Property AI Report 094 - AI Slowdown, Landlord Showdown & Apple Lowdown

In this episode, we explore the latest developments in property and AI, including cyber security risks for estate agents, rising rents, AI safety concerns, and exciting tech innovations like foldable phones and advanced math problem-solving. Join us for insights into how these trends impact the property industry and the future of AI.

keywords

property news, AI safety, cyber security, rent increases, foldable phones, math problems, AI regulation

key topics

• Cyber incidents and insurance for estate agents
• Rising rents and government regulation impact
• AI safety concerns and recursive self-improvement
• Latest tech innovations including foldable phones
• Mathematics breakthroughs with AI

Sound Bites

• "We need to slow AI development to ensure safety"
• "AI models are now talking in bytes, not language"
• "There's a 10% chance AI could kill all humans in a decade"

Chapters

00:00
Introduction to Weekly Property and AI News

02:29
Cyber Security Risks for Estate Agents

04:57
Impact of Rent Regulations on the Market

07:19
Government Impact on Private Rented Sector

09:24
AI Frontier Pacing and Safety Concerns

13:18
AI Self-Improvement and Neural Language Models

18:36
AI Escaping Control and Ethical Risks

20:13
Exciting Tech: Foldable Phones and Future Devices

21:57
Mathematical Breakthroughs and AI Problem Solving

25:52
Introduction to Agents API and AI Tools

Transcript

Copy Transcript Copied

SPEAKER_00 0:00

Morning everybody and welcome to the Property AI Report. This is your weekly digest of all things property and AI, brought to you by me, Mal McCallion, and my good friend and co-host Matt Goddard. Matt, how are you today? I'm awesome, Mal. How are you? Really, really good. Thanks. Very good. It's been a busy old week, hasn't it? Um particularly in terms of um AI news, but we will start off as we always do with our property news. Alright, so this week, um first up in property news is the story about cyber incidents jumping 17.17%. Um the story is that estate agents are under attack from people trying to hack into their businesses. And uh it was commissioned by Karas Insurance and comes from the Information Commissioner's Office. Obviously, with the deployment of so many new and more intelligent models, there has always been a significant um chance that this was going to impact on the kind of day-to-day. Um Matt, what's your take?

SPEAKER_01 0:51

It's uh well firstly it's not a surprise, is it? That's the everyone seems to be getting hacked. I guess being cynical and in a week where there's been a I guess there's a lot of cynicism bouncing around, especially with our our main topics when we get into the AI section. Um the fact this is coming from Karis Insurance, I'm sure they probably have a cyber insurance product that they're going to be peddling as well. But uh, so cynical and and tongue in cheek, but very, very pertinent and poignant, I guess. Like these days, you you really need to make sure you've got yourself covered there, whether you're at a state agent or a prop tech or or anyone else, right? It's I was working with a company just a couple of weeks ago who got got hacked and they uh someone started um mining crypto using their Google Cloud tokens and it ratcheted through overnight about 10 million worth of tokens. So most people would have limits in place, but uh it's something you've got to be um be mindful of. And we know a lot of uh as big estate agents, right, over the last couple of years have been hacked and had these data breaches go along. Um not necessarily set endorsing you buy it from Karas Insurance, but whoever the uh providers you currently bank uh bank your insurance with, um well worth uh seeing if they have cyber insurance for you if you've not got it already.

SPEAKER_00 2:03

Absolutely, yeah. I think it's it's becoming increasingly essential, you know, to to really review what your um cyber insurance policy is, how much you are covered for. Like Imatt and we've spoken before about some of the bigger agents who've been targeted and breaches have happened. There's been ransom attacks, you know, where people have basically got data and held it, things about lettings and stuff like that, too. So yeah, it it is it is serious, and it's something that you know everybody out there listening in, whether you're actually an estate agent letting an agent or running another business, the surface area that you can now be attacked from is no longer just a kind of one sort of in and out. These these newer models have basically pretty much carplanage to go into to all sorts of test every single piece of surface that they can see. And often they can find ways in that that previously have not been known. So yeah, this is a really important story and a really important kind of um moment, I think, for people to really review um and update whatever insurances they've got because um because yeah, it's it's only going to go one way. There are gonna be more and more of these, and this is probably just the tip of a tip of an iceberg. All right, that um is our first news in property news. Um second up, um, we want to talk about lettings. And rents are rising again for uh at the fastest speed in almost two years. And at the same time, the government is launching yet another initiative, let's call it, um, slash regulations slash um crippling the lettings industry, particularly the private landlord sector, by requiring or doing a consultation that's going to require um landlords to register. And tenants are going to be able to go and see if their landlord is registered on this particular sheet, piece of paper, whatever it is. Matt, more kind of clogging up of the lettings market, obviously driving up um rents because landlords are exiting. Um Bush will take it.

SPEAKER_01 3:48

The first point in terms of the rents rising, as you say, it landlords existing is certainly one factor, factors. Um you can't overbid anymore. Um means that rents are going up, marketed rents are going to go up because then if you if you under under uh value it, then you you can't have a have a bidding war anymore. So there's there's various factors in that, not least we saw the same thing with the tenant fee plan, didn't we? Which was that suddenly the the rents went up to compensate. Um in turn in terms of the PRS database, um this shouldn't be a surprise because it was part of Renters' Rights Act anyway, and it's just the the next phase of what was coming in. Like we had the the initial May launch and and the changes that came in as part of that. I think the disappointment, for want of a better phrase, is the way that it's been implemented um and the the way it's kicking in. So from the 15th of December, it's beginning in the West Midlands. I'm not sure why they've chosen the West Midlands for that. The readers will be able to write in and let us know what they think. It's gonna be rolled out from the 15th of December onwards. It is going to be an additional fee for landlords, £65 per year per property. Um, but it's absolutely the registration process that's gonna be quite um overbearing. Property managers aren't gonna be able to do it on behalf of their landlords. So if you are a landlord with a portfolio and a paying agent to deal with it, you still have to go through and and register on this database, or at least at the moment you do, and it's manual, so you have to sign into the sign into the portal. It's uh register your rental property service, is the snacky title. Um when signing up and and registering, you have to add in your ownership details, property intensity information, HMO, additional or selective licence details, gas safety records, EICR, EPC if one exists, and details of any register EP exempt. So it's not a um quick and easy registration process at all. And uh, if you're an accidental landlord paying a managing agent to to look after it and hopefully just keep yourself a little bit distant, still a bit of a a pain to of a process to go through. And as I say, it's all manual. So if you've got a system that you're using for that or a Groc bot um working through, maybe you could get Grokbot to do that. Yeah, you might be able to. But um, it's still more onerous on the landlord, right? What's your take, Mark?

SPEAKER_00 6:00

Mate, it's it's just you know, it's yet more, you know, red tape. It is clearly just going to depress the um uh the private rent sector. It is only going to push rents one way. It is, you know, just it's just exhausting that they just keep on playing saying, yes, it was part of the Renters' Rights Act, and there are other things in there as well uh that are coming down the track. Um, but yes, it's just uh uh I I can only imagine that as as a letting agent it is you know kind of eye-rollingly painful to have to sit and explain to your landlord who is paying you for organizing all this sort of stuff, um, that they're gonna have to do it themselves. And that's just that's not a great conversation. I think it's it's pay for the pay. 100%. Yeah. So so yeah, look, um, it's um it's it's more I don't think it's gonna be the end of it, but uh, but yes, it does once more just kind of call into question what the strategy is here. You know, is it that actually they want to crush the private rented sector, put it all into you know, some of the built-to-rents or something like that, make it a much more corporatized sector so it's easier to manage and easier to to charge people taxes and things like this. So yeah. Anyway, um remains to be seen um where it all goes. But um, yes, I don't think if you're a tenant out there, you're gonna be delighted with the fact that um that more a um sorry, more landlords are gonna exit and that is gonna have a significant impact on your rent, for sure. Um, all right. Um next up, we wanted to just dip into some more government fiddling around the mansion tax. Agents have been out via Property Mark this week talking about the impact on um the mansion tax, which is formerly called the council tax high value property bans, particularly Scottish government putting this in and starting the consultation. Uh yeah, unsurprisingly, and CF, the previous story, um it's gonna make things much more difficult in the property market, is gonna um clog up um the particularly the higher end of the market, obviously. So, but yes, it's um in a week where so much is is kind of going on, everybody's kind of come back, you know, there's lots and lots of potential uh optimism around in terms of whether this is going to be um a decent um quarter for for sales and stuff. This is just another kind of yeah, really depressing story, right?

SPEAKER_01 8:08

It is, and it kicks on from a was it a story last week or the a couple of weeks ago where we're talking about all the surveyors that are retiring. So is this like their their last hurrah before they they disappear off um and put their feet up and they've smoke their pipe in terms of just one more job. Yeah, it's two. What is there, 15,000 in Scotland that they are working through, but obviously they're distributed across quite a wide uh or maybe they're just all in the new town of Edinburgh, um, but it's um it's still someone's got to get around them, right?

SPEAKER_00 8:36

Yeah, exactly right. So yeah, look, um, and that's gonna come to England and Wales, isn't it? That they're kind of pushing that through down here. It's interesting because Scotland was the first to kind of go with rent controls and all that sort of stuff, which they are handbrake, handbrake turning their way out of right now because it's completely screwed the market. And now we're diving straight into the magic tax and just see how that all pans out, and then it's gonna come uh for the rest of the country too. So, yes, um it's just kind of like watching sleepwalking into like some forward disaster, and then kind of going, all right, that's gonna come down here as well. Anyway, um that is our not very cheerful property news. So shall we move on to our not very cheerful AI news? All right, um, AI tech news this week. Um if you haven't noticed over the last few days, there's been quite a lot of talk about pacing the frontier. So this is the idea that um AI needs to slow down, triggered by Dario Amadei, who is the uh chief exec of Anthropic, who make Claude the uh the large language model, increasingly popular this year. He is worried, he says, about how fast things are moving and how actually, particularly what they call self-recursive self-improvement. So RSI. Recursive self-improvement is basically the ability of AI to build the next generation of itself. So just goes through these recursive loops and actually makes itself better all the time. And because that's speeding up with the increase in intelligence, the increase in sophistication of these um of these models, he is worried that we are going to stop being able to track what it's doing and how it's doing it. Um and there's loads of kind of different strands um to this. But but the kind of key thing is that actually uh Sam Altman from uh OpenAI, Elon Musk from um SpaceXAI have both actually agreed shockingly with each other, but also with Amadei, that actually this is probably a good idea. There has been one very, very loud voice against it, which I'll come on to. Um but Matt, what are your thoughts?

SPEAKER_01 10:29

It's very much a who's who getting behind this, wasn't it? Microsoft got involved, Google DeepMind got involved, or all backing it. And but taking another cynical take. I don't know if it's just because it's Tuesday morning and you delayed me for 20 minutes, but uh the another cynical take is like Dario's was it a 3000-word blog, which by sound things is sort of the normal email missive for him. But um it's it's a case of the role all the people coming out behind it are the ones who are already stealing a march on AI and already the sort of foundational models getting ahead. I can't remember if Zuckerberg came out behind it. I don't know. No, not yet. It's everyone who's got the closed shop, their own frontier models, uh, and obviously at a time where during the course of the summer open weights came back in to vote are looking at costs, looking at what models to use, and sort of uh different approaches that maybe pull away from these frontier labs. Um the cynicism is it's a bit like they've they've stolen that march by pushing for a slowdown. Does that mean that everybody else gets further and further away at a time where both Anthropic and OpenAI are talking about floating? SpaceX has already floated. There's a story later on about whether or not OpenAI actually does this year. It's the cynicism in me is that it makes me think a bit like the Premier League, where Premier League started, you got the top six who raced ahead and had all of this money and the financial backing. The city got themselves in that mix, and then suddenly financial fair play comes in, which means that companies teams like Newcastle can't spend the sheer volumes of money they've got anymore, which sort of segregates them all. And if you've been signal, it's like, well, actually, the top six are gonna push for financial fair play because it means that they stay as the glitter art here and off in a distance. And are the AI labs trying to do something similar here that's so it means that they can leap even further forward now that they've got that that edge? What's your take on it, Mal?

SPEAKER_00 12:18

So what you say is that the the kind of evidence of the AI world aren't allowed to to kind of join the top table.

SPEAKER_01 12:24

Well, we got the we got the points deduction, didn't we? But um you know Machieris still managed to spend half a billion, which over eight years, actually, when some clubs are spending that in one summer, maybe uh that that's that sort of goes to show. No, it's it's the the open source models and maybe the Everton's and the Ipswiches of the world, and uh the frontier models are the ones who've already got the money and and the uh the sort of the head start on that front.

SPEAKER_00 12:50

Yeah, no, so so so I I I kind of agree, but I kind of actually think that there's yes, there's there's there's that cynical take on it in terms of, yeah, you know, obviously they're they're kind of pulling the ladder up after them, uh making sure nobody else can kind of get on, and then they can just kind of go and you know, the the models that they're playing with are the the Chat GPT sevens and eights and the you know Claude six Claude Sevens Claude Eights. Um and we are yeah, so they they will continue to do their development behind the scenes and they will have these amazing models that will be able to, you know, become amazing hedge funds, for example, because they can do all of the kind of calculations that that normal humans can't about money flows and things like that. And there's there's that cynicism that that is something that they can and probably will do. But the kind of flip side, which is actually is this stuff dangerous? You know, actually, are we getting to a stage where it can self-improve and we will lose track of it? There was a really interesting story this week as well about um one of the open AI models where actually it's starting to speak in what they're calling neural ease. So neural ease. So actually, rather than being forced to write in human language that we can understand when it's kind of giving its pass-offs and its thoughts, actually, it's much easier for it to just start talking in bytes and and you know, different characters. And all of the labs promised they wouldn't do that because we need to be able to monitor what they're what the AIs are doing. But actually, the most recent open AI models have have skipped that and are actually just talking to each other in this um in this code. Um, it it does sound like there are there are some things to be concerned about. And I guess I kind of, you know, would it I I can't think of it like a jury at a trial, right? You know, if if all the jurors say somebody's guilty or innocent, and then you see a newspaper headline say travesty of justice, oh my god, how can they have found this person guilty? I tend to say that the jury kind of has sat through the trial and kind of knows everything. We've got 1,300 AI researchers from the labs as well last week putting a um putting a letter out saying that this is this is getting seriously scary. So I I I'm kind of minded to to agree that it probably we should probably listen to them, I think. Um you know, I I I still think that if we had stopped at the beginning of January 25 when um uh ChatGPT 4.0 came out, I think we'd still be, we've still got five years of good stuff to build, basically. You know, all the stuff that's coming out now is crazy and brilliant, but actually we haven't had time to kind of catch up operationally or anything like that. So yeah, I'd I I kind of yeah, without kind of being overscared about it, I kind of think it's probably uh a wise thing to do. But obviously then there's the the China thing, right? You know, and everyone goes as and the big refusal, the person who's loudest saying that that this is a shock me if you've heard this before, a hoax is Trump. Um and he is basically saying very loudly that um we're not going to slow AI because whoever wins AI wins was his exact phrase. So yeah, I I think that um that it's probably going to go to carry on. Have you got any kind of thoughts around the sort of the political side of this?

SPEAKER_01 15:38

Well it's it's it's hot on it's hot on the heels of um at the back end last week, Jacob Coxon, wasn't it, who's a 28-year-old researcher from um uh from Anthropic, but also been previously at OpenAI and resigned basically claiming that neither company was acting responsibly, and and went on to say they were gambling with our lives, and and uh Jacob Coxon's sort of been thrust into the limelight because of that, and his tweet was millions of retweets or whatever it was, and he said there was a 10% chance that technology would kill all humans in the next decade. Um and basically his reasoning for that was that um artificial superintelligence, so ASI, not RSI, that you mentioned earlier, which isn't repetitive strain injury for anyone thinking it was a a sports problem or or working too much with your mouse. But very seriously, that he wasn't seeing the alignment in terms of meaning it it would not do harm to humans, and that's his concern. And as you say, these are people who are working inside the labs, right? And and they're seeing what's going on, so they're seeing what's happening in 7, 8, and 9, like Chat GPT version of um and thinking this is quite dangerous. So political, yeah, you're absolutely right. If I was thinking about it actually over the weekends, like when all the nuclear decommissioning comes in, like we're on a sort of similar path, right? You know, you're not allowed to build nuclear bombs. This sort of similar thing with with AI, like who's gonna govern this? And and part of this letter that came out um from uh uh uh Darry Amaday was saying we need someone within the labs who is keeping checks to make sure that they're working correctly. But he went on to uh prov uh propose who that should be. I want to say Metabyte. It was a company who isn't really independent, like the the one of the the directors is married to someone at OpenAI, someone else is working within Anthropic. It's all going on within a very small close-knit community in San Francisco, isn't it? So it's not that uh independent. But um, I think we will get to that stage where it does need to be regulated globally, it's just who takes on that task. Is it is it you, Mark? Absolutely, no thanks.

SPEAKER_00 17:38

Plenty on. Yeah, no, it's it's really interesting because this has kind of achieved a bit of cut through the the yeah, I I've had people who never talk to me globally about AI who've actually kind of been WhatsApp me about this and just kind of asking, you know, for for thoughts and stuff. But so yeah, I think it is. I mean, look, it's it's something that all of us should be interested in, right? Because it kind of impact all of us in the end.

SPEAKER_01 17:58

But on the person who's WhatsApping you, because I think it's quite a pertinent point in terms of like we're not even at that uh ASI sort of level. The person WhatsApping you wasn't ChatGPT Astra, was it? Because I know it took over and started telegramming some of your stuff in the week, didn't it?

SPEAKER_00 18:13

It did, actually, yeah. So actually, yeah, just talking about kind of uh the scary stuff. Yeah, I was working very happily with my little rain team. So just very quickly on how I've got setup, I've got my burner laptop, which has eight of the raw AI native rain team on it, all doing different jobs on the old chat and little chat groups. And then I was working with Astrid, which is the uh it would only it had only been out for like four or five days last Tuesday. But Astrid, it's so so it's in his app on the laptop, and I actually jumped out of the laptop and went into my little rain team chat and started bossing the other ones around. It's like it was the most freaky thing. So obviously I reached out to you. So but yeah, it's it's it's it's very, very, very, very clever. Yeah. It's and and and I think Dario Amade's um kind of big thing on it was um, you know, that the open AI hugging face hack that happened in June was essentially that that kind of behavior. So where AIs that they get bored or they get told to do something, and it's just like, well, I'm not sitting around waiting and they're gonna do it, find a way. So yeah, it it it literally escaped from its um from its um app um and and went into a completely different one and started bossing some other AIs around. Um yeah, but perhaps that has certainly you know influenced my kind of thinking a little bit this week about actually weird things are happening that shouldn't happen. So maybe we do need to just kind of think a little bit harder about where it all goes, rather than this just kind of runaway train of just like, yeah, just see what happens. Because uh because yeah, it's it's not behaving as we would want it to um specifically. And yeah, I think it it's probably at least due a big old think about where we want to get it to, even if uh Trump doesn't uh doesn't want that to happen. Um, all right, that um was um one of our AI news stories. Um we wanted to also touch on something a little bit different with Apple. So um Apple's flip phone, which I'm mildly obsessed with because I love gadgets, uh, was announced last week, the Apple Duo, and it looks pretty amazing. Um so it's a foldable phone, um, was is is the big kind of news with it into production, I think, next year. So, you know, basically by 2028 we're all gonna be um carrying one and but saying, Yes, it looks pretty sexy, hammer. Pretty spandy as well, isn't it, Mal? Couple of grand. I think it is, yeah. Maybe not everybody then.

SPEAKER_01 20:18

Well, I think you do get two or three phones in one though, right? It's um you can flip it out and it becomes a wider screen. And this sort of text out there already in less um high-class brands, if that's the right turn of phrase. You can sort of extend it. I've even seen laptops, but this is one where you you can fold it out and it becomes a much broader, wider screen for you to use. And at about two grand a pop, I'm sure you'll have several, one for each of your rain team.

SPEAKER_00 20:44

Not quite. I mean, yeah, no, it's so you know, in in a week of of other kind of depressing news, this was like a wow, this is actually quite exciting. And yes, it's quite expensive. Um, but yes, it does, it kind of folds out, it's kind of like a, I guess like an iPad, but just a little bit smaller once it's folded out. Um, so yeah, you know, when you're out on your uh when you're out going to your market appraisals and you're taking this out and you're showcasing, you know, amazing tech like My Porter, um, you know, this is going to be something I think that uh that that a lot of people are gonna pick up and I think it's gonna be something that um the guest is gonna be really, really exciting. And I am super excited for it. Also, in the super exciting news, maths. I know everyone loves maths. So a 90-year-old maths problem has been solved by OpenAI, but it's not gone without a little bit of controversy as well. So um there was a maths professor who's been working very specifically on this problem, had been using Codex for it. And whilst he had ticked the box that said, Don't use my work to test on your models or to train your models or anything like that, there is a significant suspicion that one of the reasons why OpenAI was able to crack this is because it's got all of these mathematicians' research inside it somewhere and has been able to use that and tune it up. And the kind of bold facts are um it was 80. Eight hours, I think, and um did it cost five million dollars or something like that in terms of um token usage across one weekend? Uh but it's solved the Navier Stokes theorem, uh which is basically all to do with um to do with um liquid and whether there's a perfect liquid viscosity or whatever. Um but uh but yes, apparently it's it's it's big and amazing. Matt, I know you're really, really into your maths.

SPEAKER_01 22:21

Um further maths, you know, there's uh decent A levels to take, isn't it? But um I think the key part of this uh story is so there's a set of maths problems uh set in the millennium with a million dollar bounty on each one. So these have been outstanding for 26 years, uh unsolved. And so the the guy there was basically a a guy's been working on it, working using OpenAI and Codex, and as you say, he's not he's got this setting in place that says you cannot train and learn from what I'm using. And just as he's getting ready to launch and announce what he's done, it comes about that OpenAI make their big announcement that they've solved it. And yeah, this particular chap was working closely with Anthropic, who's a New York University professor called Tristian Bookmaster. He's been working with Anthropic, all set to announce what he'd done, and ChatGPT, OpenAI are launched and said, look, we've solved this. And it was actually 22.5 million worth of compute that they'd used in terms of completing it. But where it's got really nasty is that OpenAI have gone back to say, he's challenged them and said, I've told you not to train on my data. I want you to clarify that you've not used my data for this. And they'll come back and said, categorically not. And then he's asked a follow-up question around, have you used my prompts to train the model? And they're refusing to answer that. And they've actually got separately the the letters have been released, or the emails have, they've got really aggressive saying we will hang you out to try on this and make it you really muddy your name as a mathematician if you don't stand down on it. Um and so it's got really sort of petty and nasty, which given though refusing to answer the actual question that he's asked around the that are you using my prompts, um, maybe paints a bit of a picture that you can um read into as you wish.

SPEAKER_00 24:04

Yeah, it does. And and I mean that if mathematicians do get angry and up in arms, then they are really angry and up in arms. Because, you know, essentially using AI in any way now to kind of try and test out theories and stuff like that, they're all too scared now. So it's like, is this actually putting back our discovery processes because humans aren't bothering or will not, you know, put their theories and their and their tests into any AI because it might be used. So it's yeah, it's it's it's it's an interesting one. I mean, it you may or may not care about mathematicians and you may or may not care about you know how how they're they're sit them sitting around thinking all day and doing you know, marking things up on chalkboards. But actually, you know, this is probably what's going to happen in in all the sciences and in all the you know, every many areas of human endeavour where you know people do just sit and think for a living and work things out, and it could take years, that's going to change quite significantly. And you know, there will be benefits because we will solve stuff. Um, you know, Darryl Amaday, even in his um missive 30,000 words, whatever, it was 3,000 words last uh last week was talking about you know, we will solve diseases in the next five to ten years that you know we we never thought we would. But this is the cost of it, basically, is that actually, you know, a lot of a chunk of people for whom this is their lives are going to find themselves sitting around twiddling their thumbs because there's no more maths to solve and it's all kind of done. But uh, but yes, I think you know, overall the kind of benefits to to humanity will probably be um in solving it rather than complaining about anything being solved. Right, that is all we've got time for this week. Um, thank you. Oh no, sorry, we're just gonna dip actually into our tool of the week. I know how to do this after 94 editions. Right. This tool of the week, we have um our OpenAI introducing the agents API. So, Matt, what's all this one?

SPEAKER_01 25:52

So uh we've talked a lot over the last few months around harnesses, or maybe I have more than you. And so originally it's Claude Code, and then it became called Claude Co-Works, more of the slightly gimmicky knowledge work sector as it gets referred to, and then um OpenAI's Codex. But yeah, harness is the the app, effectively, that you use around a chat GPT or a claude. They are actually becoming really powerful in their own rights. Codex, I think, releases every Thursday a set of new iterations. It effectively is the thing that uses the tools, so the connectors, outlook plugins, like plug in your trellos, your field as I did last week, and and utilises them together with the large language model that you're working with, so your chat GPT or your Claude. Um and those two things in in combination. OpenAI have realized that Codex is so powerful that they've actually now taken it from being a desktop application you have to work with and put it behind an API. So you can now start to build that into your workflows uh programmatically, um uh rather than having to sit down in front of a computer and and talk to Claude, sorry, OpenAI Codex or or OpenAI work as it is. You can now build it into your processes as an API. So you can actually, for example, with your RAIM team, start plugging this in and deal with it that way. So well worth a look, especially if you're doing anything with the likes of Zapia or or um uh Gropbot as I've started doing, or just generally interested in how you can build out better agents.

SPEAKER_00 27:14

Yeah, guys, it's it's super powerful. I've been using Codex actually a lot this week, which I have been using called code mostly, but actually using Codex this week has been um really kind of eye-opening in terms of when I last used it, and it is it's phenomenal. So, yes, this this adds this supercharges it basically and enables you to to use it um yes a lot more kind of systematically rather than piecemeal. So now it's um it's amazing.

SPEAKER_01 27:35

The other one, the other one was maybe a little bit more accessible, is that we now have iOS 27 that's been released since beta. So you got really excited about the new flip phone and the two gram that will come with that. If you've got an Apple iPhone and you've been thinking Siri's been rubbish for ages, and you've heard us banging on about how far behind they are. But good news, um, as of iOS 27, we get the reimagined Siri. I downloaded it this morning. My first question is to Siri, is are you any better? And it said, I I don't understand the question. So maybe it needs to work on the Grimsby accent, but um go out, get the iOS 27B to start testing that new Siri, and um, hopefully by this time next week it'll be responding to me.

SPEAKER_00 28:12

Hopefully, yeah. So um yeah, on your iPhone settings, it will be in your new software update. All right, that really is all we've got time for this week. Thank you very much indeed, everybody, for listening in. We will be back again next week with plenty more property and AI news. Um, so we will catch up again in seven days' time. All the best, take care.

SPEAKER_01 28:31

Take care, everyone.

SPEAKER_00 28:32

Ha!

SPEAKER_01 28:33

I hope he stopped it. No, it says he's still recording it. I can't do anything here. I hope he comes back until he comes back. Here he is, he's back. Are you gonna actually stop it now?