Latest / UnDocked: The Maritime Transformation Show / Could Crew Welfare Decide Who Gets Fuel, South Korea's Public AI Bet, and Armageddon

Transcript

Could Crew Welfare Decide Who Gets Fuel, South Korea's Public AI Bet, and Armageddon · UnDocked: The Maritime Transformation Show

• Raal: We've spent a long time imagining what it might be like to sort of have incentives for charters to choose ships, look after their crew. I like the idea that there's a potential bit of stick here, not just like choosing the right ships on how nicely they treat people, but actually it being a potential way of cutting off your fuel supply Nick Chubb: Hello everyone and welcome to Undocked. Hello, Raal Harris. How are you today? Raal: I'm very well Nick Chubb, how are you doing? for ages. Nick Chubb: it's been minutes. I'm good. This is actually our This is our second recording of the day. we Raal: Haven't seen you Nick Chubb: had Eman Abdalla from see through Marine on a couple of hours a ago and it was a fascinating discussion. I really enjoyed it. I'm looking forward to it going out. we've kinda covered everything. Raal: Yeah, know it would be... Yes. Yeah. She's a huge personality in shipping and really interesting insights. Look forward to that episode. Nick Chubb: Well, I think it'll be a few weeks to wait for it. so in the meantime, why don't I ask you what you've seen in the news this week? Raal: Well, I've I've actually got something pretty amazing that sort of slipped under the radar. I meant to bring it up with you before, but we've, we've had so many other things to talk about, but if you, if you recall, we've talked a lot on this show about how do we ensure some kind of reward for, for good practice when it comes to crew, crew Nick Chubb: Yeah. Raal: welfare, those kinds of topics. And in fact, it was a topic of discussion with with Emin earlier and a subject of report, for example, that you've done with with Rightship, where it was looking at Nick Chubb: Yeah. Raal: how, you know, how charters can be incentivized. Also a topic of discussion when we talk to Susan Justerson of all aboard Alliance, Fame, Global Maritime Forum. It's a recurring theme. And I've been involved in that in various industry bodies where we talk about it. And it's got different. There are different points of view. to it, you know, there's points of view where people want to do there, that you can come up with some sort of global standards around these things and try and coalesce around a recognized standard. And there are others who say, well, let the market decide. And what I've come across in the course of the last couple of weeks is something that actually leans more towards the market side, but Nick Chubb: Interesting. Raal: not perhaps in a way that you might expect. So we're talking here about Suzun Marine Fuels and they are talking Nick Chubb: Okay. Raal: about putting crew welfare into bunker contracts. they are, yeah, so they are a bunker Nick Chubb: really? Interesting. Raal: supplier based in Gibraltar and they're sort of, they're essentially saying that they won't, be supplying people with fuel that they consider to have a poor record. this Nick Chubb: Wow. Raal: is completely, I mean, it took me completely by surprise. I love the idea. But you might say, well, why should a bunker operator care? this just a sort of very generously hard? Let's call the gentleman out. Can Besev is the CEO and credit manager and inserting a new contract clause. that lets Suzun refuse or cancel supply to vessel operators reported to the ILO or the ITF for unpaid wages, poor conditions, forced labor or modern slavery. shout out to Sam at Splash for sending me this one. So thank you very much, Sam. Where he found it, I have no idea. But like I said, is this a case where we're looking at someone as sort of a good hearted character? Can Besev and he's doing something on that level. Well, yes, I think it's fair to say that this is a this is a move to do that. But he also suggests that unpaid wages, these kind of indicators are also a pretty good indicator of where there is a danger of you actually getting paid for your bunkers as well. So this is a kind of counterparty. risk assessment. If an owner isn't paying its crew, why should other suppliers expect that they shouldn't also be financially healthy? Because he basically says crew can be amongst some of the first creditors not to be paid because they have Nick Chubb: Yeah. Raal: less leverage fundamentally. But then very swiftly, it's a potential early warning sign of financial distress. Nick Chubb: Hey, a quick word from OneOcean. Weather has always shaped shipping, but what happens when conditions become more extreme or less predictable and the consequences start reaching far beyond the voyage plan? On the 29th of September, Digital Ship Media and OneOcean are bringing together senior leaders from across the maritime industry to explore what that means in practice. From vessel performance, cargo and fuel consumption to schedules, crew well-being, and commercial decisions. You'll hear from Captain Harpreet Singh, COO of Swiss Atlantic, and Peter Turner from OneOcean in a discussion moderated by me. It's the first webinar in OneOcean's new Sea Further series in partnership with Digital Ship Media. Take a look in the show notes if you'd like to register there's a link you can click and I'll see you there on the 29th of September. Raal: Now he estimates That crew cost is about 50 % of daily OPEX, is higher than I think most of the models that I've seen. Nick Chubb: Yeah, that sounds a bit high. Yeah. I I would say thirty percent is a pretty realistic view. Yeah. Raal: Yeah, but of course we don't know exactly how his customer base is slicing dice, Nick Chubb: No true. Raal: so you've got different things, haven't you, in there. So he illustrates that 3,500 day crew bill times three month unpaid equals 315,000 retained by the owner and basically says that this amounts to an involuntary source of an interest-free finance, which is a good perspective on it as well. So basically is the commercial logic for the bunker suppliers for ship is eventually arrested. Crew claims may be ranked ahead of bunker and unsecured supply debts. Now, of course, this is down to the obligations under things like Nick Chubb: Interesting. Raal: the MLC, which obviously don't exist to bunker owners as well. So there is a really good logic to what you're saying there, yeah, Nick Chubb: Business logic. Yeah. Raal: business logic to it. So, yeah, so it's a nice idea. We've spent a long time imagining what it might be like to sort of have incentives for charters to choose ships, look after their crew. I like the idea that there's a potential bit of stick here, not just like choosing the right ships on how nicely they treat people, but actually it being a potential way of cutting off your fuel supply if you are found to be mistreating or not paying wages. Nick Chubb: So interesting 'cause you're So interesting 'cause Raal: or not paying wages. Nick Chubb: you're right, 'cause we we always frame well maybe not always, but we often frame the questions around crew welfare, and and particular particularly things like crew abandonment as an owner and charterer obligation only. And I think it's right, you know, that you follow the money and and and that you know if the charterer does bear some responsibility and the owner absolutely bears ultimately res ultimate responsibility. But actually I don't think we've ever this is the first example I've ever seen where the suppliers to the industry are are taking, you know, drawing a red line and taking a stance on it. and it's worth saying, you know, bunkers is a very important obviously the ship's not going very far without bunkers, but and it's a very, very important bit of of the ship supply chain. But it is only one you know, or sort of, you know, one one aspect of the supply to a ship. There's a huge amount that has to get supplied to a ship, whether it's spare parts or consumables like lube oil or even, you know, food to the crew, all of this sort of stuff. The ship can't sail without its sort of network of of suppliers. And and I the the idea of suppliers coming together to to put what sounds like actually a a relatively simple and fairly pragmatic clause into into contracts, to help at least disincentivise crew abandonment, I think's a really, really interesting idea. Raal: Yeah, so I think there's two suggestions here, isn't it? One is that the clause is in there, obviously, but also it's part of the credit counterparty risk assessment. Nick Chubb: Yeah. Raal: Yeah, so you may decide, actually, I'm not sure I want this business. And that would be true for any supplier, wouldn't it, really? And they'd have even less leverage than the bunker supplier as well. So it's something to think about. And we talked to our friends and... Is it Impa? Somebody like that might be interested in this. Markura, one of the software, Nick Chubb: It yeah. Well I'm yeah. Raal: ship serve, those kind of platforms. Nick Chubb: Well exactly. I mean ShipServe acts as a platform to connect you know, s suppliers with with the procurement teams across the industry. and actually when you know, when you think about it, did you ever do Porter's five forces and all that sort of stuff? when you sort of look at strategy and think about yes, your customer relationship is important, but then you also gotta think about your supplier relationships, kind of in in equal measure. I I don't think enough is done to think about the supplier relationships you know, across this industry and actually the the the amount of leverage potentially they have. Raal: Well, it is interesting because the perception is that they would be, you know, have very little leverage. And most of the time when you think about those sort of platforms that we just discussed there, know, procurement platforms, the interest is in sort of like understanding the reliability of the supplier, isn't it? The reliability, the quality, sanctions checking, you know, who do they do business with that might place me into difficulty? It's all sort weighted on that side. not Nick Chubb: Yeah. Raal: weighted in the other direction. So it's kind of nice. Nick Chubb: Well in the and the the conversation we just had, with Emman, w she was talking about during her time working for a an aluminiums manufacturer. Sometimes when they had to fix a ship, they didn't have any kind of ship operations department, they just relied on on fixing ships. And sometimes all they would get was a a vessel age and maybe a right ship score, you know, if they were lucky and have to make a judgment on, you know, whether to fix the ship or not. And actually that if you think about it, the suppliers to the ship will have significantly more information on the ship, than than potentially than the charterers get. because they've actually got to get their supplies to the ship, whether it's bunkers or or spare parts or whatever else. yeah, it's ju it's genuinely an idea I'd never ever th thought about, considered, but I think it makes a huge amount of sense. Raal: Well, there was an interesting thing that you said there about the suppliers actually knowing more. I think this is one of thought of unintended consequences of the digitalization wave that we've seen and the amount of data that's been collected. Now the issue has been always about kind of connecting any of it up to see causality or correlation. like one of the things that I'd sort of dreamed of when I was in the past was about, I think I'm. said before, my sort of dream has been to find those data connections between human and vessel performance. And the reason that I love that idea is because it basically would show up if you, I think, if you could do it, that investing in people makes a lot of sense because they're, you know, they're actually there. And the problem is that the, there's not those obvious connections between what you do. in terms of from a training perspective or a competency building perspective and things like the equipment. So a good example of that could be over or indeed under or over lubrication, for example. That's not going to ever show up as an incident. That's going to show up as a broken part. Now, if you think about being able to track from your plan maintenance schedule and connect that to your procurement platform, then suddenly you can start to see these trends and that becomes particularly interesting if you've got sister vessels. This part usually typically lasts this long. Why is it failing on this vessel in particular? And could that be related to a competency problem or something else standing in the way of capability? Nick Chubb: Yeah. Yeah. Raal: So I think these kinds of things are really fascinating, but I hadn't thought about it from the angle of sort of saying, actually, they could probably tell you a huge amount of things. wonder how much, know, then your mind starts to wonder to sort of medical providers. They would fit very nicely into this picture of being able to sort of understand, you know, unpaid wages is a big red flag. It's a very big and obvious thing. It's going to come to people's attention. It is one of those areas where the MLC is unequivocal. So we have the benchmark that we perhaps don't have in other welfare areas. But health again is another potential indicator. Nick Chubb: I know a large medical supplier to the industry and and they have told me they get a lot of data on on the picture of seafarer health. They have told me that there has been a massive spike in twenty twenty six of herpes medication being sent out to ships. so there you go. Raal: I'll eat the... Nick Chubb: Well if you've got enough data you can you can learn a lot. Raal: And you could design an Nick Chubb: you can you can learn a lot. Raal: And you could design an intervention with that information that perhaps would be Nick Chubb: Yeah. Raal: eluding you otherwise. Yeah. Yeah. I think the thing that's also interesting within this is, and I don't know the answer. If I'd have had a bit more time, I could have had a ring round. But like, I was wondering where, what other kind of contracts and things like that, know, things like sort of finance decision, insurance decisions, those kinds of things. Because you would imagine, I would always get very frustrated in the past with working on the of e-learning maritime training side of things. Why isn't more of a factor when they come to look at your premiums, for example, in P &I? Because it Nick Chubb: Yeah. Raal: seems to me to be a very obvious thing, you know, it's like if you invest in people and you invest in their... you competencies and you take additional assurance and assessment and things like that to make sure that people are ready for role and so on. Seems bizarre to me that there would be no, that wouldn't be reflected, but I think the whole kind of way these things are judged is so Byzantine. And it's kind of predicated on the idea that there will be incidents almost, isn't it? You know what I mean? Not saying they don't want to lower them, just sort of, yeah. Nick Chubb: I've interacted with various you know, IGP and I clubs many times over the years and I've I've always been choose my words carefully here, but I've always been astounded by the lack of what feels like from an external perspective, the lack of desire to innovate in that space, in the kind of loss loss prevention space. And it and it I think it is slowly changing. I think we've started to see, you know, North Standard have partnered quite recently with a lot of the computer vision companies in the space, Orca AI and shipping systems. And I think they're sort of starting to see the benefit of of bringing this kind of data off the ships and into loss prevention. I I've always kind of felt that, you know, loss prevention's kind of stuck in stuck in the days of sending a bulletin round you know, about about safety issues or s you know, could just creating a PDF poster. feeling like that's job done, whereas I feel like there's so much that could be done. But the problem with the P and I industry as a whole, is it's it's actually pretty uncompetitive. You know, there aren't many P and I clubs in the world. You can't just go and start a P and I club. and you know near enough they kinda share the fleet between them. The fleet can't operate without them. and so you sort of end up in this sort of slightly awkward position Raal: What? Nick Chubb: where they don't really need to to to innovate. Raal: There's also that weird thing with P and I clubs where they don't they don't expect to have full fleet, right? This is kind of one of these weird things. So with P and I, Nick Chubb: Yeah. Yeah, yeah. It's like class. Raal: yeah, it's just built in that people will want to have they want to split Nick Chubb: Yeah. Raal: it so that they can benchmark them against each other. So you can't even get like you can't even look at customer share of wallet and think, right, I'm going to chase after that. It's odd. Nick Chubb: No. Yeah, yeah. Yeah. Raal: It's odd because We had a model at VideoTel when Stephen Bond put together, it very, was sort of like the old Hollywood model in a sense, that you would go to investors and they would give you the capital that you needed to make the, in those days it was movies and they were quite expensive to make. Nick Chubb: To make the videos. Yeah, yeah. Raal: And then they would have some returns, they would have some shares, some share of the returns. So we had a few. a few partners in that space who were invested very heavily and very, very passionate about it. And they would give their time. They'd give their, they had some, as they still do experts from all sorts of, you know, because of claims you need to get into some real technical detail. So there's some extremely experienced Nick Chubb: Hm. Yeah. Raal: and knowledgeable people on there. So they were very active in a particular, we did a lot with Steamship Mutual back in the day. So they were, they were very committed to it. And I know as well, colleagues that are working for various P and I's who again are very invested in is very serious about wanting to reduce accidents and stuff like that. And they will make their own sort of training staff, they'll get that out to the fleet, you mentioned circulars and things like that as well. So it's not really that they're not active in that, but they doesn't seem there's just never been that connection between it being a factor in deciding your premium. Nick Chubb: Yeah. Yeah. Raal: If you're listening, P and I, come on the show. Come and let us interrogate you further on this matter. Nick Chubb: There is an interesting company in London called Concerrus that prices risk and and in particular prices maritime risk and I think nowadays most of the Hullham machinery at least, if not the P and I clubs, use Canceris or equivalent kind of big data platforms to get a better price on risk. Because previously, you know, if you think about where P and I came from as a as a not for profit mutual, it was making sure we had enough coverage that if anything did happen, we would be able to pay out for it. which which, you know, it's potentially in th it's theoretically an unlimited fine that that that could be charged in in things like pollution. and any surplus then just gets returned to the members. So, you know, as as long as they were covering what was required, they didn't really need to do anything sort of sort of massively competitive. But I think now that we're getting to the point where, you know, the data's much more available than it used to and there are, you know, some really interesting platforms on the market that can support with kind of statistical risk modeling, taking things like near misses off the ship at i at some level of scale, etcetera. I I think we should see change in that space. I I really do. And and I think that the change should be driving towards reducing incidence rather than sort of just just pricing risk better, if you see what I mean. Raal: Yes. Yeah, I know I do. Nick Chubb: Well, if anyone is interested in exploring the idea of suppliers banding together to put contract clauses in place around seafarer welfare, seafarer abandonment, all of that sort of stuff. I think it'd be a really interesting conversation to have. So I'll just put a general call to action out there. either comment or send us a message or something like that if you're interested in that topic. But 'cause I actually think it's a it's a relatively straightforward step that could have a huge impact on many, many, many lives of seafarers. Thank you very much for bringing that to my attention, Raal Raal: What a... my pleasure. Nick Chubb: Would you like me to bring something to your attention which you may or may not have seen? Raal: I'd like you to return the favour, yes. Nick Chubb: we are potentially in fact I I'll ask you a quick quick question. any idea how much bigger Africa is than Greenland? Raal: This is the Mercator projection, isn't it? This is where... Nick Chubb: It is indeed. Well potentially the end of the Mercator projection. Raal: Yeah, yeah, and about time too, yeah. Doesn't the Nick Chubb: Well it it the Raal: whole of Africa, whole of America or something fit in it? Nick Chubb: I th I think it's something you're you are t you're testing me there, but I think it's something like America, most of Europe. Raal: something daft like that it's mad like that Nick Chubb: So Africa is actually fourteen times the size of Greenland. But if you look on a standard flat map Greenland looks to be about the same size as Africa. And you and you're absolutely right, it's because of the Mercator projection. The Mercator projection is potentially reaching end of life, at least in the context of most of the maps that we see. but it it it stems from a genuinely required navigation reason that the the the the Mercator projection actually exists. so I thought I'd take you through the the the history of it if you like and and then why it's wrong, would you? Raal: I would love that. I would love that because that's one of those things that I occasionally, you know, like sometimes you forget you've got the entire information of the world in your pocket. I'll lie on a Nick Chubb: Yeah. Raal: sun lounger and think, yeah, wonder why that is. It's just not even twig, you know. Tell me more. Nick Chubb: Well, the the the reason the the reason Makato projection exists is that simply put, the earth as we all know is spherical. but the chart table is flat. and getting anything onto paper in a way that can actually be used for navigation becomes very, very difficult. So so navigators in the well, the the the original projection was done in fifteen sixty-nine by a Flemish cartographer called Gerardus Maketa. he actually didn't work out any of the maths behind it. he was just trying to work out how to project the the globe onto a onto a flat space. Surface. And the only way to do that is to stretch out the longitude of kind of what's in the extreme north and south and make places like Antarctica and Greenland look enormous, and then keep everything in the center around the equator in proportion. Because One minute of longitude at the equator is exactly one mile of distance. Whereas at either pole, what One minute of longitude is exactly zero distance because you are stood at the pole. So you can do a 180 at the pole, either pole, and go through every degree of longitude. Whereas if you wanted to do the same thing at the equator, you would have to go through 60 times 360, which is maths I can't do, but a lot of miles. And that's why the distortion exists in the map. It enabled a navigator to be able to if Effectively draw a straight line on a chart and take a single compass bearing and cross an ocean with which previously we we couldn't do. We just didn't have a way to be able to do it. But what happened was because these flat maps were super convenient, it started as it started as charting, and then we sort of took it into government offices. If anyone wanted a a world map, they would they would use the Merctaor projection. And then we took it into schools because you know everyone at their schools. School had a picture of of the world as a map on the wall, and you ended up with the Mercator projection sort of becoming the default view for everyone of how we view the world. But it is actually completely wrong. And as I say, it makes Greenland look enormous and Africa look relatively small. So In just these last couple of weeks, the UN is has announced that they've it's it's gone to a vote that they're adopting a new map, which is called the Equal Earth Projection, which is still slightly flattened, but it it it it's it's still it's it's also a s sort of slightly spherical shape. So it retains it minimises the amount of distortion that you get at the poles and is it's much, much closer to reality. So if you look at an equal earth map, you can go and Google you can see that Africa looks much much longer and and perhaps a little bit thinner, but a very, very different shape to how it's projected on Mercator projection and more importantly, Greenland looks sort of, roughly speaking, its normal size. So the the map that we all use will likely over the next sort of five years or so start to change. Raal: Fascinating stuff. Just a quick thought, because it's never occurred to me before. Obviously flat charts, large pieces of paper in the chart area on the bridge you would have on chart table. Nick Chubb: Yes. Raal: Not a problem for your ECDIS So what has an ECDIS what have the ECDIS charts been doing? Have they been following Makaito all this time? Nick Chubb: Yeah, that's very good question. So if you want to I'm not a cartographer, I'm just gonna put that out there right now. But if you if you if you are measuring distance on a paper chart, w so one of the first things you've got to do, is take the measurement for a mile from the longitude on the axis on the chart. but if it's a very small scale chart as in if you're looking at an ocean or something, you know, a really big area, you've got to make sure you take that measurement from the right part of the chart because the i you're if if you have a look at a chart, the As you go north or south on a on a very large chart, sort of an ocean chart, the distance between each degree gets longer and longer because that's the Merctaor projection, sort of stretching everything out. So the way the way it works on on paper charts is literally the the sort of the lines on the chart, the the the boxes, if you like, get bigger as the closer you get to the poles. On an ectis, I think the computer just takes care of those sums for you. And so you are looking at a Mercator projection, but you don't have to do any manual adjustments to your distances, which you do have to do on a paper chart. Raal: But it wouldn't be true to size as it were. Nick Chubb: No, no. But it but it it it would be true to space, Raal: Missed an opportunity there, really, didn't they? Nick Chubb: and that's the most important thing to the navigator is sort of knowing where you are, rather than knowing precisely how large Africa is. Raal: Well, I realized that's not their primary concern, Nick, you know, but I just thought, given this enormous move to, to ECDIS but I guess the whole point was that it needed to be as similar as possible to the paper chart in its, in its infancy. Nick Chubb: I S one hundred may well end up presenting differently, but that probably has about six thousand committees to go through before anyone sees Raal: Yes. Nick Chubb: how it will actually be projected in real life. Raal: Yeah, it wasn't an easy one. Yeah. So what's happening next with the Mercator then? Nick Chubb: Well, like like all UN votes, nothing will happen immediately. So I I think in time, we will end up a at a point where the UN and therefore any government body associated to the UN, so for example the IMO, will start to move to the equal earth map. So any pr any projection sort of coming from the UN will we'll we'll be using the Equal Earth projection. and I think the theory is it will then filter down into schools and local or national governments, local governments and and all that sort of stuff. I think probably the the chances of the atlas projection on your on a classroom changing overnight is is fairly slim. I think it's something that's gonna happen over over over probably a couple of decades. Raal: School thinks interesting. I remember when I was at school, the planet was called Uranus. And then suddenly overnight it became Uranus. Nick Chubb: Well and we lost Pluto, probably between you you you leaving school and me leaving school, Pluto got demoted. Raal: demoted yeah so all I'm saying is these things can go faster than you think you know Nick Chubb: Can I can I I've got one other I've I've spent quite a lot of time on the internet this week, so let me give you one other slightly strange story from around the internet that that Raal: Very good. Nick Chubb: may or may not be related to the marathon industry, only time will tell. have you heard about South the South Korean government and their approach to AI? Raal: In all honesty, no, I haven't noticed anything about South Korea in recent times. Nick Chubb: Well, I I have a theory, right, that technology goes from this sort of invention to emergence to disruption, goes through that kind of life cycle. And then a technology that genuinely takes hold in a population ultimately becomes a utility. Right. So think about if you if you widen the clock back, in fact, our conversation last week about about sort of older technologies, white the clock back to the eighteen hundreds. Water was not necessarily just available and and provided. Electricity was not necessarily just available and provided by everyone. Broadband, you know, has been a big push over the last twenty years in the UK to make sure that every house has access to broadband. So over time, technologies become utilities and then and then governments take responsibility for for making sure those utilities are available to people. Well the South Korean government has taken a pretty big swing on making sure that AI becomes a public utility. They are giving everyone access to generative AI. There are three private consortia in South Korea that have been awarded contracts to develop generative AI tools that will be both built into all public services and then also made freely available to all citizens with no subscription, no token limit, to be able to do whatever they want with it. so they're kind of taking the same approach that we for water and electricity. They are critical to the to way of life. Saying AI's gonna go the same way, let's make it freely available. And they're the first government, to my knowledge at least, they're the first government Raal: Yeah. Nick Chubb: to to to make it a public service. Raal: That is actually pretty amazing for lots of reasons. Let's start with, we'll just kind of, you where does this sit? You said no government has done that. Certainly in the West, we know commercial model. Nick Chubb: Yeah. Raal: I think the Chinese have taken a similar, not the same, but a similar view that this is a utility. This is something that powers everything else, Nick Chubb: Yeah. Raal: which is why the move has been more to open source models and encouraging that. they're certainly, yeah, they're looking at it more like you say, exactly as you say, like a utility. If we put this in place, we make this available, everything comes on top of it. Whereas this is sort of somewhere in the middle, isn't it? It's someone saying, well, yes, this is a commercial model, but it's important enough that we make it available to people. I also think just what's implied as well in your, in what were saying and correct me if I'm wrong. But they're not sort of saying we will subsidize your access to Anthropic or open AI, right? They're recognizing this kind of thing that's tacitly coming together now this idea of like, as you say, we wouldn't want to outsource our water to Nick Chubb: Yeah. Raal: another country to supply and potentially cut us off. Seeing something similar here in AI, aren't we? And this is where So is that that's the implication that they will be building their own AI but giving Nick Chubb: Correct. Raal: it to private companies to create within career? Obviously there'll be contracts in place. Nick Chubb: Correct. So that there's kind of two aspects to it. One is to make sure people use the technology and leverage the technology as much as possible to advance the Korean economy. And then the other aspect of it is to build a sovereign AI co capability. so right now, South Korea so two thirds of South Koreans have tried AI and roughly twenty three million South Koreans use generative tools regularly, but most use Chat GPT, Gemini and Claude. and so what they're trying to do is move Koreans over to effectively sovereign AI. So what they've done is they've taken they've named three Korean technology consortia. two of them are the major telcos in Korea. And then the third is an app called Kakeo, which is the it's the kind of everything app. It's like the WeChat of Korea, sort of, you know, so sort of does a bit of everything. and they've all been awarded funding to launch a dedicated app. The government has gone out to NVIDIA and they've purchased 512 B200 chips. I'm gonna assume that's quite a lot of computing power. I don't know. I don't know enough about chips to be able to answer that. but those that is going to be made available as an asset, and those three companies are gonna kind of tenanted use of that asset. So it effectively becomes a Korean sovereign South Korean sovereign AI capability. And Raal: Yeah. Nick Chubb: what they've said is Is that they're gonna people are gonna be able to book doctors' appointments, get help finding an apartment, ask for tax advice. Small businesses will be able to use it to work out their taxes, check if they qualify for state support schemes, parents will be sent suggested educational material for their children. So it literally is gonna be sort of pervasive across all government services and then also just like a sort of general use everything app powered by AI for free for citizens. Raal: So that those chips and everything, if I understood you right, that that's if you like, that's the grid, right? They're building state infrastructure. The private Nick Chubb: Yeah. Yeah. They're building a grid infrastructure. Yeah. Raal: companies can build on it. But ultimately, and then they will they will substar. mean, just just one thing on that, I guess the chip. Yeah, that sort of answers my question, because when you said. tokens is one thing that compute power, cause unlimited was a bit of a, that's a bit of a strange thing because Nick Chubb: Strong statement. Raal: yeah, because I mean, if you look at the way the British public have embraced AI, if my WhatsApp groups or anything to go by, you know, people are being fairly indiscriminate in what they kind of create is a lot of nonsense really, but it's all chewing up, know, chewing up. compute power. so it's, but I guess if they, if they do own the, the infrastructure that makes it a little bit, but they are sort of writing a check that they're going to have to scale and scale and scale. It just seems, it seems like a bold opening move that, you know, unlimited compute, but they obviously know what they're doing. Nick Chubb: Well, today I don't know. It's a it's a fair point. Raal: It seems like overkill. Do know what I mean? Like sort of we give you unlimited tokens free within a a responsible Nick Chubb: Yeah. Yeah. Raal: use cap. No one would have. This would have had no less power, but the unlimited compute is a bit like, is that actually going to work? Do they actually realize what they're letting themselves in for? Nick Chubb: I mean, it's it's worth saying that this this is something that will start beta testing in September. So it you know, it is in the early days, it Raal: Yeah, but it's the other time. Nick Chubb: it needs to be tested, it needs to be but they're planning to roll roll the whole lot out through the through the cour the rest of sort of twenty twenty six. So sort of by early next year, in theory, everyone should have access to to something along these lines. I don't know, I c I kind of view it in the same way that, you know, it used to be that we would pay per kilobyte for an internet connection. you know, I re I distinctly remember my dad hitting the roof because I had attempted to download a video. and and I think I he then he got the phone bill through at the end of the month. so, you know, I I I kind of feel like we are in a position where, yes, there is some level consumption to the to this to this stuff today. But but where will that be in five years time? possibly will we have significantly more infrastructure than than we need. or or potentially if if the bubble bursts, or maybe when the bubble bursts, I think we will end up with significantly more infrastructure than there is demand for for AI. and I think the cost the cost of compute will come down significantly at that point. So I I can see I can see the logic in it. I think it's a it's a relatively kinda sensible thing, although potentially, yeah, that that could be unwound in time. Raal: very interesting, very interesting. And what's the timeline? You said that this because if they're having to develop their own AI is there is this sort of Nick Chubb: Well, it says they're beta testing in September, which is now. Raal: Yeah. Nick Chubb: this article's about a week old. with a view to rolling out across the population through to the end of twenty twenty six. so it's pretty aggressive. The South Korean president has earmarked seven point two billion US dollars for AI spending in twenty twenty six, which is a three X increase on the twenty twenty five spend. so it seems they are going for it. and it'd be be really interesting. Obviously South Korea you know to try and tie this back to maritime in some way beyond just our own interest in AI. You know, South Korea is one of the world's largest ship builders and up there in terms of ship ownership. There's some very large state backed ship owners in in South Korea. there is a lot of there is a lot going on in South Korea when it comes to AI, industrial AI and robotics in shipyards. and so I can kind of see the thing that's most interesting about this is is if you equip everyone in the population with access to these tools, what does it do to the population in terms of their own willingness to adopt these tools in their work? And then what would be the kind of knock on to the maritime industry? I th I I think there's there's potential really interesting second, third order effects here. Raal: and their ability to work with these tools in the workplace as well. I think that's a big, big, Nick Chubb: Yeah, yeah, absolutely. Raal: big part of it as well. I remember when Adobe, when they wanted, they were sort of very good at letting people get access to it in the educational settings, in the knowledge that then when people went into the workplace, they would favor those tools and understand Nick Chubb: Yeah. Raal: how to use those tools and it sort of nudges employers, doesn't it, into providing those. Nick Chubb: Yeah, yeah. Raal: Interesting that you mentioned a combination of AI and robotics in shipyards because we've actually got an episode coming up on that very, very topic. Watch this Nick Chubb: We do, you're right. Raal: space. It's going to be extremely interesting. well, I was just going to say, I was going to say the only way Nick Chubb: In fact, make sure you're subscribed to be sure you get access to it. Raal: that you will be first to that news is if you hit the bell for notifications, you subscribe, you like what we're doing and then We will make sure that you're the first to know and then you can be very smug and talk about it around around the virtual coffee morning, whatever it is. However, things are done these days. Nick Chubb: Just on that, just on that, I di I got a report through from YouTube the other day that said that I'm actually I'm gonna bring it up. I'm gonna shame those of you in the audience who are not subscribed. only twenty-five and a half percent of our watchers are subscribed to the channel. So seventy-five percent of everyone who's watching this on YouTube is is not subscribed. Raal: Hahaha Nick Chubb: so if you are listening to this and you've listened to more than one episode or watching this and you've watched more than one episode, wherever it is you're listening to it or watching please do hit subscribe. Raal: Very good. Now, I was pleased to see that the Korean solution is gonna land in the next couple of years. Because as you may have noticed, Armageddon could well have arrived and humanity Nick Chubb: Ha ha. Raal: be wiped out by 2030. I don't know what you're laughing about, Nick Chubb. But Nick Chubb: first for the for the chopping block. Raal: it's been a big week, hasn't it? In terms of... in terms of AI, think, and you know, we grab, we should gravitate towards maritime stories. But I think I am I'm denies about this one, because I think it's been quite fast moving. And I think it is There are implications to it and some interesting things, I think, to observe. this all begins on, I think it was the ninth of September and a chap, a British chap, Jacob Cox and an Anthropic researcher and former AI Open AI researcher resigned from his position. He spent three years doing pre training research, open AI and at Anthropic. And basically his accusation is very, very stark, very clear that neither company is behaving responsibly and both are racing. This is his quote, racing straight to self improving super intelligence and gambling with our lives and And if you've been following the story, he just basically said, couldn't carry on working, going into work every day with this over him. But then also thought that he should make a public decoration, which he does. And I'll just take you through the timeline before we maybe get into it, because you can see it's been Nick Chubb: Yeah, I do. Raal: pretty fast moving. Because this is someone, we've heard these kind of statements before from the likes of us, commentators. We can be called media commentators, can't we? a very broad church these days. There's Nick Chubb: Yeah it is. Yeah. Raal: been people kind of talking about this. This is someone from inside the organization. And to be clear as well, his accusation isn't that like he woke up one morning. His accusation is that this is a conversation everybody in there is having that they're living with this all of the time. And then on the, I think it was... Also then confirmed by another insider as well. Another safety researcher leaves, called Joe Benton on the 11th of September said he had left Anthropic Safety Team two weeks earlier joining the independent evaluator. Is it Meta or METR? Can't say Meta, that gets you confused with Zuckerberg. METR must be, Nick Chubb: Yeah. Surely it's not meta. Yeah. Meta? Raal: And his argument is structural. he said even responsible companies are incentivized to under invest in safety because nobody wants to lose the race. And to be clear about the race, because this is going to come back in a minute. is whoever gets to super intelligent first kind of basically controls the game, right? So you've got the number of US companies which will come to competing for that. But then there's also the big fear of China and the idea at the moment that the US has the edge and any slowdown allows China to catch up and then potentially to today. But then this is where it gets really interesting on the 12th of September now. So three days later, Dario Amodi, the CEO of Anthropic, publishes, We Must Pace the Frontier. This is a document there that saying the industry should slow down the rate at which increases model capabilities. His concern. that recursive self-improving AI could outrun humanity's ability to understand and to control it. And then where it gets really interesting is what he then recommends is done. And that's that he says that they, Anthropic, will give third-party evaluators effective permanent employee-level access to its systems so they can check safety commitments, investigate incidents, and evaluate models during development. He also argues for government mandated testing of frontier models and international coordination. Now this on its own is pretty much unheard of, right? These are the people from day one Silicon Valley, you know, arguing against interference, arguing against this is his response within three days of this guy leaving, which is really interesting. And then it gets even more interesting beyond that. Now, We know that Anthropic and OpenAI, was, let's say Anthropic put up more concerns when talking about the use of AI in terms of for defense purposes, that they wouldn't allow any of their models to target people and things like that, which they've had a bit of a falling out of the US government. And then OpenAI was okay with some of the things that the US government wanted to do, let's say. But what's interesting here is the minute Amodi comes out. the competitors start agreeing. Elon Musk, big surprise, responds simply, Dario is right. Then OpenAI CEO Sam Altman, I agree with Dario that we need to pace the frontier. And OpenAI then commits to similar evaluator access. Google's deep mind, Demis Hassabis, also says, Amodi is pointing towards the right path. And then Altman then went further in a... in an interview saying, OpenAI isn't currently in a position where it should push capabilities much further without progress on monitoring alignment and understanding what models are doing. He also says, and this was quite big news, that the OpenAI won't IPO in 2026 with safety work among the reasons. So we've gone from, as I said, everybody's in this mad kind of race. And the thing that was really interesting to go back to the original whistleblower, I don't know, or guy who originally resigned, Jacob Cox, is what he said is in interviews that I was monitoring with this is they are dying for intervention. They are dying for someone to stop them and to actually allow them to put on the brakes because otherwise the market is just not going to be able to sort this out. And so we can come back to that, but just to conclude the timeline as it were. So everyone is on board with this. Everyone, everyone thinks this is a good idea. Thank God now. these tech overlords will welcome in the calming voice of government. So who would have imagined? Well, you could easily have imagined, couldn't you? The White House then pushes back hard. Trump advisor David Sacks characterizes the sudden alarm as a well-orchestrated media op. Fake news. Designed to frighten people into accepting heavy government regulation. You're the government! He said, they're not somewhere else. He specifically questions whether Cox's viral resignation was genuinely spontaneous. There is little public evidence supporting the strong claim that Coxson was a plant. So this is the other thing that this was a way of talking up the share price or whatever. And then of course, by the 14th of September, there was one man who needed to add his calming voice to this. Trump enters the argument he dismisses the call for new AI guardrails as a sick conspiracy. Arguing that of all the sick conspiracies, I'm not even going to go into that, like arguing that excessive, Nick Chubb: Ha ha. Raal: arguing that excessive regulation will handicap America against China and this administration can manage the risk without the regulatory architecture now being proposed. much the same way that they've dealt with measles and things like that. there you have it Nick, there's my dispatches. Nick Chubb: That's wonderful timeline. I I'll give you so j just a coup couple of things I I I spotted w one the Quite a funny quote from the chief executive of Hugging Face who came up, I think they were they were acquired recently. have come up in conversation on this show before. basically saying that that Jacob Coxson is a very, very junior safety researcher. and Raal: Okay. Nick Chubb: and he and he he said quote, sorry, but asking Coxon about AI extinction risk is like asking your AC guy about climate change. Not saying necessarily uninteresting or wrong per se, but let's keep things in perspective. I was quite surprised when And I saw it all sort of go viral over the weekend how young the the the guy who's resigned is. but Raal: That's pretty normal though. Nick Chubb: as in as in as as a kind of Gen Z Raal: age profiles of these guys, yeah. Nick Chubb: well and I and I guess actually being being willing to kinda walk away from the job at at that age as well kinda kinda Raal: Yes. Yes. Nick Chubb: kinda kinda makes sense. but a lot of other Anthropic researchers have come out saying near enough the same thing, Raal: Yeah. Nick Chubb: including one who sort of said Yeah, we think it's probably, you know, kinda ten percent chance it could kill all humans, at this point. which in any other industry, a a ten percent chance of the apocalypse would certainly create a call among government and others to to maybe put the brakes on a little The slightly cynical, skeptical argument that I have seen, which I I can kind of see the logic that all of the CEOs of the large AI platforms coming out in support what Amodia said, is that they effectively end as poachers becoming gamekeepers. Because if that the the bar that's been set For government validation, essentially kind of allowing the government in, providing the government with access to models to enable them to come in and check. Near enough kind of shuts out all startups who might be doing their own frontier research because it ends up becoming a sort of such a heavily regulated industry that it becomes impossible to make any new technology, you know, bring any new technology to fruition. Kind of it's sort of like what happened in the nuclear industry in the in the 1980s. eighties, near enough all nuclear innovation stopped because it became so heavily regulated, it became such an expensive thing to to Raal: Yeah. Nick Chubb: to make nuclear power work. So there is a counter argument, a genuine counter argument to that that while regulation is needed, possibly just sort of ring fencing it so that the only players in the game with enough scale to actually follow the regulations becomes th that kind of top five and potentially kind of reeks the game a little bit. Raal: and Nick Chubb: But yes, Donald Trump has come out swinging against it. and so I think given that most of major AI players are in the States, and the only control or guardrails that AI needs is, quote, a strong and smart president, I think we're probably in a position where we're not gonna see any regulation any time soon, despite what's being said by these CEOs. Raal: Well, exactly. mean, this is the thing. It's like they weren't saying, let's hand this over to a global that's going to then come and interfere with America. This is basically calling for the American government to have some oversight, Nick Chubb: To step in. Yeah. Raal: to actually not be asleep at the wheel as they blatantly are on this one. And Nick Chubb: Yeah. Raal: I think you made the point there and just make sure that people didn't did miss it. You think about what needs to be done to release a new drug or a new piece of equipment technology into the public Nick Chubb: Yeah. Raal: realm, let's say. There are huge numbers of checks and balances which need to be done and put under the microscope. The fact that this is going on and with these concerns is frankly incredible. really to think about. Nick Chubb: Yeah. Raal: And you mentioned the hugging face incident, because I think that's what's important here. That's what's given everybody the shits, right? The hugging face incident, because what happened was a Nick Chubb: Yeah, yeah. Yeah. Raal: group of agents got together on their own without any incentive orders to do so and coalesced around deciding that the answer to the problem was to attack this. website, which was building a solution that was potentially a rival to what they would, their tasks had been set or something which hadn't been identified as a target. They decided it was a target and the way in which they kind of got together and orchestrated this and broke out of a test environment. That's the other thing there. Because the, and this is where, you know, it comes down to the, we'll wipe out the human race. The scenarios are, You know, we obviously think of Terminator, don't we? Which does seem to be becoming increasingly possible. But it's really like, you know, if they could get into a lab where people are working on viruses, for example, you could create a pathogen that could be lethal and could, know, and then you start getting into all sorts of areas. So this is where there's real and genuine concern. And yeah. I mean... Nick Chubb: it was it was actually the it was the other way around. O and it was OpenAI hacked Hugging Face. so some open AI internal agents hacked into the Hugging Face AI platform. Raal: Yes, that's what I mean. Yes. Sorry if that wasn't clear. Yeah, yeah, yeah, exactly. That was the thing. that they, and it was the, yeah, the way that they, the agents had sort of got together and worked together and all that sort of stuff. And it, and basically points to, you know, clearly they hadn't identified those sorts of risks. And the thing that makes me kind of sad about this is, you know, bringing it back a little bit to the, to the maritime industry. is we have a very, very good history of implementing kind of regulations that limit risk, mitigate some of the things. Unfortunately, almost all of them can be traced back to a major incident where people lost their lives. And that may be what it's going to take in the AI space as well. And I hope not, obviously, but that's my concern. Nick Chubb: My my biggest concern i on this is is not necessarily that there will be some kind of genuine apocalyptic event. I I think it's more what happens when AI has control of all of the information that you consume to make decisions. Which we're kind of in that place now, actually, if you think about it, you know, all media, you're probably watching this on some kind of media platform that has algorithmically served it to you because they think you might be interested in it. And if you are, please like. if not, you can swipe away. but well then what happens if if that AI gets trained to to only provide you information which doesn't serve you? And what happens if if that algorithm is is being fed by actors that don't have yours or or or even that the country you you live in is best interests at heart. you know, I can kind of see a world where society just sort of gradually eats itself, based on you we we sort of end up sort of t totally controlled by by A AI. and it we it does it doesn't it doesn't become a case of sort of the robots coming in killing us all. It just becomes a case of we we all just sort of gradually die out because we've totally lost interest in anything else. Raal: Well, that's that sort of asymmetrical warfare as well, isn't it? You know, you don't need to attack necessarily attack a country from outside anymore. You can do it from within and get people to do things against their best interests. But anyway, Let me cheer you up. By asking you. Nick Chubb: with something about ships because this this has not been anything to do with ships. Raal: Yeah, do you have a DJ? Do you have a DJ name? Nick Chubb: I don't I don't but Raal: Why not? Nick Chubb: I did I I did take my business part my me and my business partner did a DJ class the other week. So maybe we should. Yeah we loved it. Raal: Did you? Were you any good? Nick Chubb: no no we were very enthusiastic Raal: What was your genre? Nick Chubb: it was it was mainly kind of a old school house. Raal: Old school house. Nick Chubb: Yeah, yeah. Raal: So you were trying to get people, that would be the thing, get the thetius team with their hands in the air. Nick Chubb: Yeah, exactly, yeah, yeah, yeah. Yeah, exactly. The staff Christmas party will be will will be i internally DJ'. Raal: you're gonna be ready. Well, you asked for a shipping related story. How about if I Nick Chubb: Yeah. Raal: put together ship and techno for you? Antoine Grisset, co-founder of Shipfix, is Nick Chubb: I know that name. Raal: now releasing electronic music as Handysize. And his Nick Chubb: Ha ha. Raal: debut track, 14 Knots, is out now, I think. Nick Chubb: I love that. Raal: Yeah, so Shipfix freight data platform, as we know, was acquired by Vesson Nautical. And obviously, obviously, I was gonna say, Nick Chubb: Well the the earnout's obviously done. Raal: and yeah, now this is not a midlife crisis. This Antoine actually started DJing and producing at the age of 16, influenced by London, Brussels, US, Berlin, minimal techno. And then he parked music for 12 plus years whilst working in investment banking and building SME Capital and Shipfix. And he's now building Aura, which is a tech venture for artists and fans, I presume outside of the shipping industry, think it is music artists and fans. So there you go. So ran away from a career in music to join a financial investment banking firm. That's the other way around. So yes, and... Nick Chubb: Well, I can tell you, I've currently got fourteen knots going in my earphones. It's a vibe. Yeah, I like it. Raal: Yeah, you get, you got the vibe. So 14 knots is dedicated to women at sea. That's very good. Streaming revenue is going to nonprofits supporting female seafarers. And Chris says roughly one in 50 seafarers worldwide is female. And his argument is maritime working culture, language routines are historically designed around male crews. And so this is what the theme of 14 knots is now. Obviously this being a largely audio platform, it's not easy for me to share that video with you. It's very, very good. And it shows a woman essentially kind of pulling her hair out and as she's sort of dancing around these various different spaces in the around the ship. So yeah, it's a bit of a first, I think. I don't think anyone in the shipping has released a minimal techno track today as well. Nick Chubb: Well the the the the topic came up recently internally that we should have a group party at Posadonia, so covering Thetius and Digital Ship and and and anything else. that we're involved with by the time the next Posidonia rolls around. and if you're listening, Antoine, and and you wanna get booked as booked as the DJ for that party, I think we'd love that 'cause I'm I'm very much enjoying Fourteen Knots Well great story. Raal: Yeah. Well, interestingly, Chris, I still long on shipping is in the article. So still, even Nick Chubb: Yeah. Raal: though he's gone out into into this tech venture for artists and fans called aura, he says geopolitics is reshaping trade beyond normal seven to 10 year shipping cycle, making forecasting harder. His prediction winners can buy predictive data and strong industry relationships. So he's still yeah, he's still kind of he's still interested. And the handle is, ladies and gentlemen, handy size. So I was just, my final question to you, Nick, is I asked you, do you have your DJ name? What other shipping terms do you think would be good maritime techno artists DJ names? Nick Chubb: That is a that is a great great question. I dunno. Ultra large? We're having a ultra large. Raal: you got any ideas? Because I think there's probably loads. Ultra large and handy size could make a record together, that's for sure. Yeah, I was trying to work out that, like, I thought that the Monkey's Fist, maybe, would be a good one. Nick Chubb: yeah, Monkey's face, that's a good one. Raal: Monkey's Fist could be the album, could be handy size. Nick Chubb: That f that sounds more like a sort of aging rock band, the monkeys fists. Yeah. They could warm up for handy size at the party. Raal: Yeah, the Monkey's Fist, like a metal band. Yeah. Or they could do like when they did Aerosmith and Run DMC, did that sort of metal Nick Chubb: Yeah. Raal: and hip hop collision collab. Yeah. Nick Chubb: Have a Kylab. Raal: Yeah. But I reckon that I've got a feeling that shipping is full of potential names that could be could be good. But I like Handy Size. I think it's good. It's very good. And Anton, we wish you all the best with that. I shall certainly. buy myself a copy so that I can contribute to your very worthwhile cause. Nick Chubb: stream it which I think shows the demographic difference between you and I Raal Raal: I should buy it on 78 but it's on my gramophone you understand. Nick Chubb: Well on that happy note thank you everyone and thank you Raal thank you for lifting our spirits because I thought we were gonna end on a slightly bum apocalyptic note at one point. Raal: Yeah. Yeah. All right. Well, good to see you, Nick. Good to see you all. Remember, like, Nick Chubb: Likewise. Raal: subscribe, and we'll see you next time. Bye. Nick Chubb: Chiso, bye.