Delian Asparouhov: cockpit miscommunication is the hidden cause behind aviation's crash wave

talking about to try to fix the the business model of the internet. Very cool. Well, we have Delhi and Aspuhof on the show for the second time in two days. There he is. I'm surprised. Oh, he had to he had to go remote cuz he didn't want to get blasted with a confetti cannon again.

I legitimately my heart jumped when that happened. And I was like, I'm dying. This is the end. This is my final moments.

I I was saying to John after I I I didn't really mean to get you that bad, but I was imagining a scenario where you had like a concealed carry and you just like did a quick draw and like caught me twice and then that was just the end. Like I'm a Florida man at heart. You know what I mean? He's a Florida man.

He's a Florida man. It was wild. Anyway, thank you so much for stopping by yesterday. It was great. Um let's run through uh what's going on. By by the way, nothing says SF and YC are back. Median at YC demo day. Really, really symbolic, you know. Yeah, definitely. I mean, doing a good job.

City was almost dead and AI saved it or AI plus Gary Tan, you know. I think, you know, Mayor Daniel has a lot to owe to those two. Yeah. Yeah. It's just all it's YC presidents all the way down. It's Sam Alman and Gary Tan. It's true. It's Yeah. It's just YC presidents and everybody will go, you know.

Who's Michael Cyel bringing for San Francisco? Uh yeah, yeah, let's get Michael in the in the conversation. He he can work on converting Alcatraz or something. I don't know. By the way, I'm surprised you could even join this. Apparently, Google is just down.

I mean, GCP is down, but uh I you know, all of our invites are based on uh Google calendar. So, um glad you could join. Hopefully, it was cash. Are are are they sure we're not just under uh political attack? It's probably partially because they're trying to stop.

They probably said, "Hey, we're trying to take TVPN off the air. We're trying to take Delian off the air. Let's just take India. They're they're afraid. I I'm afraid. Let's keep it Let's keep it PG for anyway. Uh yeah, I mean uh the it's is a ve very sad story about the 787 crash. Uh this is the Dreamliner, right?

The first new or the most recent new major plane from Boeing. It's not as big as the 747, but it's super efficient. They're turning over the uh the the fleets. It's supposed to be kind of the latest and greatest. I remember the news around it launching.

I don't even know if I've ever flown on one, but uh it was not the it was not the 737 Max story that people were risk worried about. It's a completely different platform. Shouldn't have the same problems. Um but at the same time, there's this terrible crash. There's been a lot of crashes.

Is there something bigger going on here? Can you just give us an overall state of the union on aircraft safety? anything you know about this crash? What even questions should be we be asking about this when something like this happens? Yeah, I mean uh to like set the stage, you know, I am a you private pilot.

I've gone through some level of training. Obviously, not to the level of like, you know, commercial pilots. Not to the level of Nathan Fielder.

Not to the level of Nathan Fielder, but like you know, you you do get to train alongside a bunch of guys that turn into commercial pilots and you see what training they go through.

And while the investigations both for the DCA crash, so the Washington DC, you know, uh, helicopter plus plane crash, uh, and then this obviously India accident obviously only just happened, both those investigations are, you know, sort of ongoing.

But I actually go back to like the Nathan Fielder, you know, sort of, you know, latest season where um if you look at, you know, sort of crashes, including, you know, sort of these two and a bunch over the course of the last 20, 30 years, a significant component that fed into the crashes was [ __ ] you know, sort of uh was you pilot, you know, cockpit communication issues where um you know, a co-pilot identifies an issue um speaks up to the captain, the captain basically largely dismisses it and you continues down the prior track.

So there's like the DCA investigation is not yet, you know, sort of fully released, but there was early uh audio that was uh uh early, you know, sort of transcripts that were shown that I think it was something on the order of like 45 seconds before the crash.

The co-pilot literally said, "I don't think we have the plane in sight. We should basically be going down into the left because I'm not sure that we're, you know, uh going to be, you know, uh dodging out of the like, you know, approach path.

" And the captain, it seems like, literally just completely dismissed, you know, this piece of information. And if they followed it, they would have flown in exactly the direction that they needed to fly in order to avoid the plane.

And so that and then if you look at you know sort of today's you know sort of India's you know sort of 787 crash um super super early this is preliminary etc but from you know so the various you know sort of exac accounts and YouTubers that I follow the like you know sort of consensus theory right now there's basically like two theories one you have a like dual engine failure right on takeoff which is like a 1 in 10 trillion outcome the idea that somehow both fail at the exact same time like even if it's like a flock of birds it have to be a massive flock of birds that somehow hits both but the thing that you know quickly is, you know, lining up is like a leading theory is that the flaps weren't down on takeoff.

So with a plane like this, in order to have an additional lift basically when you're first taking off, you need to have basically flaps down. If you basically just have the, you know, sort of flaps up and wings typical, you just aren't going to be able to generate enough lift.

And if you look at the flight profile and how it's sucked down and that it was fully loaded with the passengers and with fuel, the flight profile basically ma matches that. If you're going down the runway at takeoff speed with flaps up, there are alarms.

There are things that like are clearly alerting you that you have missed a step in a checklist which is like I guarantee whenever they pull that either cockpit audio there's maybe some chance that somehow all the alarm systems somehow you sort of failed but like to me what's very clear is like in the takeoff checklist they did not put you know sort of flaps down and they clearly ignored that information um you know on the uh you know takeoff I would guarantee that there's going to be some you co-pilot pilot you know miscommunication you challenge that's going to be a huge contributor to the crash and so while the Nathan Fielder like season 2 was incredibly funny some of like the best TV that I've ever seen as a pilot.

I actually genuinely think he has like a really strong perspective here where when you're training about how to communicate with, you know, somebody else in the cockpit, all they show you is like a like, you know, basically like PowerPoint slide that says make sure to speak up.

The idea that like, you know, just being reading the words make sure to speak up somehow like trains you to actually speak up live in the moment.

That'd be the idea of and I think you know Kou and you and I have talked about this like I could read a presentation that says when you're on TV don't say you know you know you know which is like the verbal tech and I could read that and there's no [ __ ] way that I would actually be able to do that correctly.

I need to go on TVPN every practice. There's no practice. And so the whole concept behind Nathan Fielder's you know season 2 is actually make the pilots practice. Put them in these situations that are like simulations of crashes.

force them to communicate to one another and then force them to play out an alternative scenario where you actually push back against the captain and get into that an uncomfortable place where yes, this person's more senior, yes, this person's more experienced, but you have to be willing to actually push that.

And if you never practice that and that's one of the leading causes of crashes in the United States of the past 20 to 30 years, how is that something that the FAA is just like completely ignoring?

And so I have half a mind of, you know, when I'm in DC for like VA founders fund, Hillen Valley, I literally like I think next year for Hillen Valley, I want to put Nathan Fielder and the NTSB guy as one of the primary panels and be like this is a like extremely important topic coming from the technology industry.

Somebody needs to go influence this. So anyways, Nathan Fielder is a [ __ ] goat and you know, somebody's got to, you know, make role playing a part of pilot training. Is there is there a metaphor for the uh for what happens in venture?

maybe an associate who doesn't speak up when a GP is ripping a check into a you know a company that maybe doesn't have product market fit and is trading at 500x revenue role playing role playing in the partner meetings okay so let's just take a step back let's let's stop talking this specific investment let's just let's role play a scenario where a diff we were investing in a different company at 500x you know ARR and and you know maybe a junior partner had some concerns about the invest Let's just play that out.

Sidebar, five minutes, then we'll come back to this. You know, do you think that ever happens in Silicon Valley? I mean, look, you know, when uh a founder at an investment committee meeting is playing League of Legends and, you know, the numbers look a little too good. I don't know.

I think maybe like there should be a little role playing there and somebody should speak up and say, "Hey, like don't we want founders that are pretty focused on what they're doing or at least good at League of Legends? At least good at League of Legends.

Should somebody investigate the fact that he's [ __ ] bronze and [ __ ] up?

" like somebody should look at that and I think with a little bit of role playing I could imagine scenarios like that that could train somebody and so unfortunately what happens is no role playing happens and then the primary sponsor of the deal is uh you know she's shifted to somebody else and the person's quietly exited and so our solution yes it does happen I'm just referencing a hypothetical by the way just hypothetical but it speaks to the importance of training yeah training is very important junior people need to speak up and say their opinion for sure uh what about the startup angle here.

I uh there was some company that came to one of uh one of our off sites at FF. Um and it was basically like autopilot for planes. He was working with the AC130, the C130. And the idea was that uh non-deterministic probabilistic software could not be integrated into the actual control stack of planes.

And so his workaround was put an iPad there that reads the data and makes recommendations. And I was like, that maybe is bad, but also maybe I don't trust autopilot on the plane right now anyway.

So, um, this feels like, you know, whenever there's this moment, you'll probably see different companies jumping in to try and help for, you know, mercenary or missionary reasons.

Um, is there a world where the where the private sector can actually help with airplane safety or aircraft safety or is this purely in your mind like we need just changes to training like laws? Um there definitely is.

I think a lot of what you know sort of pilots do absolutely could be subtly automated you know by software over time.

I do think it's like um obviously a significantly harder problem than just like you know you know chat GPT reading the internet and then user spewing out words like it's just like the the training data is a lot harder to acquire the like outputs obviously like you can't have any hallucination and so um but like you know in 10 15 20 years could this be like fully fully automated 100%.

I think the challenge is much more cultural where it's just like the FAA other than like the DCA crash like you know basically had a spotless you know track record for like 15 plus years and so it's like really hard to justify.

I this is where it's I think like some of these like next generation like you know supersonic and hypersonic planes are going to really struggle where it's just like we're sort of stuck with this like you know you know sort of tube and wing you know sort of plane design even though there's like plenty of other net new plane designs make sense but they're just so safe and so then you're stuck in this like you know catch you like the analogy I always give is like imagine today you came to society in a world where there was no cars and you said I'm going to give everybody this like two-tonon vehicle they're going to like move 60 miles per hour while there's like pedestrians close by and by the way we're just going to kill like 70,000 Americans every single year.

Just every year we're going to [ __ ] murder 70,000 Americans. But it offers all of us a little bit convenience. We like get to places faster. Yeah. In a world with no cars, no [ __ ] way. Would we ever pass that? Like that would never happen, right?

And so there is a little bit of this like, you know, societal moral question of like what is the right number of deaths in order for progress? And right now in the world of aviation, the answer is zero. Zero deaths is allowed. Yeah.

Uh I is is there a tinfoil hat scenario where like aviation's been very very safe and then one crash happens like the first crash and and and it looks just like a blip and we return to kind of like the averages and we say okay well even including that it's still very safe but now there's been the clash the crash in Canada the crash in India the crash in DC like at a certain point there's so many like you have to wonder if there's something bigger going on and and if it was this safety training thing we we haven't been doing this type of training ever So like is it possible that there's something else going on?

I think when these complex systems deteriorate like they can deteriorate very rapidly and it's like hard to totally diagnose like you know an individual cause but like I do think that's what you're seeing a bit.

It's like Boeing culture and like their optimization for shareholder returns is like you know killing supply chain and consistency. You're seeing some of the like you know sort of craziness of co where we retired a bunch of pilots and now we're hiring a bunch of super junior pilots.

the fact that ATC is super overworked and like you know they're trying to like still operate off of like 1970s you know sort of software even though the world's getting you know more complex and you know sort of higher volume in traffic and so I think what you're seeing is just like a fraying at the edge of this system and yeah I do think there's a world where like all of a sudden you see us return to like you know 1990s level plane crashes on a super regular you know sort of basis and that may be the thing that finally makes it so that the like you know FAA and others need to start to adopt you know sort of next technologies where like they can no longer just totally rely on the safety track record.

I think there's something you have to imagine this this uh this idea this thought running through a pilot's head in a scenario where one pilot is concerned about actions that the the crew is taking collectively and then they're just basically running the the the thought process of planes are just so safe.

I've done thousands of these. It's going to be fine. We've, you know, I've been in situations that weren't great before and it always worked out and then be and then they're just like thinking about the actual statistics and realizing that no, you're not a statistic. You're in the cockpit.

You're a person in the cockpit responsible for hundreds of people's lives and I've been invested at 200x ARR and been able to spack before it can get out. So, so how bad is this for for Boeing? I mean 787 total failure. I mean I mean I don't think it's anywhere near like the 737 you sort of max issues.

I mean yeah they'll have to like go through an investigation. The like orders on this stuff are so you know sort of you know you know done so far in advance the likelihood that it like shifts revenue. It's like unless something catastrophic comes out of the investigation in like a year.

I actually don't think that'll have that much effect especially when you like look at the footage where it looks like it's a pilot error. for reference on public. com. The stock is down 5%. Or something like that. Um, should we go on to the next topic? Yeah. Uh, TS up for Shink. What's going on there? Shink.

Um, never thought I'd be, you know, sort of a fish investor, but I do love my fish. Um, I've been a long time, you know, sort of lover of, uh, the Tokyo fish market, going there, eating quality fish.

I never appreciated that a part of what makes the Tokyo fish market quality, yes, is the fish that they go catch, but it is specifically also the way the fish is prepared and basically how it's killed.

Traditionally, fish in like the United States and most of the world, you basically catch a fish and then you either like freeze it to death or you like let it flop and basically like, you know, um, you know, effectively drown in air.

The problem with both of those techniques, fish knows that it's dying, gets super stressed out, releases all these like hormones, lactic acid, etc. That significantly degrades the quality of the fish. In Japan, what they do is they basically like stab out the brain and pull out the brain stem.

So, the fish basically never knows that it's ever dying, even though, you know, you obviously are killing it. And so you have a very basic like relaxed fish meat and you bleed it out very quickly. That's a lot of what leads like omocasi level quality. That typically is like a super manual.

You need like an old Japanese chef on your boat in order to do that. Super expensive.

basically met this guy save from Shink that you guys will be talking to that like one comes from like a fisherman's family which is like you know incredible like founder market fit and two basically built these robots that automatically basically kill the fish in the way that the Japanese chefs do but rather than like selling the robots they actually just have like a consumer fish brand that is now carried in a bunch of like you know high-end Michelin star restaurants starting to get into you know sort of retail and I think it's like one of these things where it's like I just love you you you rarely meet just such an example of like this is the company that this person was like born born to build you know what I mean like you know I like you know maybe stumbled into inspace manufacturing when I was like in seventh or eighth grade but I was not born for it safe was literally like born to build this company and I think there's like a perfect parallel of like if you look at like Cargill like a 100 years ago they started off with like some simple you know equipment and then steadily basically grew into a massive like processor distributor wholesaler they're like you know the largest like pig and cow um you know sort of distributor um I think there's an opportunity here to do this in fish where like there hasn't really been a like next generation tech enabled equipment provider in the fishing industry.

It's super longtail. There's no like one conglomerate.

I actually think that this like initial angle into like high quality fish is something that like safe actually has the ambitions to go build into this like massive conglomerate that yes is serving fish but also is the wholesaler and the distributor and they have like automated fishing boats and they have like you know they're like automatic killing and the processing and like the bleeding all of that nobody's taking a look at in basically like 100 years and I'm not too worried about like other people competing because I don't think like two Stanford grads want to go get their hands dirty and deal with a bunch of fish guts you know like really I've been I've Uh, I've spent time at the facility in the Gundo.

It smells like fish. Smells like fish. You every day you go into work and for 8 hours you just smell fish every day. And every day out, you know, we just believe, you know, competition is for losers. What better way to smoke out the competition than just be like, you have to smell fish for 8 hours. Yeah.

It's a great It's a great market. You know, the market opportunity is is is massive. Uh, low competition. It's a great market to be in, but you got to love the stench of fish.

I felt bad last time last time I went and saw Safe in the team, he gave me some fresh fish and then classic LA, I had to like drive to a couple other meetings and then I realized that like I got home and like the fish had just been sitting in my hot car for like the entire day so I couldn't eat it. But I've had it.

I've had it. I mean they they're they're distributing to Nou and a bunch of amazing restaurants. Amazing. Well, I'm excited to talk to him. Thanks so much for Delian. They were so mad that you were back on the show. They took down AWS too. US town, too. Google us. Maybe not even we might not even be live.

We might not even be live anymore. We're just the three of us talking. But okay, we're live apparently. But this could be Iran. It could be Russia. It could be China. Could be anybody. They don't want us to India. I mean, Pakistan's getting in on the news. It is crazy times. Stay safe.

Enjoy the rest of your trip to San Francisco. We'll see you next time you're in LA. Bye everybody. See you soon. Bye. Uh and next up, we're going right into Shina. We got Safe who I met. He's got some big news today. And yeah, Safe YC company went through YC a few years ago at this point.

Did he raise on a safe hard tech company NYC? Yes. He