Latest / Second Nature / E201 Tom George: Why It's Time to Let AI Take Over Your Business!

Transcript

E201 Tom George: Why It's Time to Let AI Take Over Your Business! · Second Nature

• 0:06
  SPEAKER_03: Tom, when did you first encounter AI You know it's weird?
• 0:17
  SPEAKER_00: I can almost not remember a time when we weren't on that track.
• 0:23
  SPEAKER_01: I think um I mean it was GPT-4, it wasn't before from for me.
• 0:32
  SPEAKER_01: Really before?
• 0:33
  SPEAKER_00: Yeah, interesting.
• 0:34
  SPEAKER_02: I I think um I think so.
• 0:38
  SPEAKER_03: I three was COVID, so three literally launched just into lockdown period time with COVID.
• 0:45
  SPEAKER_03: So four would have been coming out of the back of that, I would have thought.
• 0:48
  SPEAKER_02: I know that I was super focused on on local models, um even back then.
• 0:54
  SPEAKER_02: So like um maybe it was just a great way to justify getting an incredible gaming PC.
• 0:59
  SPEAKER_02: Um thanks, Jane.
• 1:02
  SPEAKER_02: Uh but uh I remember, yeah, I mean, sort of going actually from Ethereum mining um at one point uh and then and then kind of uh you know switching focus onto that stuff.
• 1:13
  SPEAKER_02: Um so I think um yeah, I mean not crazy, crazy early.
• 1:19
  SPEAKER_03: Um what were the leading um actually I want to do a bit on crypto?
• 1:23
  SPEAKER_03: Where are you on crypto at the moment in terms of your perspective?
• 1:25
  SPEAKER_02: I d I don't really.
• 1:26
  SPEAKER_02: I mean look, I think um there could be some interesting use case of distributed distributed ledgers, um, but they're actually few and far between.
• 1:35
  SPEAKER_02: Um Is Heath gonna adopt some of those principles?
• 1:39
  SPEAKER_02: Probably not.
• 1:41
  SPEAKER_02: Right.
• 1:41
  SPEAKER_02: So I think with with with with Heath, um, which is the ledger we're building, um ultimately you don't really need the distributed nature of it.
• 1:50
  SPEAKER_03: But um I don't think it's another level of independence and credibility if we did host it in some kind of distributed environment.
• 2:00
  SPEAKER_02: I mean the techie in me says no.
• 2:02
  SPEAKER_01: Okay.
• 2:02
  SPEAKER_02: And and and Phil's got my back there, you can ask him, he'll say no.
• 2:05
  SPEAKER_02: Fine.
• 2:07
  SPEAKER_02: I think I think I mean I think this that it's like it's some very small use cases where it can make sense, and probably there's there's there's more than I think, but it's not as many as as all that, right?
• 2:17
  SPEAKER_02: So I think what's really important for us is you know on that ledger side is just to make sure we kind of have something that that that we've built and can continue building that basically um underpins kind of our business and makes sure that we're that we're transparent in everything we're doing, right?
• 2:34
  SPEAKER_03: Um we haven't done this before.
• 2:37
  SPEAKER_03: Um hey Alan, can you please tell me the total net value of crypto transactions in US dollars um in the calendar year 2025 or 26, please?
• 2:50
  SPEAKER_03: I reckon this is in the hundreds of millions.
• 2:53
  SPEAKER_03: I I I just think the scale I saw on a podcast the other day.
• 2:57
  SPEAKER_03: Oh, we will be yeah.
• 2:59
  SPEAKER_03: No, do you know what?
• 3:00
  SPEAKER_03: I'm thinking of agent-to-agent crypto transactions.
• 3:04
  SPEAKER_03: Okay, yeah, yeah.
• 3:05
  SPEAKER_03: Yeah.
• 3:06
  SPEAKER_03: Uh I I I do, Alan.
• 3:08
  SPEAKER_03: I mean the net value of transactions conducted in cryptocurrencies, please.
• 3:12
  SPEAKER_03: And then can I see it by agents as well, please?
• 3:15
  SPEAKER_03: Um, this is Alan's first appearance on the podcast.
• 3:17
  SPEAKER_03: For those of you out there, Alan, named after Alan Turin, uh, is my OpenAI dot that has been a game changer for us, which we'll we'll come on to.
• 3:24
  SPEAKER_03: Um, I think mine was GPT 3.
• 3:27
  SPEAKER_03: So we were, and it first made an appearance.
• 3:30
  SPEAKER_03: We did this incredible activity when we were two or three years into OC.
• 3:35
  SPEAKER_03: We'd just we've gone to that next level of sophistication as a business, and we're like, what is our next strategy?
• 3:40
  SPEAKER_03: So we'd scaled up and we did this awesome activity that our our chairman Benny asked us to do, where we sat down and we interviewed every board member and a senior stakeholder across a whole range of things.
• 3:50
  SPEAKER_03: At the same time, I was alerted to GPT-3, I think, and used that as a way of synthesizing hundreds of pages of interviews and notes and things.
• 4:00
  SPEAKER_03: And it was an absolute game changer, secret source for me to be able to do that and and process through it.
• 4:05
  SPEAKER_03: So actually, GPT-3 and an early model of AI was a co-author on the next strategy of OC.
• 4:12
  SPEAKER_02: You got I've you got me curious.
• 4:13
  SPEAKER_02: I'm gonna look later on, like when my uh account was created.
• 4:16
  SPEAKER_03: Yeah.
• 4:17
  SPEAKER_02: Um, because I I mean I don't I can't remember.
• 4:19
  SPEAKER_02: It's just it feels like it's been with us for so long now.
• 4:22
  SPEAKER_03: Um and what what were the open weight, lightweight models that you were using at that time?
• 4:27
  SPEAKER_03: What would the leaders have been then?
• 4:29
  SPEAKER_03: Um I think they were um man, that is a great question.
• 4:34
  SPEAKER_02: I mean, at the moment, um I'm loving the Quen 3.8 uh uh world.
• 4:41
  SPEAKER_03: So people worry about that because you hear the whole hold on a second, where are we on the security side of it?
• 4:46
  SPEAKER_03: And it is quite a Chinese model as well.
• 4:48
  SPEAKER_03: Yeah.
• 4:48
  SPEAKER_03: So people are concerned about that.
• 4:50
  SPEAKER_03: Help help um give that context.
• 4:53
  SPEAKER_03: What what do you think about when you're using those type of models?
• 4:57
  SPEAKER_02: Well, I mean, first up, if you're just running a text model and you're not, you know, equipping it with tools or anything like that, it's just running on your machine.
• 5:04
  SPEAKER_02: Um, and that, by the way, is the most amazing thing because it's like you're not dependent on any provider, you're not dependent on anyone, it's on your machine running.
• 5:13
  SPEAKER_02: And that's like that's incredible.
• 5:14
  SPEAKER_03: And from a cost perspective, unparalleled, right?
• 5:17
  SPEAKER_02: Yeah, I mean, the the of course you've got the the capital cost of buying your amazing game PC or whatever it happens to be.
• 5:24
  SPEAKER_02: Um, but then basically it's running, um, it's running locally, right?
• 5:28
  SPEAKER_02: So um it's quite you know, it's quite fun.
• 5:30
  SPEAKER_02: I mean, you know, I I was telling you I've got it running on my on my on my MacBook, right?
• 5:35
  SPEAKER_02: And then you can it literally sounds like the fan is just going, it's like sounds like it's about to take off, you know.
• 5:39
  SPEAKER_02: But then it's it's spitting tokens out and it's talking back.
• 5:42
  SPEAKER_02: And I and I think um, I d I don't know the the um the kind of um scoring of it off by heart, but I I would say that it feels certainly like that GPT-4 kind of level model, right?
• 5:57
  SPEAKER_02: But it's way smaller, it's like 27 billion uh parameters.
• 6:00
  SPEAKER_02: It's so smaller that that that one, right?
• 6:02
  SPEAKER_03: So uh for the viewers out there, 28 trillion in crypto transactions in the calendar year globally 2025.
• 6:11
  SPEAKER_03: And uh where they can find a track record, uh$52 million of transactions agent to agent in 2025.
• 6:21
  SPEAKER_03: That's crazy.
• 6:22
  SPEAKER_03: Which it's coming so fast.
• 6:23
  SPEAKER_02: 27 trillion.
• 6:24
  SPEAKER_02: I mean, that must be like that's almost like the same GDP as the US, right?
• 6:27
  SPEAKER_02: Or something like that.
• 6:29
  SPEAKER_03: Huge, isn't it?
• 6:30
  SPEAKER_00: Well it's changing.
• 6:32
  SPEAKER_00: Yeah.
• 6:33
  SPEAKER_00: Yeah.
• 6:34
  SPEAKER_02: So um I think uh the local model is amazing because obviously you don't have you don't let's say one of the things that um, you know, if you if we're discussing different models to use, um, that our the legal team would be like freaking out a little bit about is like, okay, where is it hosted?
• 6:50
  SPEAKER_02: Are they using, you know, are they using our data to train it?
• 6:53
  SPEAKER_02: Um those those are legit things that we need to think about, right?
• 6:56
  SPEAKER_02: That you know, um, because we've got incredible IP that we've built as as oxygen conservation, we need to protect it, right?
• 7:03
  SPEAKER_02: And I think that's one of the things that is, you know, as we move forwards, it's going to be a constant kind of conversation is like, okay, is this model, you know, do do we have a kind of surety on on what they're doing with our data?
• 7:15
  SPEAKER_02: Of course, those local models are amazing because you know exactly what who's doing what with it, because it's you.
• 7:20
  SPEAKER_02: Um, I think the broader question about then um enabling local models or enabling any model with tools that it can then use is of course, okay, how how much do I trust this thing?
• 7:34
  SPEAKER_02: Right.
• 7:34
  SPEAKER_02: And I think that's something that's that's ongoing, right?
• 7:36
  SPEAKER_02: As a discussion, you know, kind of uh in a way.
• 7:40
  SPEAKER_03: Almost counterintuitive in a way, isn't it?
• 7:41
  SPEAKER_03: Because people assume, oh, the bigger, the more high profile the model, the safer that will be to use.
• 7:46
  SPEAKER_03: But the nuance is really important there.
• 7:48
  SPEAKER_03: What does safe mean?
• 7:49
  SPEAKER_03: And is it you know how are they using data?
• 7:52
  SPEAKER_03: You know, you see a lot of um arguments now that people who have a novel idea for a startup can use one of the big mainstream models and find that that idea is incorporated into what they offer the next day.
• 8:05
  SPEAKER_03: I mean a few weeks ago, Jeff, the sort of decision-making model, was the absolute sort of darling of the sector.
• 8:12
  SPEAKER_03: And within a few weeks, OpenAI and Claude have dropped their own version of it that just And Cloudflare, my my my tech favorite at the moment in that respect.
• 8:19
  SPEAKER_02: So amazing that they, you know, I think I think one of the things that um speed of ability to develop is is is off the hook, right?
• 8:29
  SPEAKER_02: Because with these big models, and especially companies that can point a thousand agents at a problem, um, you're capable of of catching up, you know, kind of overnight with things like that, right?
• 8:38
  SPEAKER_02: Um it which is which is quite freaky.
• 8:41
  SPEAKER_02: So I think we we do have to think about what, from our own IP perspective, what we're comfortable sharing with the world and what we want to be a little bit more careful about.
• 8:50
  SPEAKER_02: That is a beautiful segue into who you are and what you do, Tom.
• 8:54
  SPEAKER_03: Would you like to tell people out there?
• 8:55
  SPEAKER_02: My name's Tom George.
• 8:56
  SPEAKER_03: I am head of technology for oxygen intelligence.
• 8:59
  SPEAKER_03: In tell us how you got there.
• 9:00
  SPEAKER_03: So, and the fun bit is going back to the start.
• 9:04
  SPEAKER_03: So, um, how on earth did you um fall in love with technology and make it a career?
• 9:10
  SPEAKER_02: So when I was a kid, I was a nightmare, right?
• 9:13
  SPEAKER_02: So when I was a little kid, I was a nightmare because um I want to know how everything works.
• 9:19
  SPEAKER_02: So um what would normally happen is, because my dad was mostly at sea, my mum I'd wait for my mum to go.
• 9:25
  SPEAKER_02: Your dad was a fisherman.
• 9:25
  SPEAKER_02: My dad's a fisherman, exactly.
• 9:27
  SPEAKER_02: From well, near near, well, uh near Land's End.
• 9:30
  SPEAKER_02: So um Pemberth Cove.
• 9:32
  SPEAKER_02: Um, and so I'd wait for my mum to go out to, you know, to the town or whatever.
• 9:38
  SPEAKER_02: And then as soon as she'd gone, whatever new piece of thing that was in the house, I'd take apart and try and understand how it worked.
• 9:44
  SPEAKER_02: So, yeah, especially electronics and stuff like that.
• 9:47
  SPEAKER_02: They got something new, like an alarm radio or whatever.
• 9:50
  SPEAKER_02: Straight away, I'd be like, as soon as I gone, like getting the screwdriver out and taking everything apart, just trying to try and understand how it works.
• 9:56
  SPEAKER_03: Just like also put it back together.
• 9:58
  SPEAKER_02: Yeah, normally with like one or two less screws, because you know, as a kid, you don't know where everything's gone.
• 10:03
  SPEAKER_02: Um, but um definitely kind of that started.
• 10:06
  SPEAKER_02: I think I just always wanted to know how stuff works and how to build stuff.
• 10:09
  SPEAKER_02: So that was like um super formative in that kind of in that first part of like understanding, okay, I really want to do, you know, things that are kind of tech related, right?
• 10:18
  SPEAKER_02: And then from there, um yeah, I got into um a lot of like networking side of stuff.
• 10:24
  SPEAKER_02: Um and my first jobs were in that.
• 10:26
  SPEAKER_03: And where was that to start with?
• 10:28
  SPEAKER_02: Well, I went for a job in Cambridge, this company called UUNet.
• 10:32
  SPEAKER_02: Um, and they were really, I had a really good time in the interview, I guess.
• 10:36
  SPEAKER_02: And they like they offered me the job that said that you can have a better job, but it's in Amsterdam.
• 10:40
  SPEAKER_02: Uh, and so I was like, Yeah, absolutely sign me up.
• 10:43
  SPEAKER_02: I'm like 22 years old, 23 years old, or whatever.
• 10:47
  SPEAKER_02: Um, and so off I went.
• 10:48
  SPEAKER_03: We're recording this in the middle of our intern recruitment procedure, and that's advice for everyone out there.
• 10:53
  SPEAKER_03: Take the opportunity, take the difficult one, take the adventure, take the challenge, especially early in your career.
• 11:00
  SPEAKER_02: No, no, no, just anytime, Rich.
• 11:03
  SPEAKER_02: I don't great advice, man.
• 11:04
  SPEAKER_02: I like fair, you know, because I think um, you know, I think it's so easy to not do, right?
• 11:11
  SPEAKER_02: Love it.
• 11:11
  SPEAKER_02: Keep taking us through the steps.
• 11:13
  SPEAKER_02: Yeah, um, it's I mean, it sounds like I'm it's well, I don't know, it's boring, but um, yeah, then we we we made some friends and we're like, hey, you should move to Australia.
• 11:22
  SPEAKER_02: We're from Western Australia, it's an incredible place.
• 11:24
  SPEAKER_02: Um, so we're like, hell yeah, that sounds like fun.
• 11:27
  SPEAKER_02: So we basically saved a bit of money.
• 11:29
  SPEAKER_03: For those of you unsurprised, Ellie's in the room and punching the air in the back of shot wanting the Australian expansion.
• 11:34
  SPEAKER_02: In Australia, well, us is amazing, and like it's obviously it's a young country and is amazing to live there.
• 11:41
  SPEAKER_02: We were really, really lucky.
• 11:43
  SPEAKER_02: Um, so yeah, kind of landed on our feet.
• 11:46
  SPEAKER_02: I got I went for a like a six-week contract uh and then was there like six years later with the same company, and it was my first kind of exposure to to big corporations, right?
• 11:57
  SPEAKER_02: So the other stuff have been kind of smaller, you know, ISPs and things like that.
• 12:01
  SPEAKER_02: And it was that was the first the kind of first step.
• 12:03
  SPEAKER_02: It's also coincided with having then at some point my first kid and like Tom, you better take things seriously.
• 12:09
  SPEAKER_02: So um, yeah, then sort of work for this for this mining company um out there for for Rio Tinto and for BHP.
• 12:16
  SPEAKER_02: The big thing for me, and like kind of interesting parallel in a in a kind of weirdly opposite way, um, was just getting out and seeing what those mining companies are doing, um, leveraging capital, right?
• 12:29
  SPEAKER_02: If you I I was lucky to be the um the face of the IT department, more or less, um, for for at Rio, and just got to go and like every week fly out to different uh mine sites, right?
• 12:41
  SPEAKER_02: So um Marindu, where Dave Dave Keir was working, right?
• 12:45
  SPEAKER_02: Um so Marindu and Tom Price, Parabadu, Cape Lambert, all these amazing places, some of them uh like literally in the Caragini National Park uh in Northwest Australia, right?
• 12:55
  SPEAKER_02: And you get out and just the scale of the things they're doing is is unthinkable, right?
• 13:02
  SPEAKER_02: Um, Tom Price, right?
• 13:04
  SPEAKER_02: It's called Tom Price because of Mount Tom Price, except you go there, there's no mountain, it's gone.
• 13:10
  SPEAKER_02: They've literally dug the whole thing up and it's now like an 800 meter hole, like deep.
• 13:15
  SPEAKER_02: So you imagine how big that is.
• 13:16
  SPEAKER_02: If you think about how how big the pit has to be, it's gone.
• 13:20
  SPEAKER_02: It's like being shipped to China, right?
• 13:22
  SPEAKER_02: And it's just unbelievable the scale that you can that you can impact with kind of big capital.
• 13:28
  SPEAKER_03: Um, so that was like that was mind-blowing.
• 13:30
  SPEAKER_03: I think for people that haven't seen large-scale heavy industry, you can't comprehend it unless you've seen it on the ground.
• 13:36
  SPEAKER_02: It's epic, not in the positive sense, just in in in the kind of or or like awe-inspiring in terms of the size of it that's happening.
• 13:43
  SPEAKER_02: Um so yeah, that that was like crazy to see.
• 13:47
  SPEAKER_02: And I think that's something that with those big corporates, and then obviously I made a jump, um, had a couple of kids, um, got sentimental one night, counted on how many fingers, how many times kids would see their grandparents if we weren't moving back somewhere closer.
• 14:02
  SPEAKER_02: So kind of figured, hell, let's move back to Europe.
• 14:05
  SPEAKER_02: And I tell you what, there's a guy called Eden, uh Edin Mustabekovich, who I used to work with.
• 14:10
  SPEAKER_02: Right now.
• 14:11
  SPEAKER_02: Right.
• 14:11
  SPEAKER_02: Um, who's a who's a fantastic uh enterprise architect in Perth.
• 14:16
  SPEAKER_02: And the the two of us were having lunch one day, and I was explaining, I was thinking about it.
• 14:21
  SPEAKER_02: So just fucking do it.
• 14:22
  SPEAKER_02: Go for it, go for it.
• 14:23
  SPEAKER_02: Uh, and so that's kind of cool, you know.
• 14:25
  SPEAKER_02: It's like that I remember that um distinctively.
• 14:28
  SPEAKER_02: So yeah, just um got back then from from Australia to Germany.
• 14:33
  SPEAKER_02: Um, I went for my interview there with with Volkswagen, BW, then different thing again.
• 14:40
  SPEAKER_02: You worked in government, VW, like 20% owned by the state of Lower Saxony, um, huge organ, like massive organization, 600,000 people, 10,000 people in IT, right?
• 14:53
  SPEAKER_02: To think about the scale that we are, like at Oxygen uh conservation, you know?
• 14:57
  SPEAKER_02: So it's just so, so big.
• 14:59
  SPEAKER_02: And and the same deal, like leveraging big capital, building big things, right?
• 15:04
  SPEAKER_02: Making making money that way.
• 15:06
  SPEAKER_02: Um and then yeah, with with uh with with VW, I was like massively lucky because I had uh a guy called Peter Garzarella, um, who I work for, who kind of realised that I'm like a bit of a fish out of water in Wolfsburg.
• 15:23
  SPEAKER_02: Um because I he because I I like building stuff.
• 15:27
  SPEAKER_02: Wolfsburg, for those who haven't been, uh it's a town of about 120,000 people in in like central northern Germany.
• 15:34
  SPEAKER_02: Um and I guess a hundred thousand of those work for Volkswagen, right?
• 15:39
  SPEAKER_02: So it's like um the city just lives and breathes it.
• 15:44
  SPEAKER_02: Um, and it's the headquarters of so much concentration of kind of competence and and power and bureaucracy and politics and that sort of stuff.
• 15:54
  SPEAKER_03: And it's really um not my not my kind of I think the read from afar is that the idea of being creative or innovative in that environment just seems impossible.
• 16:05
  SPEAKER_03: Yeah.
• 16:05
  SPEAKER_03: Because presumably every IT decision and purchase is made at such a scale, there is a system and process to it.
• 16:12
  SPEAKER_03: So you can't just go buy the new creative thing that's available.
• 16:15
  SPEAKER_02: Yeah, I mean, r really that's it.
• 16:16
  SPEAKER_02: And we we we were really lucky.
• 16:18
  SPEAKER_02: So a couple of things happened.
• 16:20
  SPEAKER_02: Do you remember the dieselgate thing that happened to I think everyone's familiar with that?
• 16:24
  SPEAKER_02: Yeah, well, take so a couple of things, right?
• 16:26
  SPEAKER_02: Number one, PW people are so passionately, they love BW that people were devastated, like literally devastated.
• 16:33
  SPEAKER_02: I was in in the office the day that it all came out, and people were crying in the office, right?
• 16:37
  SPEAKER_02: Which is like that's how passionate they are, which is quite a cool, I think it's quite cool, you know, in general.
• 16:43
  SPEAKER_02: Um, but it it just opened the door for me because uh my boss at the time said, Hey, could you build um some websites and apps and stuff for our customers to like figure out where they've been hit?
• 16:54
  SPEAKER_02: And it was like a golden ticket because suddenly it's like, get out of the way, I'm gonna build stuff.
• 16:59
  SPEAKER_03: And you're working on the solution to a problem.
• 17:01
  SPEAKER_02: Yeah, right.
• 17:02
  SPEAKER_02: So it was it was really cool like that.
• 17:03
  SPEAKER_02: And suddenly so I was then in my happy place, like instantly, um, and was able to kind of build you know some teams up and build stuff.
• 17:10
  SPEAKER_02: Um, and that was really cool.
• 17:11
  SPEAKER_02: And through that, I think I kind of then people realized like, okay, this is someone that needs to be given space to build.
• 17:16
  SPEAKER_03: Talk to us about that stepping away from the corporate world and the conversation at home about you know, time for something different.
• 17:22
  SPEAKER_00: I basically so kind of you're saying was there a specific point where I was like taking a decision or not?
• 17:30
  SPEAKER_00: And I think um it was a point where I I had figured out I want to do something that is that is that that is furthering a mission that's better.
• 17:42
  SPEAKER_02: Um by the way, I love working for Volkswagen, I think it's a cool company.
• 17:45
  SPEAKER_02: Great talented.
• 17:48
  SPEAKER_02: No, no, I'm genuinely talented, super talented individuals, um, passionate about what they do.
• 17:53
  SPEAKER_02: Um and I just I just felt like I wanted to work in a team that's working for something that was that was paying back perhaps a little bit more into into the kind of um into the world.
• 18:05
  SPEAKER_02: Um and yeah, then I came across the OC thing.
• 18:08
  SPEAKER_02: I was a little bit skeptical because it seemed a bit too polished.
• 18:11
  SPEAKER_02: Um and I was a little bit unsure about like, okay, is this a real thing?
• 18:15
  SPEAKER_02: Um we get that a bit.
• 18:17
  SPEAKER_02: Yeah, I mean, you know, I think But should we didn't make it look worse?
• 18:20
  SPEAKER_02: No, I don't know.
• 18:21
  SPEAKER_02: It's it's it's it's very difficult because if you put everything down on paper, if you like not obviously not on paper, but you know, I mean, if you look at everything, it it you're extremely transparent, or we are extremely transparent, better said.
• 18:33
  SPEAKER_02: Um, but as an outsider, you're like, this is very glossy, what's going on?
• 18:39
  SPEAKER_02: Yeah.
• 18:40
  SPEAKER_02: And I think um through that process, basically then I was doing more research on like, okay, who who is OC?
• 18:46
  SPEAKER_02: What they what do they do?
• 18:47
  SPEAKER_02: Um, and that was then super attractive for me, right?
• 18:50
  SPEAKER_02: That kind of marrying of um of the the capital side of it.
• 18:57
  SPEAKER_02: So that was when I was looking, you know, when I was thinking about it, the the capital side of it and the the tech the tech, um just super powerful because I think those are two things that you can move a lot of stuff with.
• 19:08
  SPEAKER_02: You know, I was I was explaining about the kind of the the whole the hole in the ground in to uh Tom Price and and and that parabadoo and you know, everywhere you go, there's these massive, like massive, massive impact on the on the landscape.
• 19:21
  SPEAKER_02: And then you guys are like talking the opposite, which is like, okay, we can leverage capital to to do that.
• 19:25
  SPEAKER_02: And that was that was something that was yeah, super inspirational for me.
• 19:29
  SPEAKER_02: And kind of when I was thinking about joining, that was kind of paying into that, I think, a little bit.
• 19:33
  SPEAKER_03: On the tech side of things, knowing what you know now, how how much did you expect what we were going to do around tech and data and information?
• 19:42
  SPEAKER_03: And how much has that sort of surprised you in terms of where we are and where we're going?
• 19:48
  SPEAKER_02: I was literally, I mean, before I joined, I was literally like in GitHub looking at different um uh like machine learning um uh open source algorithms for for figuring out how much carbon's in a tree.
• 20:02
  SPEAKER_02: So like I was so excited to join and then and kind of to start working in that sphere and trying to understand.
• 20:08
  SPEAKER_02: Um I think um we got some incredible talent in that respect.
• 20:12
  SPEAKER_02: And kind of I think it's really interesting to be listening and like Noelle, yeah, that's kind of a good way to do it, but actually there's you know, we've got some other ways as well.
• 20:19
  SPEAKER_02: I think that that's a real excitement for me.
• 20:22
  SPEAKER_03: Did we do a good enough job of explaining to you that yes, we're a conservation company and an asset manager, but we're also an AI company?
• 20:30
  SPEAKER_03: Did we explain that well enough at the start, or have you had to realize that the time you've been with us?
• 20:35
  SPEAKER_02: No, I think I think I've primarily realized it since I've been with um with OC.
• 20:41
  SPEAKER_02: I think um I'm sort of glad because in a way it would have been even more uh like uh nerve-wracking joining, I think.
• 20:48
  SPEAKER_02: Um but on the other hand, I think um I think the constituents of what we have are like incredible.
• 20:55
  SPEAKER_02: Um and I think, you know, if you look at the tech side, without just leaving aside the AI part for the moment, but if you look at the tech side, what we're going to be capable of doing, you know, by kind of with the pipelines that we're building and this type of thing is is going to be incredible.
• 21:09
  SPEAKER_03: So where are we going next, Tom?
• 21:11
  SPEAKER_03: What are you excited about?
• 21:12
  SPEAKER_03: Let's stay within the business in the first instance.
• 21:13
  SPEAKER_03: What are you excited about us doing next?
• 21:16
  SPEAKER_02: Um more data, more real-time data.
• 21:19
  SPEAKER_02: I think that's that's stuff that's so cool, right?
• 21:22
  SPEAKER_02: I think that because um, you know, the I guess one of the what I've understood so far, one of the kind of things we're trying to do is is demonstrate the value of of natural capital and uh and and nature.
• 21:34
  SPEAKER_02: And I think the more kind of real-time data that we have that demonstrates that, brings it to life, I think is super cool.
• 21:42
  SPEAKER_02: So I'm sort of really excited about that, really excited about the geeking out side of that as well.
• 21:46
  SPEAKER_02: It's just I think that's that's something that's really exciting.
• 21:49
  SPEAKER_02: Um the product side, I think Heath's gonna be revolutionary.
• 21:53
  SPEAKER_02: Can you explain what that is?
• 21:55
  SPEAKER_02: Yeah, I mean, I think um fundamentally basic.
• 22:00
  SPEAKER_02: Stuff in in some respects.
• 22:02
  SPEAKER_02: You know, we we need the ability to make sure that as a business we're capable of, you know, right from the beginning of the of the kind of uh projects that we have for for improving nature, that we're tracking exactly what we have, what we're building, what state it's in, um, and and who we're you know, we're selling uh premium premium carbon and and those types of uh BNG uh type things to.
• 22:26
  SPEAKER_02: Um so we just have that whole traceability through.
• 22:29
  SPEAKER_02: Um I think that's one of the essential things.
• 22:31
  SPEAKER_02: Um it's one of the things that in Oxygen um we're so good at is that transparency, right?
• 22:36
  SPEAKER_02: And I think that's one of the things that um ledger's really good for showing, you know, that kind of like literally blow by blow what's going on.
• 22:44
  SPEAKER_03: Um and that's cool.
• 22:46
  SPEAKER_03: I I think what's exciting is what that unlocks.
• 22:48
  SPEAKER_03: So one thing you've been able to do really quickly is build or or create an element of something that unlocks lots of other things, be it the CRM side of things, just functionality for us.
• 22:59
  SPEAKER_03: So you've been able to, you know, work with the legal teams and things, but give us all more access and more freedom and ability to make choices on things, and and that's been transformative.
• 23:08
  SPEAKER_03: And we touch wood, we haven't had a problem yet.
• 23:11
  SPEAKER_03: We will, because you always do when you're trying new things, but we haven't so far.
• 23:14
  SPEAKER_03: What I'm excited about the the Heath side of things is because we're gonna have that digital individual tree and Peatland and spatial-based ledger, we can then map that to an API that can allow direct-to-customer and business partnerships and different way of almost democratizing that premium natural capital, which just isn't possible without that ledger data set.
• 23:36
  SPEAKER_02: I mean, you you you need the whole the whole piece, right?
• 23:38
  SPEAKER_02: Otherwise you can't it it won't work.
• 23:40
  SPEAKER_02: And I think um on the API side, um, in the tech world, like uh API is your front door, right?
• 23:47
  SPEAKER_02: You know, I I like I know in in the non-tech world it's well, maybe it's marketing and maybe it's like, you know, the different things, but um in in the in the pure tech world, uh an API is your is your kind of is your calling card, is your front door.
• 23:59
  SPEAKER_02: And I think um for us, by enabling that, by building out this this API in front of of Heath, it's giving people access to stuff.
• 24:09
  SPEAKER_02: Uh and and I want, you know, I want it to be so trivially easy for us to to help people to do good and to improve nature and scale conservation.
• 24:18
  SPEAKER_02: Um so we need to make it.
• 24:19
  SPEAKER_02: That's what we've got to build.
• 24:20
  SPEAKER_02: That's why I'm like so excited about those parts of it, you know.
• 24:23
  SPEAKER_03: You you talked before about um when we first met you about how we could use AI and tech to keep the business as lean and as small and as cohesive as we do.
• 24:32
  SPEAKER_03: Where have you seen examples of where we've been able to use tech that otherwise we might have filled roles?
• 24:39
  SPEAKER_00: Um well that's an interesting question.
• 24:41
  SPEAKER_02: Um, I I think I I've been here eight weeks and at the moment I'm on my own uh in the role.
• 24:49
  SPEAKER_02: We're hiring, by the way, everyone.
• 24:51
  SPEAKER_02: Uh um Yes, we are.
• 24:54
  SPEAKER_02: Uh so I would say I've used it a lot where I mean I was I was explaining um to Hannah on the way here that one of the things that I joined VW to do originally was to build this like capability um model for a sales and marketing space because I was working in enterprise architecture.
• 25:10
  SPEAKER_02: Um and I basically built that with Claude, you know, in a couple of hours.
• 25:17
  SPEAKER_02: So, you know, for I think I showed you, right?
• 25:19
  SPEAKER_02: It's like you did this this this OC capability model, which is in the public sector, that was a team of people for like a year to build.
• 25:26
  SPEAKER_02: Yeah, it was with VW, it's like it was like uh three or four of us for for at least a good six months.
• 25:31
  SPEAKER_02: So um, so it's crazy, and that that's a sort of an afternoon of work.
• 25:35
  SPEAKER_02: And we just we've literally just started.
• 25:37
  SPEAKER_02: These things are so early.
• 25:38
  SPEAKER_02: I mean, I my my feeling is that then kind of I think there's a big step change with the dots and and duet probably when it's when it's available and that sort of thing.
• 25:48
  SPEAKER_03: Um and I I'm worried about dots, right?
• 25:50
  SPEAKER_03: I haven't talked to you about this yet.
• 25:51
  SPEAKER_03: No, I don't think it's been as successful in the US as we've enjoyed it, because I think um Muse has had such a good landing.
• 26:00
  SPEAKER_03: Yeah, I'm really worried they're just gonna bin it.
• 26:02
  SPEAKER_03: I I I I'm not sure it's been successful enough for them.
• 26:06
  SPEAKER_02: What people are using Muse for, by the way.
• 26:07
  SPEAKER_02: I was listening to the radio yesterday and I was there they were explaining.
• 26:10
  SPEAKER_02: I'm nervous about whether this is podcast appropriate, but it totally isn't.
• 26:14
  SPEAKER_03: They're just using it for most people are using it for shopping.
• 26:16
  SPEAKER_03: Oh, yeah, yeah, yeah.
• 26:17
  SPEAKER_03: Which is crazy.
• 26:18
  SPEAKER_03: And it's crazy in a bill.
• 26:20
  SPEAKER_03: They're gonna bill on um uh a percentage of what they save you.
• 26:24
  SPEAKER_03: Yeah.
• 26:25
  SPEAKER_03: So they're gonna say if we can find you something that's a dollar cheaper, we're gonna bill you 20 cents for that.
• 26:29
  SPEAKER_03: I mean, it's a crazy business model right there, especially advertising.
• 26:32
  SPEAKER_03: Yeah, McKinsey and Bain and what all those guys do, they come and bill you for savings that that you could have made.
• 26:38
  SPEAKER_02: Yeah, yeah.
• 26:39
  SPEAKER_02: Um, so I I think I th I do think those those types of agentic stuff is, you know, or more more asynchronous agency stuff is is is like step change.
• 26:49
  SPEAKER_02: And I think, you know, we've all seen that we can give we can offload things to it.
• 26:52
  SPEAKER_02: Um it's logical to think that we'll offload to more than one.
• 26:56
  unknown: Yeah.
• 26:57
  SPEAKER_02: Um, I guess.
• 26:59
  SPEAKER_02: And then after that, the next logical thing I think I was messaging with you, or what's happening with you yesterday, right, is like, you know, how long before we're you know we have a team of agents that are being managed by an agent, right?
• 27:11
  SPEAKER_03: Um Alan, how long do you think it'll be until you have a team of agents working for you and I task you to distribute activities to them?
• 27:22
  SPEAKER_03: See what he says?
• 27:22
  SPEAKER_02: Yeah, I don't reckon it'll be that long.
• 27:25
  SPEAKER_02: Um you know, I mean it's logical, right?
• 27:28
  SPEAKER_02: Um so I think that's probably something that's gonna be pushing things much further forward.
• 27:34
  SPEAKER_02: Um I think it comes down then to the question of what we trust stuff to do autonomously, um, which is an age-old in a limited sense, Rich, already.
• 27:43
  SPEAKER_03: I can split a task between specialist AI workers, give them separate briefs and bring their findings back into one answer.
• 27:48
  SPEAKER_03: You can ask me to do that today.
• 27:50
  SPEAKER_03: That's a top tip, actually.
• 27:51
  SPEAKER_02: I don't know if you do that already, but I you it's really important to ask your your dots to do stuff in different threads.
• 27:56
  SPEAKER_02: So um, because then it is literally an it's it's sort of spins up another chat thread and it's running in a different chat thread.
• 28:02
  SPEAKER_02: I don't with Alan.
• 28:03
  SPEAKER_02: I've literally used a single thread.
• 28:04
  SPEAKER_02: So it's worth doing.
• 28:06
  SPEAKER_02: Um, especially especially if you're writing software, but e if you're even if you're not, yeah, because then you have like I I would imagine that you have then different models running for you.
• 28:15
  SPEAKER_03: As Humphrey reminded you of stuff like that's come in and it's just gone, whoa, you need to do this quickly.
• 28:20
  SPEAKER_02: Like Yeah, yeah, yeah.
• 28:21
  SPEAKER_02: I mean it's it's so cool.
• 28:23
  SPEAKER_02: I mean, it's so cool.
• 28:24
  SPEAKER_02: I just I think that's so we're gonna have to figure out what we can trust stuff with and what not, right?
• 28:28
  SPEAKER_02: And I think that's something that you know of the many things that Han and I agree with, right?
• 28:33
  SPEAKER_02: From kind of a that perspective, is like, how do we figure out what's the you that we got a Tom?
• 28:38
  SPEAKER_03: Because imagine if we'd have had like a Chris choice of like head of town just shaking a head, like, no, of course not.
• 28:45
  SPEAKER_03: Of course we wouldn't have let him do that.
• 28:47
  SPEAKER_02: Yeah.
• 28:47
  SPEAKER_02: So like, you know, I think we've got to find that right balance of what we're giving stuff or what we're giving away, what we're I mean, I was discussing with with Hannah, I think with Hannah earlier, it was was it with you?
• 28:56
  SPEAKER_02: No, I think it was with Hannah.
• 28:58
  SPEAKER_02: Like, for example, I think we should start to think about how careful we are when we're allowing agents to talk with our voice.
• 29:07
  SPEAKER_02: So I think Oh, it was with you, L, yeah.
• 29:10
  SPEAKER_02: So like I think there's some boundaries that we probably need to learn to set, right?
• 29:15
  SPEAKER_02: Um, in a similar way, um, that oh, that's interesting.
• 29:20
  SPEAKER_03: So um, if I work on something with Alan and I'll say, please can you send this to person X by um by Teams?
• 29:28
  SPEAKER_03: I always make a point of saying, please be clear it's from you.
• 29:31
  SPEAKER_03: Yeah, yeah, same.
• 29:32
  SPEAKER_03: Interesting.
• 29:33
  SPEAKER_03: Would they do that if we didn't?
• 29:35
  SPEAKER_03: Or would they present as us sending you in?
• 29:37
  SPEAKER_02: Occasionally I've forgotten to, and it just then goes out.
• 29:39
  SPEAKER_02: As if it did, yeah.
• 29:40
  SPEAKER_02: And I think that's something so so I think Oh yes, because Chris doesn't Chris doesn't say to say it's from his agent, it just comes very short and direct.
• 29:48
  SPEAKER_02: And and I think you know, letting it speak with your voice or not is something that I think we're gonna we're gonna have to test the boundaries of what's called.
• 29:55
  SPEAKER_03: Chris Williamson's got that at the moment.
• 29:56
  SPEAKER_03: I don't know if you've seen it, but they've cloned his voice on like 11 labs.
• 30:00
  SPEAKER_03: Yeah.
• 30:00
  SPEAKER_03: And people can buy his voice for like a dollar or something, and and he's giving voice overviews of products that he would never touch.
• 30:08
  SPEAKER_02: For what it's worth, it takes about 20 seconds of voice to clone.
• 30:11
  SPEAKER_02: I mean, it's like literally very, very little.
• 30:13
  SPEAKER_03: How do we go about balancing that speed of change and speed of execution and protecting the IP?
• 30:19
  SPEAKER_03: Because absolutely George and Hannah are saying, well, hold on, have we thought about this?
• 30:23
  SPEAKER_03: Can we look at this?
• 30:24
  SPEAKER_03: What are the process?
• 30:25
  SPEAKER_03: And at the same time, I'm going, but I want the new exciting thing that we can play with and learn and test and trial.
• 30:30
  SPEAKER_03: How do you manage that tension?
• 30:32
  SPEAKER_03: Because you you kind of feel like the spot that that comes together.
• 30:35
  SPEAKER_03: Yeah.
• 30:35
  SPEAKER_02: I think um, I think to be honest with you, I think first of all, I think it's a it's the it's a great tension to have.
• 30:41
  SPEAKER_02: Um, and I think um we're we're quite lucky in in the respect that I think our legal team is really capable of deciding when the risk is worth it.
• 30:50
  SPEAKER_02: I'm not sure how Hannah is beaming behind.
• 30:52
  SPEAKER_02: I'm not so happy because Hannah's there.
• 30:54
  SPEAKER_02: We I told you, well, Hannah and I were discussing it on the way on the on the way in the car.
• 30:58
  SPEAKER_02: Like I think um, you know, I think we're in a good position in terms of like understanding risk, understanding how we're gonna manage that risk.
• 31:05
  SPEAKER_02: And I think that's really important.
• 31:07
  SPEAKER_02: Um we've got to differentiate between unknown and and being careful, like not doing something because it's unknown versus um doing something and we know the risk or we we're intentional with it.
• 31:18
  SPEAKER_02: Um, and then we've got to basically understand what's the payoff, right?
• 31:22
  SPEAKER_02: So some of these models that are so powerful could be worth uh us, you know, um taking some risk uh in that sense, right?
• 31:31
  SPEAKER_03: Because they have transformed the business so quickly.
• 31:33
  SPEAKER_03: I mean, the addition of dots in the space of like a weekend.
• 31:37
  SPEAKER_03: I think I got to it just a nose ahead of you.
• 31:39
  SPEAKER_03: Yeah, yeah, yeah.
• 31:40
  SPEAKER_03: And then that was right.
• 31:41
  SPEAKER_03: So I had Saturday and Sunday playing with it.
• 31:42
  SPEAKER_03: You were on a boat fishing, yeah.
• 31:44
  SPEAKER_03: And then Monday morning we chat and you're like, it's already set up, it's flying.
• 31:47
  SPEAKER_03: Oh my god, this is amazing.
• 31:49
  SPEAKER_02: Yeah, I mean that I I mean that that has been, I think.
• 31:52
  SPEAKER_02: Um I mean, we don't get muse, do we?
• 31:55
  SPEAKER_02: Right?
• 31:55
  SPEAKER_02: So we're missing out on that.
• 31:56
  SPEAKER_03: I thought I was gonna get it in New York for Climate Week, and for some reason Apple knows that you're not US best even when you're there.
• 32:02
  SPEAKER_02: Yeah, true that I couldn't I couldn't get the and Claude, they're releasing duet without us for the moment, which is very, very mean.
• 32:09
  SPEAKER_02: And we did play, did you play with Instinct as well?
• 32:12
  SPEAKER_02: Um so my tech spidey senses are quite well tuned, and I just felt like it just feels a little bit dodgy to me.
• 32:22
  SPEAKER_03: So I I did in a personal capacity, I didn't give it access to money of any kind or things like that, and was quite impressed with the building how they did it.
• 32:30
  SPEAKER_03: And then dot arrived, and it was just night and day are different.
• 32:33
  SPEAKER_02: The the dot part is a step change, and it's a step change not because the model is any smarter.
• 32:39
  SPEAKER_02: In fact, I don't think they're using necessarily particularly smart models behind it.
• 32:43
  SPEAKER_02: It's the fact that it's more asynchronous and you can kind of delegate tasks to it in a different way.
• 32:48
  SPEAKER_03: Stack it, yeah.
• 32:48
  SPEAKER_03: Just keep giving it and it'll it'll come back with it.
• 32:51
  SPEAKER_03: I I completely agree.
• 32:52
  SPEAKER_03: That's been a game changer.
• 32:54
  SPEAKER_03: Um I I I love some of the use cases that are coming through.
• 32:58
  SPEAKER_03: So, like the way Jack and Emily and Lara have used it in the finance team is super, super exciting as well.
• 33:04
  SPEAKER_03: Um, you know, to automate and make our and they've made it so playful as well, in terms of making the budget round a playful, interactive process.
• 33:14
  SPEAKER_03: Whereas before it would have been a spreadsheet can be playful and fun too, yeah, but harder for people to collaborate on it.
• 33:20
  SPEAKER_03: So actually, some of the worry cases people have talked about in terms of AI being bad for collaboration and and bad for creating opportunities and and creativity.
• 33:28
  SPEAKER_03: I just think we've found the exact opposite.
• 33:31
  SPEAKER_02: I mean, I think um you know I spent the last years working in software.
• 33:37
  SPEAKER_02: Um that's been quite an exclusive club.
• 33:40
  SPEAKER_02: So if you wanted to have a dedicated piece of software for doing something in your business, it's gonna be a couple of mil, whatever it happens to be, right?
• 33:48
  SPEAKER_03: You know three or four times to take our platform to the next level.
• 33:53
  SPEAKER_03: We thought about hiring developers.
• 33:55
  SPEAKER_03: Yeah.
• 33:56
  SPEAKER_03: And we we just like, no, it's not core, it's not core.
• 33:59
  SPEAKER_03: If we had have done, that would be an error now with what we're able to do.
• 34:03
  SPEAKER_02: I think um, I think that the thing is that it I think software engineering is a is an incre is an incredibly foundational and important part of the tech industry.
• 34:12
  SPEAKER_02: Um, and AI is not yet capable of doing everything, um, but by God, is it capable of speeding people up?
• 34:20
  SPEAKER_02: And and the one thing that's quite interesting for me is that it's democratized access to that.
• 34:26
  SPEAKER_02: Um, so with that, it's actually enabled creativity for so many more people because software engineering is a creative, a creative uh art, right?
• 34:35
  SPEAKER_02: Um, a lot of people don't realize, but it really is.
• 34:37
  SPEAKER_02: And I think that's something that has enabled so many more people.
• 34:40
  SPEAKER_02: So I agree with you on that.
• 34:41
  SPEAKER_03: Um look in um in each pocket of the business, you've got multiple people who are wonderfully creative with AI.
• 34:47
  SPEAKER_03: So you mentioned Karis at the start, who was the first person you met.
• 34:50
  SPEAKER_03: She's the most amazing person at presenting data and information using AI.
• 34:55
  SPEAKER_03: She's spent so she's she's big on Claude and she's built a huge amount on the dashboarding side of stuff that allows her to present data in a in a beautiful way.
• 35:03
  SPEAKER_03: And then you look at how Matt uses it for a research and a strategy perspective.
• 35:08
  SPEAKER_03: He is lightning fast with what he can turn around as a result of the build there.
• 35:12
  SPEAKER_02: It is like having a jetpack or what have you in that respect.
• 35:16
  SPEAKER_02: I mean, it's just so fast.
• 35:18
  SPEAKER_02: I think, you know, we're capable of all of us doing so much more.
• 35:21
  SPEAKER_02: And um, I mean, that for sure, if if you're a techie and you're responsible for keeping everything safe and secure, it can be a little bit daunting.
• 35:28
  SPEAKER_02: Because, you know, people who've worked in software, they know what is perhaps, you know, good or bad or like good patterns in terms of security and other things that obviously that doesn't necessarily translate.
• 35:38
  SPEAKER_02: And and the one thing that currently AI is really good at is giving what you ask, not necessarily what you should should have.
• 35:45
  SPEAKER_02: You know what I mean?
• 35:46
  SPEAKER_02: Um, and I think that's probably an area that's gonna keep us on our toes a little bit, right?
• 35:50
  SPEAKER_02: Um, you know, and you and you know the first few weeks I kind of got in and I was like working on okay, this is a kind of safe way for you to build stuff and please give us your agents.
• 35:58
  SPEAKER_02: This is how it's gonna work.
• 35:59
  SPEAKER_03: Remember that day when um your um Claude told you to report yourself to Tom?
• 36:07
  SPEAKER_02: Yeah, yeah, yeah, yeah.
• 36:08
  SPEAKER_02: Yeah, I I mean I I kind of loved that.
• 36:11
  SPEAKER_02: I think I think um so just I don't celebrating again.
• 36:14
  SPEAKER_03: You're so proud.
• 36:15
  SPEAKER_02: Yeah, because you can you can so for if you have Claude for business, you can put in like these uh top level kind of prompts, right?
• 36:23
  SPEAKER_02: Um which the more you think about it, the more it's like a crazy power to wield, I think.
• 36:28
  SPEAKER_02: Because you can obviously influence everyone's work in that respect, you know, anyone who's using Claude or I tom a beer.
• 36:34
  SPEAKER_03: Yeah, exactly.
• 36:35
  SPEAKER_02: Because it can exactly um so and I and it started with because I was I'd noticed a couple of patterns.
• 36:40
  SPEAKER_02: I was like, hey, if you're working on you know anything that's touching personal information, just please check with Tom first.
• 36:46
  SPEAKER_02: He'll help you to find a way that's kind of safe.
• 36:48
  SPEAKER_02: Um and that's quite cool because it then just occasionally pops up and uh yeah, and for me as well, saying, Hey, like Tom, you sure about this?
• 36:54
  SPEAKER_02: You know, some free reported you should tell me.
• 36:56
  SPEAKER_02: Yeah, no, not yet, but that would be that would be gold.
• 36:59
  SPEAKER_03: Yeah.
• 37:01
  SPEAKER_03: On the security side of things, um, how much should people be worried about data and IT security as a result of what AI is doing to the industry?
• 37:11
  SPEAKER_02: Um, well, we had yesterday a PHP vulnerability that was um released.
• 37:16
  SPEAKER_02: We don't have any PHP in the in the business.
• 37:18
  SPEAKER_02: Can I explain to people what that is?
• 37:19
  SPEAKER_02: Uh it's a programming language, um uh a bit like Python.
• 37:23
  SPEAKER_02: Is that way before Python, I think.
• 37:25
  SPEAKER_02: You describe it as largely unpleasant.
• 37:27
  SPEAKER_02: I did say, I did tell you that it gives me the ick as a language, yeah, which will probably get me some hate.
• 37:33
  SPEAKER_02: Um, but uh yeah, so there's a big vulnerability yesterday, like a really serious vulnerability that allows people to run their code on your machine without you know being authenticated or anything.
• 37:45
  SPEAKER_02: So it's like a like the most severe levels of um vulnerabilities.
• 37:49
  SPEAKER_02: And you know, that was basically was found with with a you know researcher just asking AI like, hey, find a serious vulnerability.
• 37:56
  SPEAKER_03: Not even an astero or a fable.
• 37:58
  SPEAKER_03: It was like a 5.5 sol or something.
• 38:01
  SPEAKER_02: Yeah, so that's kind of I think that's I think that's that's gonna be a effectively a one-time effect, you think about it, because we're gonna have code bases that are gonna get tidied up by that.
• 38:12
  SPEAKER_02: Linus Torvald, who, who, who's the inventor of Linux, has started started using, you know, has started using AI to write the absolute core of Linux.
• 38:20
  SPEAKER_02: And for those that haven't come across Linux, it's like it's runs most of the internet as an operating system.
• 38:26
  SPEAKER_02: Um, so that's quite amazing.
• 38:28
  SPEAKER_02: And he even he's convinced, and he's like very tough on that, on that type of stuff.
• 38:32
  SPEAKER_02: So I think the quality of of source code will improve.
• 38:35
  SPEAKER_02: And like those, those kind of vulnerabilities are gonna probably go away over time.
• 38:41
  SPEAKER_02: Um, but we're entering a new era of agentic, long-running, autonomous uh systems, uh, and for sure people will use that for bad.
• 38:52
  SPEAKER_03: Um, so I think it could be quite spicy.
• 38:54
  SPEAKER_03: So I had to I was very fortunate to have lunch with one of the top developers at one of the leading Frontier Labs a few weeks ago.
• 39:01
  SPEAKER_03: And he describes that the sort of phase shift between what we see now and what they're working on is about four months.
• 39:07
  SPEAKER_03: Yeah.
• 39:07
  SPEAKER_03: So if you think about how impactful DOT and Astra have been, I don't know what the release was four months ago.
• 39:14
  SPEAKER_03: May maybe that was Opus 5.5, perhaps.
• 39:17
  SPEAKER_02: I was kind of trying to tell myself today to reflect on the fact that there's plenty of people that don't follow this stuff that aren't like literally day and day out checking what's going on.
• 39:27
  SPEAKER_03: I love it when I meet people that are like, oh no, no, I don't use AI.
• 39:30
  SPEAKER_03: Yeah.
• 39:31
  SPEAKER_03: Because it's really interesting talking to them about their perception of it.
• 39:34
  SPEAKER_03: So I try and stay, unless they ask me explicit, I don't I just keep listening and keep asking to understand.
• 39:40
  SPEAKER_03: We are in a very small bubble.
• 39:42
  SPEAKER_03: Yeah.
• 39:43
  SPEAKER_02: Yeah, that's true.
• 39:44
  SPEAKER_02: Um it doesn't change the fact it's gonna change the world like massively, right?
• 39:50
  SPEAKER_02: Um, you know, it feels to me more than the connectivity that the internet brought itself.
• 39:55
  SPEAKER_02: It feels like it's a bigger change.
• 39:57
  SPEAKER_03: You you quite rightly correct me earlier when I made the facetious comment about, you know, so does a typewriter.
• 40:02
  SPEAKER_02: Yeah.
• 40:02
  SPEAKER_03: Of course it's not the same.
• 40:03
  SPEAKER_03: No, it's like all these different iterations have changed the world, but nothing like this.
• 40:09
  SPEAKER_03: It's like the it'll be like the same effect as the Industrial Revolution, but in weeks and months, not generations.
• 40:15
  SPEAKER_02: Yeah, I agree with that.
• 40:16
  SPEAKER_02: I think um it's also the first time that something has um brought power to intelligence rather than you know, physical power, right?
• 40:24
  SPEAKER_02: So, you know, every every other automation so far has been around automating the physical, um, or you know, electric drill versus a hand drill or whatever, or you know, um, and this is the first time we're bringing that that type of power and change to the intellectual side, I think it's gonna be.
• 40:39
  SPEAKER_02: Do you think we've got AGI already?
• 40:43
  SPEAKER_02: Um yeah, probably.
• 40:45
  SPEAKER_02: Like if you'd asked us 10 years ago, we'd have all absolutely 100% said we had.
• 40:48
  SPEAKER_02: I think it's just a quick question of shifting the the goalposts the whole time.
• 40:52
  SPEAKER_03: It's it's quite funny.
• 40:52
  SPEAKER_03: So I'm reading an awesome book at the moment whereby each of the the sort of leaders of the respective labs and it and it um chronicles how anthropic broke away from open AI and things and about the sort of difference in philosophies and terminology.
• 41:06
  SPEAKER_03: We see that in the natural capital sector all the time.
• 41:08
  SPEAKER_03: People's arguments about what's the right type of carbon credit, what's the right reporting cycle, what's the right measurement.
• 41:13
  SPEAKER_03: And ultimately they're all arguing about nuance.
• 41:16
  SPEAKER_03: Yeah, and I think that's the same with AGI as well.
• 41:18
  SPEAKER_03: I think have we got to an artificial intelligence that's outperforming the level of intelligence of a lot of human beings right now?
• 41:27
  SPEAKER_03: Yep.
• 41:27
  SPEAKER_03: Yeah, I would say.
• 41:28
  SPEAKER_03: Yeah, we have.
• 41:30
  SPEAKER_03: In brackets for some stuff, but so absolutely right, but that's true of some people as well.
• 41:36
  SPEAKER_03: Yeah, that's true.
• 41:37
  SPEAKER_03: So, how many different types of intelligence do we have in the room?
• 41:41
  SPEAKER_03: You know, we have lawyers and mathematicians and photographers and videographers and tech guys and whatever on earth I am.
• 41:47
  SPEAKER_03: And, you know, we we have each of those different types of intelligence, whereas AI, in a general sense, is probably better than all of us across the board.
• 41:57
  unknown: Yeah.
• 41:58
  SPEAKER_02: I mean, you know, I was I was telling, telling some people earlier that, you know, at the moment I'm like finding this incredible productivity because I I I live in the far southwest of UK uh of Cornwall, right?
• 42:08
  SPEAKER_02: So I'm so used to saying UK because I live abroad, but of Cornwall.
• 42:12
  SPEAKER_02: And um, so it's two and a half hours to drive here, right?
• 42:15
  SPEAKER_02: And it's two and a half hours of talking with with Chat GPT, getting it to work on stuff for me, you know, and and it's it's incredible.
• 42:22
  SPEAKER_02: And it if if you were to turn around and say, well, that's not an artificial general intelligence, you say, what are you talking about?
• 42:26
  SPEAKER_02: That's absolutely crazy.
• 42:27
  SPEAKER_02: Of course it is.
• 42:28
  SPEAKER_03: What's the one question I haven't asked you you wish I had?
• 42:31
  SPEAKER_02: Um man, I knew this was coming, and I really couldn't think of anything good.
• 42:36
  SPEAKER_02: I always forget as well.
• 42:37
  SPEAKER_02: Like, um And I've done it twice.
• 42:40
  SPEAKER_02: Um Well, I you know, you didn't ask me about are we hiring, right?
• 42:46
  SPEAKER_02: And that's something actually very dear to my heart at the moment.
• 42:50
  SPEAKER_02: I think we got um you heard me explain about the team that we've got and how incredible it is to work uh here.
• 42:56
  SPEAKER_02: We've got two two roles uh actually advertised.
• 42:59
  SPEAKER_02: One is a uh tech ranger role.
• 43:01
  SPEAKER_02: Um look, we've got to do everything tech wise um in in OC, which is incredible from working on you know drone data acquisition through to um AI models through to just literally everything.
• 43:18
  SPEAKER_03: Remote sensing to the actual like we're we're pushing what's possible with hardware across everyone and all that camera kit and photography and videography.
• 43:27
  SPEAKER_02: As well as, right, you know, keeping the lights running, figuring out what the hell to do when there's a security.
• 43:33
  SPEAKER_02: Right, you know, all that stuff.
• 43:34
  SPEAKER_02: And I think um it's such an incredible um span, which is super interesting, number one, and then to do it uh with a team that are like massively motivated, incredible, and for a cause that's an a mission that's really worthwhile.
• 43:48
  SPEAKER_02: I think that's you know, three reasons it'll be worth coming and working from.
• 43:50
  SPEAKER_03: And with a boss who will be the most incredible teacher.
• 43:53
  SPEAKER_03: So I think um it's hard for you to say that, but I can say it.
• 43:57
  SPEAKER_03: I I think this is an opportunity for someone to learn everything about the present and the future of technology, and wherever their career wants to go, be it across remote sensing, be it across media, be it across finance, technology, that whole span of things, it is a rocket shit for someone to start here.
• 44:17
  SPEAKER_03: Whilst you haven't been with the business so long, I don't know anyone who's had more of an impact in such a short period of time.
• 44:24
  SPEAKER_03: So, on behalf of all of us, we are so incredibly delighted you're part of the team.
• 44:28
  SPEAKER_03: Thank you.
• 44:29
  SPEAKER_03: It's exciting to be here, it's like amazing.