Latest / UnDocked: The Maritime Transformation Show / $600 Ship Data Collection, Digital Literacy, and Preventive Healthcare
Transcript
- Nick Chubb: Hello everyone and welcome to Undocked. Hello Raoul Harris, how are you today? Raal: I'm very well Nick Chubb, how are you? Nick Chubb: I'm good. It feels like ages since we caught up and had an episode like this. Raal: It's been a while, it's been a while, a lot of big, well, we had our two Bergen specials, if you haven't caught them, one of which, with the great and good, rather than us and Natalie. Nick Chubb: This episode of Undocked is brought to you by GTT Marine. Today's shipping companies are drowning in data. Performance dashboards, compliance feeds, voyage tools, emissions reports, and yet, when the time comes to make a real commercial call, teams still fall back on gut feel. Why? Because more data hasn't meant better decisions. The Great Integration is a new report from Danalek and the team at Thetius. It looks at why fragmented systems are quietly eroding, decision quality, and what ship owners can do about it. Raal: but I took a look across the news today and yeah, I think the big news today is our friend, Nick Chubb: It. It's a practical read for anyone who centers their tools should be doing more. And if you'd like to learn more, head over to gtttmarine.fr. Raal: Michael Scov stepping down from as CEO of Hafnir. So Michael on the show, great episode if you haven't seen it, give a look to that one. But yeah, shock. Nick Chubb: Yeah. Well, he's done it all now, I guess. He's done the Undocked podcast. He's been CEO of a public listed company. That's it, as well. There's no peeking that, so he may as stand down. Raal: There's no greater mountains to No worlds left to conquer. But yeah, I mean, he's been running Hafniel's and he found a CEO from 2010, I think, something like that. Yeah. Nick Chubb: Yeah, 16 years he's been running the show. And one hell of a run, they went, by the way, a great episode, well worth going back and listening to, but they went from startup to the world's largest chemical tanker operator and became a publicly listed company. And consistently, it's seemingly every single quarter produce a pretty stunning set of results. And I think Michael's on how go about that and what to pay attention to and what not to pay attention to, which we discussed on the podcast, is well worth listening to. So yeah, he's quite an operator and he's done an incredible ⁓ Raal: Hmm. Yeah, and I think as well, a sort of challenging market, the spot market where it's all about kind of, you know, dealing with volatility. so it's almost kind of like, there's a lot of lessons in there as we not I guess we're all in a volatile market nowadays, ⁓ sadly. yeah, how to thrive in sort of uncertainty and be live and agile and make, you know, decisions as you go. and he succeeded by... Nick Chubb: You Raal: Steenberg Jensen, was the executive vice president, head of asset management there. So he's the next CEO. So we wish him all the best. Best of luck. And Michael also not exiting from the business. I understand he's too subject to board approval or whatever. Join HFNIA's board of directors, which they'll do a general meeting. So yeah, big news, big news for big personality. As you say, he's done a fantastic job, really left his mark there. So wish him the best with whatever he's going to do next. So, yeah, good luck, Michael. Nick Chubb: Yeah, good luck indeed. I take on a I saw that I thought was quite interesting? ⁓ Raal: Yeah, sure. Nick Chubb: There's a, came out a couple of days ago, this company called Smart Ship Hub based in Singapore. They're ⁓ platform that puts a server onto the vessel that allows you to capture high frequency data and then to the shore then it can be used for all the kind of performance purposes, all that sort of stuff. They've just come out with an onboard logger. that they're pricing at less than 600 US dollars. And the whole idea is to make high frequency data collection accessible across a much, much larger share of the global fleet. I've been involved in various projects. I haven't done anything in this space for a few years, but the last time I was kind of meaningfully involved in this space in the early 2020s, getting a kind of equivalent setup. to what they seem to be offering for $600 was easily a kind of best ways 10 at the most expensive end 25,000 US dollar kind of commitment in terms of the kind of equipment needed on board the vessel. And so it's just remarkable how quickly the price of this sort of stuff has dropped in literally a space of in the space of five years. it's possible to transmit, ⁓ let's just with the basics. You could transmit a every second. So that one data point every second. You then take a whole bunch power readings every second from all of the different equipment on board. can take fuel flow meter readings every second. You take temperature readings from equipment every second. you can take like telematics data every second. so ship actually ends up, if you want to try and capture it all, can produce enormous volume of data. And there's kind two challenges there. One is actually just getting the data around the ship because you've got steel bulkheads. can't, the wifi doesn't really work on a ship in the traditional sense that we think of it. It doesn't penetrate through ⁓ And so you need clever ⁓ to enable you to get data of of the sensors onto a central server on the ship. And then you either need to do something intelligent with that data on board. And certainly what it seems like is being included in this kind of $600 is what they're describing as a kind of edge intelligence tool to allow you to do some data processing and get some intelligence on board. Or, and or, you've got to then transmit that data back to shore. And that used to be the major problem. These days, it's much less of a problem because of Starlink. But even with Starlink capability data processing capability, you still potentially be taking that entire with high frequency data kind of leaving nothing else left. So a lot of the kind skill, if you like, in making this work is in how you manage the data coming off the and how you kind of package it up so that it can be transmitted ashore in an intelligent way or how it gets processed at the and how moves around the ship. So it's like it's a super complex area and it used to be, it has traditionally been super expensive to actually kind of put all of this kit on board to measure the right stuff. And so we're now, it looks like certainly, or directionally, we're moving towards this becoming a tiny, tiny part of the overall investment in improving the performance of the ship, which is quite exciting. Raal: I remember this, I'm going to say 2017, 18, 19 maybe. One of our sponsors, KVH, a guy called Robert Hopkins who was with KVH at time. KVH had acquired Videotel, the company that I was in. And he was sort of the lead on a project they had called Watch. And this was, do you remember this? Nick Chubb: Yep. Yes, that's that IoT platform, right? Yeah. Raal: Exactly, yeah. And this was exactly the idea, was that it would be that there was this, all this data was hugely valuable for planned maintenance, i.e. So you can actually do effectively condition-based monitoring because you're constantly able to see the condition of equipment and so on. You can see when things need work. again, bandwidth was way more constrained ⁓ then it is now. So the value was in this watch box, taking all that data and ⁓ sending back what you needed to know. Because that's the thing, you know, when you talk about like... Nick Chubb: Yeah. Raal: constraints of the bandwidth and the huge volumes of information, you going to clog the pipe with all that, but also as well, you don't need it. What you actually need is the bit that tells you is something running as expected or is there some material change, big temperature change, big sort of, you know. Nick Chubb: Yep. Raal: from the, let's say the norm that you're expecting or whatever is, or indeed a kind of a broader diagnosis a longer period is what you get back. So that makes a of sense in any scenario because just, know, all data has to kept. It has to be... cleaned, passed, et cetera. if it's not gonna add any real value, then don't want that anyway. So that's It's interesting the model that if I understood correctly there, because you were sort of saying there's a, unless I misunderstood, a unit cost for that rather than the service cost for that. Nick Chubb: Well, what they've said is they are making the onboard vessel data logger less than $600. I don't know the exact detail on the price as covered in Digital Ship. But yeah, essentially, mean... That near enough, in the grand scheme of things to me, it kind of takes it as close to zero as you possibly can. Because certainly a lot of the high frequency data collection systems that we see in the market will charge for the equipment and the ship owner will buy the equipment and then own the equipment on board and then charge a subscription fee for the kind of ongoing data services. And as I say, that equipment charge can be fairly high. they charge on the other end of it, once got the $600 box on board, I don't know. But that's worth checking. Raal: So interesting. It's always nice when when things you remember from the past and think what happened to that as an idea. It's good to know that that's still out there, still happening. And I imagine as well. probably even more data points coming through ever before. So yeah, interesting, keep an eye on that one. Takes to a story, similar kind of thing that we've talked about often in the past now, when you talk about sort of tech providers and that kind of stuff. And this is a discussion piece ⁓ in Splash, 24 Splash tech, I think call their sort of tech roundup. And it shipping isn't ready to bet on a single provider. So this is the idea that we've talked about like whether the increasing sort of amount of digital solutions, technology adoption, digitalization, broadly the name use, onboard chips, points towards the single platform play where you want everything in one place. So that makes ⁓ seamless retrieval of data or getting it, know, being able to connect all that operational data and decision making and so on. having a sort of an ecosystem of different partners working towards together. And we've we've you and I have talked about that. We have our own views on it. I know you have strong views on it, but it's really interesting to see because in this particular piece, put together a number of people like Bjorn Hogard also been on the show. Hafnir mentioned there as well. Sedna being on the show. Everybody, everybody who is anybody basically Nick is kind of in this common thing. Yeah. Nick Chubb: I notice we weren't asked. just looking at the list of people on it. I'm slightly insulted. Raal: Well, I read that as they watched our show when we talked about this and thought, there's a story in that. So they didn't want to come. So Sam, if I'm right, you've been run board, son. that's how I read it, because there's exactly the themes and arguments. But they sort of, you you and I talk good game. They actually analyzed it and cross checked it with a number of different opinions. indeed, you ⁓ it shakes that people are. Nick Chubb: You That's a very flattering read. Raal: critical of a single vendor approach that they consider fundamentally that different people are best in class at different things. And so want to be able to pick those. And I the other thing that's borne out in the article as well is the thing that you and I have talked about around. Well, actually the way technology is moving with AI and its ability to translate and act as a translation layer from different data sources and that sort of stuff, it's actually easier than ever before to connect data up. And there's also a point Tabitha Logan of Cetus Marine also been on the show where she makes the point that actually as well, if people need to ⁓ middleware software now, it's never been easier. ⁓ Nick Chubb: Yeah. Raal: to engineer code. She does make the caveat, of course, that that then takes you into you better be prepared to, you know, to have a software on going forward. Yeah, maintain it. And, you know, these days, software is never finished, you know, so. that's a good point to bear in mind. But I thought it was interesting to see an article that that sort of takes out sort of how can I put it out slightly? ⁓ Nick Chubb: maintain it. Yeah, yeah. Raal: hypothesized view and actually tested it out with a a bunch of interviews and a bit of research behind it. Nick Chubb: I have a kind of theory on this it's ⁓ in part by conversations that I've had with Singh, who's another friend of the pod, that we through phases of ⁓ of particular where we end up with consolidation in the market and then ⁓ we through periods divergence where those... converged offerings start to move very, very slowly. breakthrough technologies become available, like what's happening in the AI world at the moment. And get the kind of startups coming through with the point solutions that are just so much than the kind of big platform, the monolithic platform, that then the starts to diverge and move towards that kind of best in class for different point solutions joined up by APIs. And that's become super easy. And then that cycle probably up repeating. itself and moving back into some level of convergence where they all get swept up by and the like and everyone ends up close to a single platform. But the that we always give ship owners is that you should have... an exit route planned from your software and kind of really understand how your wires in to the rest of your organization, how different pieces of software wire into each other so you can unpick it if you to. Otherwise you do end up just entirely locked through apathy of nothing else. ⁓ Raal: Yeah, and I think there is a mention in there about that trust. mean, if you were going to entrust to somebody take care everything for you as well, would it be likely that the person that you trusted the most necessarily had the best software as well? So it's not necessarily a given, right? it might work that Nick Chubb: that's an interesting point on the direction procurement is in ⁓ general, right, when it comes software. know, lot of the work that do at Thetis is, you know, we do kind of pre-RFI supplier evaluations, so kind of helping a ship owner go from, there might be 200 solutions in the market could. suit their fleet. We try and take that down to five that you would then go out and do a kind of full RFI, RFP type process with. And it certainly used to be the fact that the relationship between the salespeople and industry was hugely, hugely important. I think over time that's becoming and less important and being able to really understand needs of fleet internally and understand where your organisation is going and then being able to kind of thoroughly obviously unbiased right coming from a research company but going out to be able to research the market to understand what actually suits. I think the quality of information that's available to be able to make those kind of determinations is better than it's ever been and so actually you know it's much less about who you golf with on a Friday afternoon and much more about the solution, which I think is the right direction, is where we should be going. But I think it definitely, kind of flipping that around to the vendor side of things, you might need to rethink your go-to-market and possibly how your salespeople work, all of that sort of stuff, because that definitely is a changing pattern that we've seen in other sectors and it's coming to Maritime. Raal: think it's changed a lot, it? Like certainly when I came in, the golf thing would have stood quite well, I think. I don't think it's been like that for a while. And I think when it comes to software these days, you're absolutely right. I don't think you can hope to be an effective salesperson if you don't really have ⁓ real in the product itself, which is always a given. I mean, it might sound mad, but I've seen... salesmen who have been and women who've been very effective and haven't known a bloody thing about the ⁓ But yeah, not just, know, but also knowing, really understanding the customer's business and stuff like that. And that is one of the areas where I think the platform play where of the areas where that comes from. Nick Chubb: You Raal: because the understanding as a software vendor is you look at it and you say, well, of course all these things should be in same platform because you're seeing it as one, we sell this, we sell this, we sell this. ⁓ in reality, and I'm not it's just a maritime thing, by the way, but in reality, fragmentation isn't just a software problem. Fragmentation exists all the way through different. ⁓ Nick Chubb: Yeah. Raal: businesses and the way in which different departments interact and stuff and the way we set people up with their own sort of KPIs and their own you know and objectives could well be that they take into account a whole bunch of stuff that might you know it might make sense to exist together. And I think one of the examples where we saw that was when you and I were working on the report that we did on human capital management report, because we were looking at it going like, this doesn't make sense. All these things are related. But departmentally, they weren't related. So ⁓ weren't thinking about it in kind of ways. And so when it comes to making big decisions on software, Nick Chubb: Yeah. ⁓ Yeah. Raal: a good example of that is the fleet management space, because what you'll see with fleet management software, it's a very nebulous term, but what you'll see is someone will be really good at the crewing bit and someone will be really good at the plan maintenance bit, procurement bit might be their star best in show product. And the reason is they normally start with a point solution and they build out from there to try and win, you know, either more share of wallet from a customer. or because the customer is saying, I want less software overall, so I need you to do this bit and that. So they kind of, they try and keep growing what their kind of core is, but then they also try and add these other in. So it's normally the case that, you know, they've got a very good, one part of the solution is very good, the other bit is not so great. And something always kind of gives when those decisions are being made. So if when you're buying in your software, Nick Chubb: Yeah. Raal: If you've actually got different departments that are working and that's typically the case with procurement, there won't be the same people that dealing with plan maintenance, the same people that are dealing with HSEQ or crewing. So the individual platform just might not make as much sense as the vendors actually think it does. So I think that's, Nick Chubb: I once I've got to be very careful on the details here, but I once sat in a boardroom and a nearly almost fight break out between different departments who wanted to go in different directions when it came to software. ⁓ Not even the went in the room. was internally everyone kind fighting around direction we should be going and... Raal: With a vendor, yeah. Yeah. Nick Chubb: Yeah, it was mad. to exactly your point, have one department that's very, happy with the bit they've got and another department that's very happy with the bit they've got. And ⁓ no actually looking at the organization as a full system and how ties together. And that's where all those silos start to be generated. And it becomes so to change. to switch out, particularly when you're dealing with the mission critical stuff like crewing and training and ERP. Raal: And to your point about sales people really understanding the organization, I'm reminded of one particular time in my career where we thought we had a pretty major deal sewn up. ⁓ had a very enthusiastic backer, all all of the boxes, everything he was looking to do within the business. And we didn't even know that there was like this other guy whose requirements were almost, you know. to be considered and I'm trying think how I could describe it giving a game away. fundamentally, this guy had completely different view of how it needed work, completely different view around needing to own it versus a vendor in place and And within about 15 minutes of this, we'd had three, four meetings, I think, at that point, we'd flown out to sort of sign and finish the deal. And it was like this new chap had been invited and I wonder who this is, you know, and just killed it, you know, there and then. And it obvious within a few minutes of the meeting starting that this whole thing was dead. Yeah. ⁓ And. Nick Chubb: Ha ha ha ha. It wasn't gonna happen. Yeah, yeah. Raal: So it's important, you know, it's also important look at it through that lens. And when you think about the complexities of any company, there's a lot of different departments ⁓ who, I said, are all set up with different objectives and different parts they're trying to fulfill. They may be looking after completely different customer relationships terms of the services that they're offering. A lot of these, once you get of an owner's sort of set up, particularly in a ship manager's set up, They may have completely different customers that they're serving or completely different sectors, but when it comes to software procurement, they're all got, know, this is deceptively all sounds like the same thing. So they all get put in there as well. yeah, know your customer, know their problems, understand the market and the landscape and the commercial relationships between them and their customers. And that's going to give you a much better idea of how to take it forward, I think, and how to. Yeah, and whether you're even a fit, you know, that's the other thing. Nick Chubb: Talking of know your customer, can I share my, I had a glimpse the future of generally, but I think potentially the future of crew health over the last couple of weeks. And ⁓ ⁓ up a lot of data about myself as a customer with the company. ⁓ Well, I back. Raal: Thank ⁓ really? Nick Chubb: I came back from Macedonia thinking not sure I can do these multi-day drinking anymore. It used to be that I would never get a hangover and now that I'm my mid-30s I came back much more tired than I was expecting. So, feeling jaded, I booked myself a NECO HealthScan. you heard of ⁓ Yeah. It's N-E-K-O. basically go Raal: Okay. Nick Chubb: very posh building in and you kind of greeted, taken into a little waiting room where it's like ⁓ relaxing music and sort of, you know, it feels like you could be in some kind of a meditation retreat. And then this like attractive medical team takes into a room and basically scans, pokes, prods, like single bit of your body. ⁓ Raal: He's sure this is a medical event. You didn't pick up this number in a phone book, did you? Nick Chubb: Hahaha! can assure you it's legit. you spend half an hour, you're in there for an hour, you spend half an hour getting enough every single test done that you could imagine. So they blood, you go into this big 3D scanner that kind of scans your whole body, captures all the that might be on your skin. test your blood pressure, they scan your heart with some kind of laser, and for the second half an hour, you sit with a doctor, the doctor, she up your results on a screen, and then literally talks you everything that could be wrong with your body, and then what you need to about it. And it was, honestly, was one of most hours of ⁓ my life, both like personally enlightening, in terms was, honestly, was one of the most remarkable hours of ⁓ my life, both personally enlightening, just in terms of you I might need to make some lifestyle changes also ⁓ where tech is taking us in terms of the ability to collect data really quickly and it around instantly to be able to give someone a kind of full view on on like their body and how it works and of you I might need to make some lifestyle changes also ⁓ where tech is taking us in terms of the ability to collect data really quickly and it around instantly to be able to give someone a kind of full view on on like their body and how it works and and I sat in the room thinking about the it's not far off the cost of an ENG1 right to get this assessment done but it is so more comprehensive can kind of see a world in the not too distant future where before going to sea or putting anyone into an industrial environment, they're put through a test like this to check what the risks would be. can kind of see a world in the not too distant future where before going to sea or putting anyone into an industrial environment, they're put through a test like this to check what the risks would be. Raal: Yeah, DNG1. is the well it's like the medical checkup that you have to have as part yeah Nick Chubb: Yeah, mean, it's pretty, you know, it's been near enough. check if you can see colour and they check if you can breathe. And you pass both of those tests, congratulations, you're to sea. ⁓ Raal: really, it's not. I mean, well, I've only done one. I did one, as know, working with OSM Tome. So when I was last in the Philippines, went to they got a fantastic medical centre down there and a very good medical services that they do. I think it's sort of best in show. they, I, they took me through it and ⁓ I kind did it. And it was much more, yeah, was much more comprehensive. It wasn't like, as you described it, by the way, I'll have you know that, but it was. ⁓ Nick Chubb: Right, okay. today. Or maybe I'm being unfair. Hahaha. Raal: I had to kind of crawl through, know, crawl between some spaces to show you had the mobility and everything to do that. And I check, yeah, a whole bunch of stuff. don't think they took any bloods from me, but that's probably a good thing. It was during crew connect. I don't think I would have had enough blood. I probably would have blown the system up. Yeah, yeah, yeah. Nick Chubb: Really? Wow. Didn't do a drug and alcohol test. Raal: Well, this sounds terrifying. I'm not sure I'm ready for that kind of intrusion. don't know if I want to know. mean, how did they? I mean, you're a young man. You don't need to worry about these things. Nick Chubb: Well exactly, but the whole premise, I'm actually going to bring up, I don't know if I can get into it, I now have like this little app with all scan information on But there is my 3D scan of me that they took. Raal: ⁓ my god, that does look like you're in the nude. Nick Chubb: Yeah. So it's it's it. Raal: I mean, I imagine, I wouldn't know, ladies and gentlemen, he's never got that far. But it looks like, you know, it looks like, so he held up to the camera, sort of a, what would I, what would he say? Like a, imagine it was a 3D model, I think, Joe, in the nude. Yeah. Nick Chubb: Those winter nights, Yeah, like a 3D laser scan of my whole body. And even like, you know, the doctor spotted a mole that they weren't 100 % happy with. They took a photo of it and it went off to a dermatologist about halfway through the appointment. By the time I'd finished the appointment, the dermatologist had looked at it and had written a little report back being like, yeah, it's absolutely fine. Don't worry about it. So it's just like... In terms of preventative care and going from being in a position where number of seafarers that their career early because of a of preventative care and a lack of picking up on this stuff ⁓ and either end ⁓ getting a condition prevents them from to sea because they then fail the ENG1 or end up in a position where ⁓ they get sick on because they've got an underlying that was never picked up. quite staggering numbers. And yeah, ⁓ I absolutely see world where our healthcare generally, definitely the healthcare seafarers kind of goes through scans like this. And you can, with the technology, you can get just an incredible amount of data on someone's health. ⁓ Raal: Well, I think you're right to call up the seafarers one two reasons. think that obviously the main one being that you really want to try and pick up on anything that might be a problem before people go to sea when they're a long way away from. medical services and so on. Second point, remember VKand had some very good data on the number of people whose careers end prematurely, people who want to carry on working, who are perfectly capable of contributing fall out of the system because of failing these. So a question of people's livelihood as well. You mentioned costs roughly the same as the NG1. I find that... Nick Chubb: Yeah. Raal: to get my head around but were you able to say ⁓ Nick Chubb: ⁓ yeah, it's It's few hundred pounds. I think an in the UK is about 200 pound and this was like 275. So they're not a million miles apart. I know you can, the ⁓ very basic they can be in the Philippines very, very cheaply for kind of like 70, 80 dollars. ⁓ So it does on where you are in the world. But yeah, it's really a million miles apart. ⁓ Raal: Right. Yeah. And did they recommend any, any, reconstructive work? if you're this thing major done, like that, cause where I always think these things, if it, particularly if it's sort of like, ⁓ it sounds quite reasonable. ⁓ 275. I'm like, is like a sort of like, yeah. Is this like the kind of, don't worry. We'll make up the money later. You know, Nick Chubb: That's how they get you. Yeah. ⁓ Raal: I popped in, I can't remember what was I doing, I was getting something looked at, I was at skin clinic and I think it was like, he was saying it was like a mole or something, was getting checked out. And while I was in there, the guy was like, know, so you're a surfer, mountain climber, and I thought, what's he, where's he going with this, you know? He's like, oh, it's just, oh, no, no, no. He's like, oh, no, it's just, you know, a bit of sun damage there, you know? Nick Chubb: Hahaha Raal: He obviously hoping to kind of an endless amount and if I went to see your friends there I'm quite certain I'd come out of there with sort of you know a few thousand pounds worth of stuff to ⁓ try and salvage what's left of my body. Nick Chubb: Well we can sort that out no problem. What? Yeah, I mean I came out with some recommendations basically around basically around diet and exercise I think we all know just we choose not to follow but there's you'll pleased to know if you're a long-term of Undocked ⁓ I'm not about to dead ⁓ Raal: Good know. Well, I'm glad you're healthy and well. is good news, good because we have sold sponsorship till the end of the year. Nick Chubb: Yeah. What else have you got? Raal: Well, I again, you know, as you know, I your publication and ⁓ story I think we've we've sort talked around it a little bit before around, shall we say, the for organisations to manage more dimensions of digitalisation they are thinking about. Normally these are technology decisions. Nick Chubb: Okay. Raal: And you and I said quite a few times that strong advice is not to overlook some of the more cultural dimensions. And we've also talked about that in the context of leadership and how roll out things like, you know, big technology changes, things like AI into their businesses without really sort of considering cultural dimensions and things like that. So I was very interested to really get an article that Arnell had done. in digital ship, importance of digital literacy in maritime digitalization and straplined digital ambitions be held back less by technology itself and more by the industry's ability to use it effectively. This is an article of kicked off around the ideas of Dr. ⁓ Kyriakos ⁓ Pavlou who the director of the Diploma in Digital Transformation in Shipping. managing partner of Electi Consulting and director of the Electi Academy, Busy Man, who in organisations continue to make mistake technology adoption for true digital transformation, which is obviously these are all big. lovely subjects for us. So yeah, so looking at it and obviously the clues in the name of things is the course director of the Diploma in Digital Transformation and Shipping. He's talking about actually managing some of these new skills and actually actively working on those, which is obviously something that I would support. You we can't expect these things to happen in vacuum. We have to think about how we're going to give people the skills that they need. And I think one of the points he makes in there is that, you know, yes, all this technology coming in may displace people or that's the idea but he's sort of arguing that actually this amplifies at least a portion of the workforce to actually pick these up and make a success of it so is that I mean I presume you're aware of that article Nick Chubb: Yeah, I read it and I think it underpins quite an interesting point around data literacy specifically than of ⁓ overall digital because we're all now good, user interfaces are pretty we've all kind of developed skills to be able to use these ⁓ software, generally speaking, if they're designed. But actually really understanding the underlying behind something I think is becoming increasingly important. There's a couple of reasons for that. One is that in my view, what generative AI has caused is that the cost of presenting information, the cost of presenting data is near enough gone to zero. So you can present any data any which way you like. You can build a dashboard. It takes 10 minutes. You can build a special UI. You can build a particular workbook you can have your favorite AI tool do that for you. So we end up in a situation where you are inundated with data that's coming at you from all sorts of different ways, that's been very, very kind of cheaply put together. And the ability to really understand, like get under the skin of that data and really understand the... both the value that might sit in it and also of the of awkward critical thinking questions about where the data came from, ⁓ recent is it, how complete is it, ⁓ all of those kind of basics, I it's becoming significantly more important because it's so easy to get caught up in the UI and kind of forget ⁓ about the of data behind it and then you consequently may well end up a wrong decision. I've some good, I've been ⁓ on data literacy a topic recently. actually got some good data literacy questions for you if you're up for it. Raal: Why not? I'll humiliate myself live on the podcast. Nick Chubb: Hahaha Okay, alright. so this is a question on the kind of ⁓ law large numbers. Right, so gonna give you a scenario, right? There's a cancer screening. it works well. ⁓ It correctly flags 90 % of people who have cancer that take the test. woman receives a positive result. What is the chance? that she truly has cancer. It correctly flagged 90 % of cancer cases. Raal: what your question is... ⁓ Nick Chubb: What are the chances that she actually has cancer? Raal: Well based on what you've told me I would have no idea. You said it's 90 % so 9 to 1 or something like that? I don't know. Nick Chubb: The actual answer is a little under 10 % chance that she has cancer because Raal: Okay. Nick Chubb: the chances of someone actually having, of 100,000 people that might get screened in a cancer test, the number of people that actually have cancer might be something like 1%, which means that nearly every single positive result is actually a false alarm if the test is 90 % accurate. ⁓ it's a example of like some of the counterintuitive things you get when you're looking at. statistics of any nature. Same with averages. the ⁓ is a terrible, terrible descriptor the vast majority of data sets. And yet we use it all the time without really thinking about it. ⁓ And we also the other one that comes up all the time ⁓ is, know, and causation. You know, the that ice cream causes sunburn because sales of ice cream at the same time as cases of sunburn hospitals, at cases in hospitals. Raal: Yeah. Average is a terrible. Yeah. Yeah. Yeah. Yeah. Nick Chubb: So back to that article, like getting the real kind of fundamentals of data literacy, how statistics work, all of that sort of stuff under organization's in general, I think is now becoming far more important than investing in, you know, on specific digital tools. ⁓ Raal: Yeah and I think that it becomes much much worse when you start to do some of the things that we're trying to do these days which is to put together vast sets of data and have them work on multiple metrics and then give you you know, something that's kind of a fairly crude measurement of all those things. Because if you've got any mistakes within there, or indeed the averages are biased, or it just clouds the picture considerably, it just gets worse and worse and worse the more you kind of work on it and multiply it. So, the other thing is that I've... Nick Chubb: Yep. Raal: I think is really understanding what the data really represents. And that's your point, I think a little bit towards the causation is inferred by that as well. So we and the reason I say that is because often use proxy measures to think that, you know, I'll give you an example like we we quite often use the measures that we have rather than the measures we actually want to know. So if you take if you think about competence, for example. What we'll typically gravitate towards with competence is something that's easier to manage. Like did, people able to complete an assessment? Which is actually not a judge of competence at all, but it's a proxy for it. And then once that then gets embedded into your dataset and people can then run away with that, you know, and, and so it's really important that you understand what the underlying data is doing. The other thing is to say is when we're dealing with particularly things like AI where Nick Chubb: Yeah, yeah. Yeah. Raal: hallucinations will occur where the system will start to kind of do its own out and come up with its own hypotheses and stuff like that. It's even more important that you understand the data so you can spot where something just doesn't. look right, feel right and actually interrogate the system and say, how did you arrive at that? Let me see you're working out. And then you're able to kind of get a sense of whether or not you trust it and whether it needs further verification and stuff like that. all kind of parts of really understanding data and what data is actually doing. Nick Chubb: Yeah, think on that point around competence. The kind of difference between compliance, competence, compliance certification, competence, confidence, know, all similar words, they have very, different meanings. ⁓ We actually published some with OneOcean, actually it came out a couple of weeks ago, looking into the issue around, the gap between design and reality. We've done it as a research series. The first looked at the ⁓ gap between design of software and the of they're actually used at sea. out it's quite a big gap. ⁓ then the second looked at the of crew training and the of how competence is actually built and applied at sea. I'll give you a stat from that report. We did a survey and 82 % of described the current training and certification models that are in use across industry today as either neutral, poorly aligned or very poorly aligned with reality at sea. only 18 % of people believe the existing of training and certification models that exist today are well aligned genuine life at sea, which is just a slightly kind of mind blowing stat. ⁓ Raal: Yeah, yeah, I can believe it. Yeah. So, I mean, just coming back to that one there, other points that it made, I think I dealt with these in the preamble, but it was really talking about that digital transformation being a people challenge as well and really thinking about how you're going to do that. So when approaching the way in which people have relationships to data and what they do with that data and all that sort of stuff is part of that as well, not just the people that are actually hands-on in the data, let's say, well. So from that point of view, it recommends, I think, in the article that data literacies should be developed across all roles and with the IT teams. I think that this is something that's massively overlooked in our when we talk about competence. I think there's a big, I think the reason for it is that there's this idea that new generation are digital natives and the technology really needs to keep up with the, this new generation of digital natives who are all like born in the saddle, let's say, as far as digital is concerned. Simply evidence for that whatsoever. because the difference, there's a massive, massive difference between being able to fly your way around your mobile phone and all the apps and games and stuff that you there on a daily basis and actually using the kind of software systems that you'll find on modern ship or indeed in any workplace kind of commercial software. They're really different and they require different disciplines, different understanding, all that sort of thing. We can't infer that people, because they very savvy and know how to use Facebook. that or TikTok or whatever the equivalent is. therefore infers that they're A, literate in using the kind of software that most people would be using on board ship and also B, that they understand ways in which you, they understand required of working with software in that kind of environment and how you. how to use it safely and securely and all those kind of things as well. I think it would benefit us to think about that as a skill. I haven't checked. I'm going to check off the back of this whether there's anything about digital literacy that's currently timetabled for SDCW revision. I've seen lots of stuff around technology, but I haven't seen it as a discipline, as a data discipline. But it could well be there. Nick Chubb: Surely there is. Wow. Yeah. Raal: Answers on a postcard. Arvin Natrojan from ICS, if you're listening, he'll be to tell me. So I've seen so many of the papers that were put in and I can't remember which ones got kicked out. Although I did, ⁓ I was to just talking about this week, about the revision to STCW, of the things that was declined was mandatory training. Nick Chubb: Mandatory training on what, sorry? In yoga. Raal: Yoga. This was our submission for the SDCW that they should be open. Nick Chubb: Who put that forward? Raal: I mean. Nick Chubb: That is a way to put off the vast majority of seafarers ever going back to sea, it? ⁓ Raal: I like bit of surely some of them. tried it on holiday once, emphasis on once. yeah, I like idea of it. I don't ⁓ think you live long enough to kind of master this sort of stuff. But yeah, so if there wasn't room yoga, maybe it got pushed out by digital literacy or something like that. Nick Chubb: How did it go? Is it worth just... It Yeah, I mean, if yoga makes it in above digital literacy, will be, I I will scream. It's worth just for anyone isn't aware, SDCW is the standards of training certification of watchkeeping. So it's the kind of international convention that governs how. seafarer training is managed and kind of the levels to which the various ranks and roles on board need to be certified to and kind of what competence means. It was first introduced in the 1970s. It was then amended in the Manila amendments of 2010, I think, and then those amendments came into force in, I want to say 2019, and which now in line for a major overhaul. you'll guess when I'm talking about regulation that came in the 70s, was amended 2010, technology has moved somewhat. And one of the that STCW has and the IMO has generally is regulating for ⁓ technology does not exist but almost certainly exist in five time as a set of conventions that will be, think, coming into force. I think ⁓ the will run... ⁓ through to October and then I think the actual meeting doesn't happen for another two or three years to kind of adopt the convention and then you get like a sort of 10-year adoption period. So we're talking about stuff that will be maybe in the mid 2030s will actually become law that the industry will have to follow and you know who knows what the technological picture will look like and they will stay law until well into the 2050s possibly the 2060s as well. So it's a really really difficult thing to regulate, when you consider this is a kind of consensus convention, you need to get every country to agree to it, but have rare opportunity, these papers that are being submitted that Ryle's talking about, ⁓ is a very rare opportunity for the industry to make an intervention what they think should be regulated ⁓ and how it be regulated and obviously someone feels strongly about yoga. Raal: Well, yeah, you're absolutely right. The other thing just to kind of add to how much worse it is than it even sounds is that Nick is talking about this coming into force in 2035, which I think is good estimate. I've heard some people saying 2029, which is for me is too optimistic. The papers are going in now. So it's already going to be just I have 10 years out of date. with what goes on now. Now, obviously in the 70s, that was true as well. So we talk about STCW, the PREA prior to the Manila Amendment. So it's even older than that because it was sort of fixed some period before that. And the Manila Amendment took care of some of that. But again, they were, I you said 2019, it was 2010 that they were looking to. Now, let's consider the difference or the changes in maritime and in what was required from mariners from the 70s to 2010 from 2010 until now. And I think that's where you start to see the problem. And so we're actually looking at trying to update this mechanism based on the way that it's always sort of worked in the past. And I challenged that myself and said, well, actually, is there any point in an update to SDCW? Don't we need to change? the mechanism, the instrument, you like, of how we do it and how we manage this going forward. So we don't have this same problem. And I was told in no uncertain terms, forget it. It's never going to happen in a million years. You're never going to get that through. So it is a problem because we are, it's a bit like we've talked about with the education system. We're trying to prepare people for a world in which we can't get a really solid steer on. And the job of being a mariner could be very, different the time this new version of STCW, well, could be, it will be very different by the time this new one comes in. And need to take into account these kind of skills and a good many others as well will be around assisted navigation and all those kind of things. You look how, how, how ⁓ Nick Chubb: Yeah. It will be. Yeah. Raal: what a big deal it was to make a jump from paper charts to electronic charts and that sort of the shockwave associated with that and how different that then became and there's other there's lots of examples as well and yeah it'll interesting interesting thing to watch see what happens Nick Chubb: We've got time for one more. Raal: Well, ⁓ have you got one more? Nick Chubb: I can do one more. it's nothing to do shipping. ⁓ Yeah, go on, do that. ⁓ Raal: just wanted to talk about Fable before it was too late. Now, one of the things when we talk about like, how does the future, you know, how is the future going to look and the future of work and all that sort of stuff, clearly, once again, the biggest sort of impact that we've seen. And one of the things that shipping will have to try and get his head around in terms of what happens with SCCW is what is, what is... AI going to do to all of that. Now, you and I have been playing around with this for a long time. You've done, you've embedded it into some workflows. I've recently upgraded a bit in terms of, because I was getting to the limit of what I could do with it. So I've experienced some of these tools, but we saw, I think it was about two weeks ago now, we haven't had a chance to talk about it, but I think it's well worth us putting it on people's radar in case they missed it. And is the... the brief, the all too brief life of fable. so fable was the, was it, it, what's the correct word for it? it a framework or? Nick Chubb: foundation model. was Fable 5.0. I had a wild 48 hours with Fable. It was incredible. And then yeah, 48 hours after launch. this is... Raal: Yeah. I just think it's sort of ⁓ a world that people can understand because they might be thinking, well, if it was new, why is it version five? Because this is is ⁓ tool, So people might well maybe have heard almost certainly have heard of Claude. They will have heard of things like open claw, which we had ⁓ Peter Rossi on, taking us through agentic AI and what that's going to mean. And so this was this was like the kind of the new kid on the block, basically, from that perspective. Right. ⁓ sorry. should say sorry. We should say because I think we might have talked about it. Mythos was the big one. Right. So mythos was going to be this huge kind of it's the clues in the name sounds like the Zeus proportions. Nick Chubb: Yeah, so it's... Raal: was they pulled it themselves and swap it on the basis that this was going to cause too dangerous from a cyber yeah ⁓ but it was going to be that it was going to cause major security breaches because it could be used by a bad actor to to infiltrate security they then brought out fable yes Nick Chubb: too dangerous. Genius marketing from Anthropic. Fable is of the same family. So the of general set of models that we have open to us, that are open to us at the moment is Opus 4.8, which the top Claude model. Fable was of us version five. So it's a kind of sister to Mythos, it was supposed to be unrestricted available to And yeah, I a really interesting weekend working with it and then came in on the Monday morning and to the notification that it had been withdrawn with almost no explanation. And then go and look it up on the news. And it turns out the US government wasn't happy with how advanced it was when it came to cybersecurity. to kind of jailbreak it and make it do things it's not supposed to do like go and hack government systems. Raal: Yeah. But wasn't it that they withdrew it for outside of the US? Nick Chubb: It's everywhere, it's global. Because they couldn't, decided they couldn't reliably stop it in the US while keeping it available internationally, because VPNs and all that sort of stuff. So they just decided to just pull it, for the being, pull it and have to for maybe at some point in the if they can solve these jailbreak issues, it'll come back online. ⁓ Raal: Good. Before we get into that means, what different about it, having had 48 hours to play with it? Nick Chubb: ⁓ felt with these models like a narrow number of fields I'm smarter than the model. You know, the majority of things, the model ⁓ get there ⁓ than me and quicker than me. This the first time I was able kind of work with something and just be like, ⁓ no, actually, I'm toast. Humanity's I may well give up. And also just the way in which it the of volume of work it was able to produce to a genuinely exceptionally high standard was... completely mind-blowing. And so yeah, I was able to produce fully working web from the kind of chat interface, which bridged, not even months ago, six weeks ago, was just not possible. So yeah, it's ⁓ Raal: Yeah, and we had a couple of on our WhatsApp chat groups, had a couple of other people validating very similar kind of things from this. Now, this has spooked people because of the government interference into this, the government deciding when the pipe can be switched off and who's allowed access to what. And I believe there's something very similar happened. I'm not sure if it's... a firm request or an outright ban, I think the latest chat GPT has also ⁓ ⁓ Which one is it? Sorry. Nick Chubb: Yeah, 5.5 I think. Check GPT 5.5. Raal: Yeah, has also been back as well. And so I what's spooked people genuinely is the idea that these are commercial models which are available and we building infrastructure around the use of these tools. We're building businesses around relying on these et cetera. And the genuine possibility that that could be off. And the examples that have been given are, you know, restricting power, restricting water, restricting these sort of fundamental kind things. And so is when we were talking earlier on about the vendor. vendor lock-in and all that sort of stuff. Again, it gives us pause for thought there in terms of, you know, can we rely on technology in the way that we think we can fundamentally it's being controlled by states. Nick Chubb: Yeah, if you think of AI as... a lightweight infrastructure layer, very, similar to kind of broadband internet. know, AI could be used anywhere in your business, in any process, in piece of software. That's of the world we're moving to. It would a bit like ⁓ turning off internet, which, you know, if you think about what that would do to global commerce generally, you know, they wouldn't able to listen to this podcast, for example. It would be a total, total disaster. Whereas turning off AI, we're probably not quite there yet. we've definitely reached a point where a lot of people are starting to feel like they could depend on these models and the models are becoming very dependable. And so being built the infrastructure and becoming utilities in same way as you mentioned, water, energy, broadband I think going to end up in same kind of way. What's really interesting about this is that for the two years, Everyone in the US in the kind of AI acceleration camp has been making fun of Europe the EU for taking the kind of slow and steady, we're going to properly regulate AI approach and like, you know, in the US it's built, baby build, we're just going to go for it ⁓ and left to get on with it. And then ⁓ it is US government that is the one that's the kind of first to just outright shut down a major model ⁓ with almost due process, very, very little to the organisation, to Anthropic, who who are behind it. that's kind of thrown everyone for a bit of a turn. ⁓ building AI models in the European Union is actually looking a little bit more attractive. ⁓ Raal: Yeah, and I think as well, it's also significant in the sense that, We did an episode where we were talking about the fact there was that great article in The Guardian that was talking about the AI arms race. And it was basically saying all regulation was off and everything was kind of, you know, peddled to the metal. We've got to get there first. We've got to get to agentic AI. because if we don't, the bad guys will get there before us. ⁓ And this is kind of like there was a huge amount of fear and Nick Chubb: Yeah. Raal: about the fact that the guardrails weren't there because we were just sort of steaming ahead to try and, you know, to try and reach this event horizon sort of thing. And this kind of speaks a little bit to that having been reconsidered, let's say, and that governments might actually start to say, do you know what? We're not confident about this, actually. We're kind of concerned about this. And it's interesting that it was the... sort of takedown of infrastructure and cyber security side it that was, you know, that's been more considered to be the real and present danger. Whereas, you know, a lot of other people were talking about the ability to create bio weapons and all that sort of stuff and all that kind thing. And the point that you made there, I think is a good one as well, which hasn't really been that bigger part of the debate before, which was idea that these models might be prevented, might have some security in them to stop them being used for those kind of purposes. the fear being that they could be hacked in some way and they could be overridden do some things that they weren't there. So I think it's a really significant development that has happened. Nick Chubb: Yeah, I fully agree. Right, that's a happy ending. sorts. Humanity will have a role. shipping will still have a role, so we're all right. Raal: Good. It is, yeah, of sorts. ⁓ anyway, yeah, thank you for sharing your intimate details. ⁓ Nick Chubb: You're very welcome. I hope you enjoyed it and a lovely week and ⁓ look forward to doing it all again soon. ⁓ Raal: Yes. See you later, bye.