Old Boats Meet New Bots - How AI Helped Fix a 1970s Fishing Boat in Alaska

Jason: I was way up north in Alaska with very shitty internet connectivity. And I'm using Chat GPT to solve a problem with a 1970s fishing boat. that is just such a cool, amazing, practical way to do this. what else can I use AI for in this scenario?

Jeremy Grater: Welcome to Robots, the podcast that tries to help you be a better human by being smarter about the way you use your technology. Jason, you just got back from a luxurious trip to Alaska where AI certainly came in handy.

Jason: It sure did. So I spent two weeks in a while. neighbor who's got a commercial fishing boat and it's up by the Kenai and we fished the upper for salmon, sake. and it was awesome. And it's a combination of sport fishing, it's a combination of commercial fishing, work in the boat, kinda like the deadliest catch things. And we started to work on this big project with getting the b the boat repaired. And lay down and the boat rocks me to sleep and it was a great night of sleep. And the accommodations were not comfortable. They were not plush. but it was great. And I wake up at five AM, my buddy's getting his boat ready and the other the other crew hand is sleeping in the other bunk there. And the guy leads down and goes, Hey everybody get up. Time to go. We gotta get start getting things ready. So we're like, okay We wake up and we go outside and I take a look and there's just a beautiful sunrise going on. You can see the mountains all around you. Absolutely wonderful. So we start motoring out and fishing starts at 7 a.m. and we get to this spot and the co-captain's like, Hey, this is a great place. Let's let's hang out here for a few and we'll drop. So, like right at 7 a.m., we drop the net. Roll this thing out, sit there for about two hours, and Pull it back in. We pulled in like seventy, seventy-five fish. It was great. you gotta know how to pull them out of the net, like move things through. I was in the way. and I had to learn, right? So they're trying to teach me as they're going through this. And it's it's awesome. You know, I'm having to exactly.

Jeremy Grater: You're the you're the hold the flashlight guy. Show let let us see over you. Hold the flashlight, yeah.

Jason: Yes, I I'm I'm definitely the green horn. so there's there's fish all over the deck and I'm reaching down and picking them up and I'm picking them up with like both hands and trying to be gentle with them. In the deck, Handlexy Mane goes, Yeah, don't. They're about to be meat. Just be violent. Pick them up and throw them in the ice tank. You don't have to be soft. So I'm like, all right. And they started picking these things up like a bear. Tossing them in the tanks. It was awesome. We get to that first run. And the boat captain goes, Hey, let's motor out a little further and get to a different spot where I know there's some interesting cross currents and where fish Tend to circulate because there's a thing called the flood when you get to high tide, and the more of the fish have moved their way back in and they come up in the water as opposed to sitting on the bottom. And quite often you're looking for jumpers. And you're looking for these schools of fish that are moving around this direction. So we weren't seeing any of that. Because everything was running so deep, and this is like one of the worst fish harvests on the Kenai in decades. We get ourselves into this spot, and the captain comes out and goes, All right, roll the net out. We roll the net out, get it dropped off, and you you can let sit there and soak and like float around, you can tie them up to a cleat on the boat and drag them around or hang them off the back of the reel reel winch and have them move around. But there's just different techniques to use it to get the net to stretch and spin and turn in different directions to try to improve your harvest. So we're out there doing that. We're about forty-five minutes in, and the weather starts getting tough. typical swells there are like one to three feet. we were at about six feet and the wind really picked up. And we were no longer just getting waves coming like front to back. We were getting waves left to right and kind of kitty corner as well, because we were in a weird tidal zone and the winds kept shifting and turning because we were sitting. And they went from six feet to nine feet real fast. And the boat was just going up and down. And and nine foot waves on a thirty-two foot boat, it's big. Like you slam, slam. And when

Jeremy Grater: yeah.

Jason: they're really tight together and they're not just up and down, but they're left and right, and you're you get caught in these wave motions and you just get tossed. Well, the boat captain looks back and goes, Hey, It's getting rough. Let's pull the net up and we're like, Yeah, no shit, it's getting rough. We're both banging around on the deck, you know, trying to ride bulls out there.

Jeremy Grater: Getting rough. It's been rough.

Jason: Yeah, it's been rough. So, we go to back the net up and the hydraulic reels do not turn. We're like Well, what do we do? And the other deck can't, like, well, I guess we can try to hand pull it, but the current was so high, there's there's no way. So talk to the boat captain and he goes, All right, I'm just gonna back up over the net, and there should be enough hydraulic pull to be able to get it to rotate through and you guys just sit there and pick fish. So normally something that would take like ten to fifteen minutes to pull up, took forty five minutes. And it's not because we were stopping to pick fish out, it's because it was just a long process to get through it. And we're getting thrown around. Everything is off the shelves in in the boat cabin. you know, nothing stayed. I mean it was it was tough. we get through that and we motor back and I'm just exhausted. The other deck can goes under the into the cubby and falls asleep and I'm up on the the crew chair behind the captain's seat and I just lean my head against the wall and fall asleep. Like you would do on an airplane, which I don't sleep on airplanes. I slept on the boat 'cause was exhausted. so we start motoring on and it's about another ninety minutes of just being tossed on the water and the boat captain screaming it's weird because it's it's dangerous. Like shit's happening. And other people like this weather came out of nowhere. So we go back and we get in and it was like eight o'clock by the time we get up to the mooring buoy and and get a boat actually set up so it doesn't move so we can actually just take a look at the hydraulics. And we're looking at it and we're like, well, we can't really tell exactly what's wrong, but we can see that it it it seems to let the net out just fine. So the hydraulics working in the opposite direction are okay. It's when it's trying to pull it back in that's the issue. And we're like, okay, well, let's go figure out what this is. So we go to like these You know, the thousand the millennium of experience that all these old fishermen have in this this little camp and started asking questions. At the same time, I'm on Chat GPT, just like, hmm, all right, let me see what qu let me see how they're asking these questions. Let me see what's not figuring out. Let me see what's in there. And like the things that ChatGPT was asking me is like, okay, I kind of see how this is. What's the earmake of the model of the boat? What's this? What's this? Nobody knew because these boats are friggin' old. And I'd taken a picture of the nameplate of the It went through and found the manufacturer of the boat from the nameplate. it was made in Burnaby BC in the seventies. And it's like, I think it's probably this and this, and that manufacturer isn't around anymore. And we're talking about the hydraulic pieces, and the old guys there go, so you're gonna need to find a replacement engine for it, or motor for it, a hydraulic motor. most likely, because it's probably threaded out. but you might have more than one problem. It's like, all right. So I'm on ChatGPT going back and forth of this, and ChatGPT is like, yeah, you can probably find this part here, here, and here. And these guys go, Yeah, you should call this place in Homer, Alaska, because they probably haven't drived in and get it, which is one of the places ChatGPT looked. I'm like, hmm, all right, this might be pretty cool. So we go down there and we we go down there the next morning, wake up, drive down to Homer, we go to this place called the Salty Dog Saloon, which is awesome. If you ever get a chance, it's well worth had great fish and chips and picked up this part. And take the part, get it back, plug it into the boat, hit the hydraulics, and same problem. Alright. Well the next step is to start pulling the hydraulic lines and trying to understand what they're actually doing and seeing if there's valve problems or anything else. Well the boat captain and the deck hand are they moving those pieces out and I'm inside holding flashlights and cleaning the boat, as the greenhorn does. And the boat captain goes, Hydraulic handle break. And we're looking at this thing and the boat captain goes, you know, I think the right thing to do is to just kind of, you know, reverse the polarity on on the hydraulics, put the out to in and the end to out, because letting the net out works just fine if we can flip those things around, we can start dragging stuff back in, but we could still gotta do something about this handle. So we get off the boat, go consult the oracles again. They're telling us, you know, I

Jeremy Grater: Please tell me they are referred to as the oracles by by

Jason: They should be. They're not. They're not, but they should be. I'm gonna start referring to them with that. We call it Camp Dignity, which is pretty funny because it's just run down shitty trailers and cigars and then bottles of whiskey everywhere. It's crazy. but we start going through and talking to these folks and I'm at the same time doing the same thing. Like I'm on Chat GPT asking these questions and the oracles are coming back and saying things like, Yeah, I'm not sure if you can do that because the braking system and the hydraulics, the way these things are built, sometimes they're There's interlocking linkage inside of them and you can't change the polarity for certain runs because of the way it circles itself back to the system. And some of them are like, well, no, it's it's before all that stuff was made, because it you know, all that stuff was made in the eighties. And nobody has a definitive answer because there's no manuals for any of this shit online because all these things aren't made anymore. They haven't been made for twenty five plus years. I mean, realistically they were before the internet was a real big thing. so I Pink Jack GPT. I'm like, okay, sh can you reverse the polarity on this looking at these pieces, looking these things down? And it comes back with, yeah, like I'm like 99% sure that you can do this. And then here's the steps you want to take to see if it works and to make sure it doesn't break anything else. And it gave you an ordered list of what to do. That was better than what the oracles were telling us. Like, huh? So obviously not a replacement for them, right? Like I didn't know the right questions to ask. But listening to them ask the questions, I could plunge that into AI, and it gave me better answers than what the other people were giving, or sorry, gave me faster answers than what the people were giving and it helped me to validate the other answers that were there. And then we got the part, this this lift handle. And I'm like Nobody knows where to find one of these lift handles. I bet if I take a picture of it and tell it what's going on, I I bet it I bet it's helpful. So I get the handle and I take pictures. And I am gonna share those. Here is the first picture that I showed it. And then here is the second picture. The same thing, slightly different angles. Here is the third, a slightly different angle. Here is the fourth, a slightly different angle. Not too much information there.

Jeremy Grater: Yeah.

Jason: It told me it could read the serial numbers that I couldn't read very well when I'm trying to look at it in light. And I'm looking at the picture right now and I can kind of see it. It went through and it analyzed what was broken and if you look you can see, you know, the parts up top are are clearly cracked and and not working right. Parts of it are rusted through. there's a beautiful picture of of of of

Jeremy Grater: Lovely. Lovely scenery.

Jason: the moon, believe it or not. And then it asked me to take pictures of the mounting spots. And here's the mounting spot where it broke. Like it that's all I really gave it. You know, and a a couple of different angles on it. And it came back and it told us the part, number that it thought it was, who the manufacturer was, and it started searching inventory globally of what we can of if there's any of these parts available. Hint, there's not. it says you can go to a boat salvage yard and try to find them, or you can take these other existing parts that are manufactured in mass and you can try to modify one of them and use them as a replacement, and here are the welders that are close by you that can do it for you.

Jeremy Grater: And how many people at this point are like, Fuck it, it's it's broke like game over. There's nothing I can do.

Jason: nobody said that. Everybody, I mean this is remember, these are fishermen that have been doing this for decades. So you you're always solving problems. You're always fixing things. You're always moving through those cycles. So nobody said that like at all. Everyone was convinced we can figure out a way to solve this problem. What this thing did was confirm the break on it, how those pieces would work, how to get those pieces fixed, what parts to go and pick up, and how to go and do those things. All things that these guys probably would have come to It just would have taken them significantly more time and it would have been done by committee. And, you know, ultimately speaking, we went through, we we didn't find our placement handle, we didn't have to. The boat captain went through and used some different epoxy steel resin to go through and kind of rebuild it, cle clean things back up, because we knew we weren't really gonna have enough time to get this thing bought and refabbed and everything else before we had to go back out. So we take it, put it on the boat, flip the hydraulic lines, and boom. Everything's working. It's like, right, this is awesome. So we went out and did some sport fishing the next day, then the next morning or that that evening we went out and camped again in the boat and rocked back and forth and had a decent day of commercial fishing on Thursday. Didn't pull a ton in, but you know, again, worst sock ice season a long time. But yeah I was way up north in Alaska with very shitty internet connectivity. And I'm using Chat GPT to solve a problem with a 1970s fishing boat. It was awesome. Like that is just such a cool, amazing, practical way to do this. So of course it leads to the next question, what else can I use AI for in this scenario? Hmm. Alright, well, what about going out and buying an unmanned aerial vehicle, aka a drone, that can take pictures and can start looking at the water for fish jumping or changes in color when it like looks like their school of fish. And I start looking at all this stuff and everyone's like, nah, that's gonna be like hundreds of thousands of dollars. I'm like, nah, I don't think so. I think they're like 10, 15 grand. So I went through and I found them. I found a whole bunch that were like right in that price range that have most of the tools that you need to do this. Like there'd have to be some software already, but nothing Claude couldn't do in in two days. And everyone's like, This is awesome. You know, they're like, We used to have an airplane that would fly around and do that for us, but then they stopped letting us do it about twenty years ago. I'm like, Hang on. So I took the whole idea and I put it in chat GPT and it kicks back a response and the response says, Yeah, you can't do any aircraft over the water an hour before, during, or an hour after the opening of those fishing seasons. No aerial vehicles are allowed. I don't know why. I didn't look into it, but it is an Alaska regulation. Probably because it was too easy for them to find fish and take them out. So

Jeremy Grater: I was gonna say, is like a is it like pr like preserving competition among fishermen? Like if one person's got that tool

Jason: Maybe.

Jeremy Grater: and somebody else doesn't, may maybe it's that.

Jason: I don't think they're that concerned about that. I mean if that were the case then a lot of the boats that were on the water wouldn't it be allowed in the water. For some being for really good and some being yeah. I mean like there were several boats that had blood motors and transmissions this year. And I'm like, is this normal? They're like, yeah. It's just such a crapshoot. But it was cool 'cause I could use AI in all these different ways and really practical ways to figure out how to make things work in meat space in a more productive way. And yeah.

Jeremy Grater: That's the thing that's so cool about it is that it's like I you know, at first I'm thinking like, who's the boat mechanic who's gonna be out of a job because ChatGPT can tell you how to fix these things? But this sounds like a bunch of people that wouldn't go to a boat mechanic. These guys are self-sufficient. They're gonna they're gonna stick some gum and some duct tape on it and make the thing work one way or another. This is just fast tracking them to a to a better solution to problems that they would fix on their own, much slower, much clumsier, much more trial and error rather than just here's a step by step. Here's how to get back in the water today.

Jason: Yeah, exactly. And and these guys whole thing is like they're constantly troubleshooting problems over and over and over again, at large scale and small scale. And there are problems that can happen out there that can cost you your life. And you know, having the hydraulics blow like this wasn't one of those. Like if the hydraulics went, fine, okay. Then worst case scenario you cut the net and you just get out of there. But if the engine goes and you know, you're suddenly twenty miles out to sea and nobody can find you and the wind chops and the water's forty eight degrees, and the boat takes on a little bit of water, you're all gonna Like that's just what's gonna happen, you know?

Jeremy Grater: But that's that's more the natural order of things. It's not the robots coming to gun us down in our sleep.

Jason: It that that's true. I mean, I don't know if fishing boats are the natural order of things, but it's definitely historically something that's occurred for a long time.

Jeremy Grater: Yes. Yes.

Jason: And and it's it's fascinating because Google Gemini is on everything, right? So as I'm going through and I'm searching for like restaurants or things to do and Kenai or like what are good fishing spots, even the AI then is picking up and going, yeah, like you should look at this here, here, and here, here's what we've seen. Here's what the fishing reports are from these different areas. Here's what the fish count is. I'm like, dude, this saved me from having to go to like 15 websites. No, do I trust it? I don't know. I'm gonna follow that, but it at least it aggregated the information in a way that I could go through and track. But you know, it's it's cool. I gotta put on my resume that I'm now a I I've been a commercial fisherman, which is fun. I had to buy a license for it. and I've gotta go and experience something new and different, and I gotta see how AI really did make a difference in this kind of industry.

Jeremy Grater: And I'm excited for that because of all the implications that it has for helping humans be better humans, which is what we're trying to talk about here. I'm also excited because I think this is two episodes in a row where we haven't predicted the the end of humanity thanks to AI. So we're out we're on a good track. I don't know if I can handle this much optimism. This is crazy.

Jason: Yeah. We're on a roll. We're on a roll. You've just jinxed it.

Jeremy Grater: All right. Well, I did not expect just keep going. Yeah.

Jason: We should probably go ahead and record another one right now. Just get out of the way.

Jeremy Grater: well, I did not expect to learn how much AI can help the commercial fishing industry. That that seems like one of those jobs that like a computer would never touch, but it's exciting to to hear how you were able to do that with boots on the ground or on the boat in this case. very cool. We hope to share more stories like this of of AI and technology improving the lives of human beings like us and like you. if you're into that sort of thing. Give us a follow. We're we're our website is Brobots.me. We're on all the major podcast platforms and of course YouTube. all those links are at our website, probots.me, where we'll be back next Monday morning with a new episode. Thanks so much for listening.

Jason: Thanks everyone. Bye bye.