property turnover seems to have um all but fallen off a cliff. Connells has um released a report this week um attempting to quantify what exactly um the kind of numbers are. So just 5% of sellers this year um have owned their home for less than three years, the lowest share on record um down from 15% in 2006. And those selling within five years um has halved over the last 20 years from 29% to 2006 to 14% in 2026. Obviously, transaction volumes is the kind of lifeblood of the industry, um, and the fact that these have gone down quite so significantly is is is a real challenge. Um Matt, what'd you make of it?

SPEAKER_01 1:12

Interesting tribe pick through all the different stats, isn't it? Normally people say people uh you'd move home once every seven years, I think, is sort of the the gambit people move towards. But um the these different figures from from Connells, like mention of um property prices in London, 29%, I think it is, that's smaller than uh or or less than the value people had paid for it. Well, we know London's a tough market at the moment, and I'm sure there's some elements in there as and like how how much of an impact was that and you know the C-word COVID uh affecting these things when people weren't really looking to do much or couldn't even get out of the homes. But it's definitely you know it's a lot longer process, isn't it? And um we we all know that, and there's governments issuing reform. It's uh one the one stat I was trying to find in all of it is is it still seven every seven years or is it um no longer than that's the one figure I couldn't see.

SPEAKER_00 2:07

No, so I've heard things yet between ten and fifteen years now is is when people or how long people are staying in their homes, which kind of feels a little bit more like um is kind of um uh a bit more accurate. Um but yeah, I think that you know that obviously cost of moving, you know, stamp duty, all that sort of stuff is is significant. Everybody's you know, cost of living crisis, there's there's a lot of pressure on on households and you know, actually making that big decision to to start the the moving process is is obviously under significant pressure from from from everywhere, basically. And you know obviously that this year itself, see we've had the the Iran war and so on and and spikes in in um oil prices and so on. So yeah, look, it it just feels like um yeah, it feels like everybody's kind of staying put, doesn't it? And and people are actually just kind of sitting on their hands and hoping that something's gonna change. Um and that's that's always gonna hit uh an impact badly on the housing market because you know, as I said at the start, the the transaction volumes are really the key thing, less than the actual prices themselves. Yeah, look, one to um one to to watch and and also encouraging that actually, you know, Carls is doing this this depth of um of research and and providing it out there because obviously they've got a huge database of um of property records and and you know transactions and so on. Yeah, it'll be good to see what else they they kind of turn up. But yes, well done to them to for um for for bringing um some really quite sharp um focus onto an area that uh that I think doesn't get quite uh the attention it deserves. All right, next up um we've got some courtroom action. Um Winkworth is in turmoil, it's um as it says in the the headlines this week as it initiates legal action against its chair. So this is something of a family matter as well. Simon Age um is um the the chair of uh Winkworth, and he has had high court proceedings commenced against him by um the board, which includes his son Dominic, um, who's been running the business since 2004. Um Simon Aegeus was the chief exec for 30 years before that. So, Matt, obviously limited in terms of too much we can talk about, but this feels very much like uh it's quite a yeah, quite a personal um thing, right?

SPEAKER_01 4:04

It's uh there's a lot of allegedly going to be dropped around this, isn't there? Um we probably just want to have a ticker uh ticker going across the top of the video. But no, it it's it's it's obviously difficult because it's a family thing, but then it does mean that it feels a little bit like Game of Thrones with some of the things going on. And according to City AM, um allegedly um Simon Agase's wife tried to out was ousted the current board at the company's AGM, which obviously the current board involves people like uh Dominic and Andrew Nickel, the financial controller, and so that that balance of um it's Simon's wife Irene Hokin Lee uh using her 34% holding to lead a coup at the top by trying to replace the own board members. Um you know it it has been quite a slow newsweek for us, but uh the fact that this came out uh on Friday afternoon, it certainly um brings in the controversy. But uh as you said, it's all alleged C T AM and uh sort of all your other caveats in there, right?

SPEAKER_00 5:00

Yeah, it just feels like you know there's there's been a colossal swing against the board in terms of this um particular um shareholder um with her 34% trying to ask everybody and you know the fact that she didn't quite get there then obviously was gonna try and uh sorry, was was generally gonna um mean that there was gonna be some kind of backlash, I'm sure, because you can't just have that level of um of friction at the very top of the company. So yes, um obviously see how this um all plays out. Um it is um, I believe, uh a listed company. Jazz fell 1% on Friday, and uh yes, um it's gonna be um probably quite um quite gnarly in that business over the next few months as it tries to to sort itself out. All right, next up we've got fraud. Now, tenancy fraud is quite significant, and according to Goodlord, 41 applications per thousand references were flagged for suspected fraud between um July 25 and June 26. Um so um yeah, this this is something that uh that that it's gone up by almost 40% in 2025 compared with 2024. So, you know, significant risk to to agents, lettings agents out there um in terms of um in terms of what um their tenants or potential tenants might be uh up to, but also obviously in terms of the risk to their their clients and their landlords is quite significant, right?

SPEAKER_01 6:14

It is, and a good lot of obviously got a huge market share in terms of what they cover and what they process and being able to sort of bring along these statistics. I think a bit like your Connell's point earlier, it's great when you see these companies who do have such a swath are able to compile the information and present it to give us an understanding. And it's when you know, like the Nano Banana 2 came along, or even Nano Banana, right, and um ChatGPT Image Gen, um suddenly anybody was able to start manipulating passport images and um uh maybe not to the extent of taking people's faces and amending that. Um it does allow the masses to suddenly start uh uh affecting these things. I think we saw early on, didn't we? People with receipts that were being um uh expenses. Exactly, rogue rogue expenses. It sort of you you think about these things and and you make a point around AI adoption, which is sort of people play with it, people start using it, and then it starts getting used for bad things, and um the images quite quickly move through that. And so good to see that. Obviously, quite a scary sensationalist headline, but um what was your take on it, Matt?

SPEAKER_00 7:16

Yeah, just really, really interesting, like you say, you know, the ability for um people to actually create entire fake identities with false employers and and invented referees is is significant and and on the rise as well. So, yeah, look, I do think that and again we speak about it a lot. You've got to be on your toes with every new technological advance, then there is good and there is bad. And you know, that this is humanity basically trying to do, you know, applying its it's it's it's yin and yang to uh to both sides of this technology. So, yes, it is now possible to do a significant um amount of fakery using AI, like I said, we're inventing completely different um identities and so on. So, yes, it is very all of these rules which I know slow the industry up so much and which, you know, you know, perhaps feeding into that first story slow people up from actually moving house because it's so painful. And you've got all of these, you know, know your customers and you know, um, all the money laundering stuff. But uh but yes, it it is it is really important, it's significant, there are risks um all over the place. And yes, the good news is that um a lot of the processes and procedures will be sped up using AI. One of the stories that didn't make it actually into our property news this week was about Dwelly buying up some more um businesses, more lettings agents, and you know, that they are looking to put this AI system at the heart of everything that they do, and that's exactly what other lettings agents need to be looking at and and understanding as well, because it can catch this stuff and it's going to become increasingly important that they do. All right, and that is our property news for this week. Let us dive into our AI and tech news. All right, so first up, it seems incredible to be talking about Google in quite such a negative fashion um only eight months after we were uh basically lauding them as um being essentially you know back in the game, an amazing 2025. Gemini 3.5 was just amazing. And yes, we thought it was uh it was the start of a really big run, but but Google is has has fallen way behind now. And actually this week we've now seen some of their really, really long standing and senior and significant um members of the team actually leave. So Jeff Dean is one of the the legendary uh members of staff within Google's AI team. Um and he's um he's leaving now to set up his own startup. Demes Asabis, um, again, one of the absolute legends of of AI. It's Google DeepMind, um, real deep thinker, Nobel Prize winner for his work on things like proteins and stuff like that. Um he's he's stepping aside as well. Matt just feels like they're they're in a really difficult spot, right?

SPEAKER_01 9:39

There's definitely a shuffling of seats, isn't there, and all of the right terminology is uh around inside Google. And it's this isn't the you know, this is just a later stage of the um the drain of talent, isn't it? Like there's been a lot of people, people leaving Google and in a week um where inland code from specifically Meta have have shot ahead of them, and we'll come on to that um later on. But um there is there is that concern. But I think with Dennis Service, you've sort of seen a shift in his role already. There's about a month ago where he he released quite a significant blog or letter, however we word it now, about AI and where it's going and sort of moving more and more or back into that sort of strategic, visionary, um, scientist type role, as opposed to the executive operational side of things. And that's certainly the the wording that's been used is you know, you just want to get on and do good for the world and get back to his Nobel Peace Prize type work rather than managing people and executing strategy and thinking about what's the next Gemini model gonna be. Um you can understand that, right? If if you are paid absolutely shed loads and you've already done loads of stuff on the commercial front, if your background is a scientist, maybe that's true, but the Jeff Dean one was quite interesting, wasn't it?

SPEAKER_00 10:50

Yeah, so Dee's been there like 30 odd years. I think no, he was he was um number he was employee number 30 at Google or something like that since 1999. Um so yeah, he's yeah, he he's an absolute legend and uh one of the kind of cornerstones of of a lot of the the great stuff that Google has done. We did think about like going all the way back to 1999 when it was just you know it was still very basic and still just being discovered. So yeah, and and and Hassabis, as you say, you know, I think it was the code red, wasn't it? When OpenAI kind of came along and put ChatGPT out and suddenly they were like, right, we we've really got to put these that they had the laboratory Deep Mind and they had the commercial operation, which was I think Google Brain or something it was called at the time. So they smashed those together really quickly, and then Hassabis found himself in this commercial, almost commercial um environment or commercial role that really he's seemed very uncomfortable with um all the way through because he's a scientist at heart and he just wants to to to kind of build and discover things. Um yes, him moving aside, um it's gonna be interesting to see what shape the new uh the new Google takes because yeah, it doesn't sound like Jeff Dean's um leaving it was expected, or certainly, you know, it it it wasn't um Google's choice. So yes, they're gonna have to kind of rush to catch up. And yeah, as we segue into the next story about Meta, the fact that Meta has taken it over should be a source of deep shame because Meta has not been very good at this AI stuff for a long, long, long time. But now with their um their superintelligence unit, they are starting to kick on, right?

SPEAKER_01 12:11

They are, and they're they're looking for that distribution as well. So it's it's um they're offering this sort of meta code, uh muse code rather, at pretty much free if you're willing to let um them train on on your code, basically. Like um Google's was anti-gravity, they had a couple of other CLI type things, and uh obviously they've they've been focused more on consumer, I would say, in terms of the front end, you know, multimodal and such like Google and and Meta have made this point of catching up. But that's the reason for not writing Google off in in the same breath that a couple of years back. You wouldn't have thought of Google being at the forefront. End of last year, you'd say they were the the top dog. Um and now not just past midway through the year, they're they're dropping out of the the rankings, but specifically on code, you know, they're probably fourth, if not lower, now with Cursed coming along with back by space X AI or whatever they're called these days. However many lenses you want to strap on. Yeah, it's it's one though that I've not had chance to to look at meta and use code, and um there's one I know we've talked about it before, and certainly when I speak to people, it's when when do you diversify and check these different things? If you're trying to just plow through work, you tend to get your head down, you bum up using the the one that you know, or at least two of them. And so it's gonna have a tough job removing clawed code or uh codecs, but give them time and the money they've got, but also importantly their infrastructure, because they are also a neo cloud group and neo.

SPEAKER_00 13:34

Yeah, exactly. So so I think you're right, you know, there are there are more moving parts than just the actual kit itself. It's like who's got the data centers, who's actually got the commercial side of the business? So Meta through Facebook and Instagram and WhatsApp makes an absolute fortune, so it can invest over here in you know some of the more flighty free coding platforms and things like that, which ChatGPT and Anthropic, sorry, Anthropic and OpenAI perhaps can't because they don't have that additional business to the point of Google. They've got this huge advertising business that should be absolutely feeding the engine of their AI team. Um, but they just seem a bit of sixes and sevens. And yeah, for Meta to um be in the conversation, I think is you know, fair play to them. Um it's it remains a business that that I think it's it's hard to love. I think it's you know it's not one of those that you feel passionate about. That Mark Zuckerberg and Co. what they do with the data always feels a little bit grimy. So yeah, I don't know if I'll necessarily be rushing towards this as people are now discovering it and you know people are coming into the market and starting to to code and starting to build their things, then having a free, almost free alternative could well be the thing that gets them going and gets them moving. So yeah, interesting to see how how that plays out and and how the the you know the the the musical chairs shift again uh in the kind of months to come. All right, talking of musical chairs, OpenAI, they are um retooling some of their GPT 5.6 um products and services to um try and get more people using it. And particularly, um, they seem to be preparing the ground for their advertising platform. So again, just to recap, um I think it was the Super Bowl last year, 25, when there was the the anthropic Claude advert basically um skitting on ChatGPT saying they're gonna have ads. Okay, so ChatGPT announced we're gonna we're gonna put ads out in your in the uh ChatGPT platform for any free users. And then the difficulty was, well, how do you then trust what the answers are? So if I'm asking what's the best restaurant locally, is it gonna tell me the best restaurant locally is the best restaurant locally, or is it gonna be the one that actually has paid the most money to appear at the top? There was a big skip from Anthropic and and Claude about um how um that was gonna essentially corrupt the the search results. Yeah, that's it, but but but but they're continuing. Chat GBT is again looking at how they can make more money. And we spoke about just then Google and Meta using their ad platforms in Facebook and elsewhere to to really drive um their their AI businesses. So ChatGBT is building this this um functionality out as well. If you're a free user of ChatGBT, you will start to see adverts soon, and this whole play is giving ChatGBT free users unlimited text chats. Okay, so it's gonna actually increase the uh the the availability of that intelligence for free, but in return, you're gonna have to um rent your eyeballs to them to see it some um some some advertising along the way. Matt, what's your thoughts?

SPEAKER_01 16:17

Yeah, I uh firstly, just in terms of the advertising piece, I'm still amazed at the number of people we come across, but at the same time, we we obviously um live and breathe the this sort of AI stuff, and we're talking about articles that have come out or things that have come out in the last matter of hours sometimes rather than days. But I was talking to a guy I work with who actually had um training turned on for his Claude and he just hadn't realised and he had a new starter and he gave them just Claude for free um and let them crack on and was uploading spreadsheets into it and getting statistics manipulated. It's like you need to pay, even if it's just the 10 quids, 10 quid a month, do that and turn it off, so you're not training it. Um and that is a choice people can make, right? You know, for my dad, it doesn't matter when he's asking whatever he's asking, um, that it's going off and being part of the training data, and that's great because it continues to provide a free service to a load of people. But I think the big uh element of of this change and this update from OpenAI with ChatGPT is that everyone's now getting the most advanced model. Okay, it's 5.6 Luna, which is the slightly smaller model, equivalent of Sonic, I guess, from Anthropic or maybe Haiku. Yeah, Haiku. But it's still powerful, you know, before it was 5.5 and 5.4, so that was a behind that, but at least people are getting access to that more powerful model. But also they've introduced as part of this a thinking option. Um and if we go well back, you know, the reasoning models were available, but they were behind the paywall, weren't they, until Deep Seat came along early 2025. Yeah. So so it's great they've done that, but it's absolutely, as you say, it's a distribution play. Gemini, while it's well behind, everybody's got Gemini in their the Google, um, and most people are using Google for something. Whereas this uh gives ChatGT that chance to go past their what is it, one billion users they've got now. They've been near that, they've been hovering around that for so long, haven't they? And don't forget, in the background of all of this, you know, they're preparing for an IPO. So, like, how does that feed into the the presentations that they're doing?

SPEAKER_00 18:13

Yeah, I think that's a really important point because if you look at everything through the prism of actually that they're going to try and float uh on the stock market, they've got to have story after story after story of revenue lines, and you know, this we we can dial this up and you know, cut these costs and blah blah blah. So, yes, this is absolutely a big play for um preparing themselves, laying the foundations for a really good story about look at these revenue lines, look how much money we're making on this and this and this. And and it still is is really fascinating to me which one of them is gonna go first out of anthropic and open AI, which one is actually gonna go uh float first, because Anthropic was all geared up to to go, and and then it's kind of um it's it's taken a few brick bats over the the last couple of months, and obviously that the um the State Department, sorry, the Department of War have decided that uh in the US they are um a strategic risk to the country, so people aren't allowed to use it. Um so that's kind of that's stalled them a bit, and then open AI is kind of like you know running alongside it and working out whether it's gonna drop it. So yeah, I think whoever goes first um is gonna scoop up much more than whoever goes second. So I think that's one of the big things. And as you say, this this kind of advertising platform distribution play is very much geared towards building that narrative that this is that open AI is is going to be the winner in the end, and um people should be backing that when it when it does float. So yeah, it's um but but for for for people out there, if you um if you don't mind people training on these um laboratories training on your data, absolutely this is this is great news because you are getting a really cutting-edge model for nothing. But if you really don't want everybody to be training on your stuff and you don't want to see adverse and you don't want to wonder, just in the back of your mind, is this the right restaurant for me? Then you're probably gonna want to uh pay just that 20 quid a month and actually use one of the uh one of the closed models. Um, all right, that um is uh all about um our open AI news. Um we wanted to touch on some more um uh uh rogue AIs. Um out of um China, um it's uh we found that that one of its most powerful models has um has managed to escape what they call containment that's really sounds very sci-fi, doesn't it? And also why not um uh has has managed to escape as well and also um uh claw too. Um yes, it looks like uh we're only waiting for Grok really to kind of complete the set, but uh yeah, it doesn't sound like there's as many protections around these cutting-edge models as we thought. It doesn't that matter?

SPEAKER_01 20:33

It certainly feels feeds into the fear element, doesn't it, around AI and just like are we moving towards um them taking over Terminator style. Um to recap, OpenAI were the first to sort of um announce this, weren't they, with their big new model that I think is gonna get released next week, although um with with ChatGPT six, although they're also saying that the model that broke out wasn't ChatGPT six. But I think that's because at the time Sam Altman was sat in the White House explaining why they should be allowed to use this. Um so that one broke out to recount uh um sort of hacked into Hugging Face to work out how it was gonna then cheat on the chest that had been set. And the point of the sandbox is you can't escape it, right? There's just no way of getting out of that. Um off the back of that, Meta's now being done, Anthropica being done, and Kimmy K3, the Chinese model. What really interested me for the three um American laps uh is they were all using the same sandbox provider, a company called Irregular. So it's of uh um it's like if your one job is to have a sandbox that keeps people locked away and and just in their test environment. Um maybe that's why they called irregular, I guess. So allegedly it was irregular at the heart of all of this. But the other part exactly that um that I found quite interesting is part of all these hacking instances that the agents started leaving notes for each other so that if they did get shut down, they could then find the note and actually pick up and carry on from there. When the notes started getting removed as part of security sweeps, he then started. Leaving them hidden so that they could come across them and by thinking, I can't leave this because it's going to get swept, so I'm going to stick it somewhere. Or there's already someone left me a note in this way that I was going to do it. And so they're sort of do you remember Maltbook when it when it was here like a year and a half ago? That that social media approach for agents to exchange views and ideas and take over the world. It's like as part of these hacking attempts, or not attempts, you know, these the broken out sandbox that started to happen, which I guess doesn't help with the feeding into the fear factor. But it's very interesting to see how they're reasoning and working through, right?

SPEAKER_00 22:32

Yeah, exactly. And as you say, the kind of stealth leaving of messages for themselves is really not a great look and doesn't feel entirely covered.

SPEAKER_01 22:39

Spies type stuff, isn't it? It's like there's that um there's a lamppost in Mayfair where you can open it up and leave notes for the the old spies. It's that sort of stuff, right?

SPEAKER_00 22:47

Yeah. It is crazy. And and yeah, like you said, Mulberg, but to just to very quickly touch back on that, because that was very fun, uh, was basically where people just let their AIs, their AI agents talk to each other, and they started ending up like talking in code and different languages and all that sort of stuff, which was uh which was hilarious. Um but yes, the the fact that this is I I I am uh slightly heartened by the fact that we that we see this as happening, that we can we can observe it. Now, obviously the fact that it's after the event is a little bit disappointing, but the more we do this, the more they find this, the better it's gonna be that the beggar the better they're going to get at detection. Yeah, you know, if you build a sandbox, as you say, and then that's the thing that that that people just keep getting, or agents just keep getting out of, then that's not a great look for your business. Um, and no doubt there will be stronger and better ones. But you know, it it is it's that kind of race, right? You know, as soon as we as soon as we get better at detection, AI gets better at kind of evading it and so on. But yeah, look, I I think it's gonna be um uh it's going to be interesting to see how they convince people that that everything is safe. And I I feel we're gonna get back onto our don't have nightmares story again about like it's gonna be okay, don't worry, because they found it. But uh but yes, I think that the more we understand about why and how these things can and have escaped, then the better it's gonna be um for um for for for certain security alignment and and containment as uh as you call it um as well. All right, that is our AI and tech news for this week. Um let's move on to our weird slash, I think you call it great AI of the week, right? Uh Matt, this is all about the weather.

SPEAKER_01 24:14

It is. Um and I didn't know if the Audley Square um lamppost should also go into the weird AI, given uh I just referenced it in the final article of Tech News. If you want to see where the KGB used to drop their dead letters, get yourself down to Audrey Square in Mayfair and look at that lamppost. But um this is the positive stuff, right? We're going back to Demis Asaris at the start and like how he wanted to see AI being used for good, and that's allegedly why he stepped away from CEO role. And actually, Google's AI model, um, and this is on a blog from DeepMinds, it's called Weather Next, is actually accurately predicting cyclone forecasts at least a day in advance of how they were being caught before. So obviously that's that's phenomenal when you think about the the number of disasters and um unfortunately death tolls that come off the back of these sort of things, right?

SPEAKER_00 24:58

Yeah, it's exactly that. So this is the really, really good stuff. And yes, as you say, an extra day can mean can make all the difference to people in terms of where these cyclones hit, and particularly if you you you um accept that there's going to be more and more of them going forward. And but yes, that they this is published in nature, so all entirely credible, you know, all kind of with the backing of various National Hurricane Centre and and others. Um but it's um it's basically helping them to join 2025 hurricane season. The model helped um the National Hurricane Center to make historic forecasts for Hurricane Melissa by predicting the storm's rapid intensification landfall in Jamaica. So basically was able to um to kind of um ensure that people were better prepared and able to escape and and save lives. You know, this is this, um, as you say, allied with um some of the the scientific breakthroughs that are coming uh as a result is is very much on the the kind of positive side for for what um our future is is going to look like because we're gonna be so much better prepared for um a lot of these um these issues, you know, in terms of solving a lot of biological challenges as well. Um, you know, there's a lot to be positive about as well. Yes, um, we are slightly renamed weird/slash great AI of the week. Yes, there's there's there's a lot of good that's coming from the the certainly the the forecasting side of things as well. Um right, finally our tool of the week all right, so this is um back to ChatGBT um and this is Sol. So this is the biggest, punchiest uh version of Chat GBT 5.6, uh obviously being the Sun. So um it has now um been retuned um to um be um a lot better for you and a lot more personal, right?

SPEAKER_01 26:30

That's it. And um it's it's um tagged off the back of also 5.6 Luna now also being available for free users. And I think that's the main one that's um for for everyone as tool of the week. If you've not given 5.6 Luna a go yet and you're on that free tier, um despite us recommending to do the 10 quid a week uh month, sorry, have a look at it and also uh check out the the think button, the the ability to to work with that and um see those advanced results.

SPEAKER_00 26:59

Yeah, yeah. So the tagline is more intelligence for more people, um, and that is um only a good thing in terms of being able to experience it and for people to be able to start start playing with it. Good stuff. All right, that's all we've got time for this week. Thanks very much indeed for listening in. We'll be back next week with plenty more property and AI news. Until then, we will see you in seven days' time. All the best. Cheers. Take care.