No. Shipbuilding isn't an area for startups and unions are a very very small hurdle compared to getting the necessary materials. Just think of the steel. Anyone can order a few tons of rebar, but ask for 10,000 tons of plate steel suitable for ships and you will be laughed at. That isn't sitting in a warehouse or a back lot waiting for customers. An order out of the blue might take years to fill. The people that make steel work with regular customers and will always put those regular paying customers above a random startup who might not ever pay. If anything, they will want insane down payments.
Then think ship engines, the ones with blocks bigger than most houses. Order one of those out of the blue and, again not joking, it might be 10+ years. Those engines are targeting at future ships still on drawing boards, each is accounted for long before any metal is assembled. Any startup's order would be at the back of a very slow line.
A startup wanting to get into ship construction would have more success launching a new social media system, getting lucky, IPO, then use that money to purchase an existing shipbuilder. But at that point the "startup" is really just another a hedge fund investing in ship construction.
Totally different scales. A falcon 9 rocket weights in at 550 tons, and most of that is fuel and the rest can be broken down into small parts. The average cargo ship might be 50,000 tonnes, literally orders of magnitude larger, and none of the major parts are road-mobile. The logistics are just totally different. Space X also competes in a rocket industry that is maybe 60 years old. Shipping is a 1000+ year-old industry making it, again, a couple orders of magnitude more mature. One might even argue that commercial cargo ships have existed for two, three or even four thousand years. Pharaoh paid someone to ship those blocks down river for the pyramids.
Then do the obvious thing - don't start with 50kt cargo ships. There was a funny article the other day about how some group - possibly the Columbians? I forget - was getting into sub manufacturing because they needed to transport drugs. That isn't military grade nuclear subs but it is on the path to them if for some weird reason they keep chugging away.
No company in the world starts with massive state of the art. It is a recipe for failure, the engineers need to learn in a lower-risk environment.
Building crab fishing boats doesn't have an evolution that leads to building nuclear submarines anymore than manufacturing charm bracelets or ready-to-eat delivery meals. The requisite materials, logistics, and labor expertise are effectively unrelated to one another and separated by mountains in terms of capital and material resources.
As was said before, the only way in is basically to buy into an existing operation, at which point where you got the money from originally is irrelevant.
> Building crab fishing boats doesn't have an evolution that leads to building nuclear submarines
You're putting words in my mouth because that is probably not how people would start out.
But the literal incorrectness of that statement is so wrong it is puzzling. The shipbuilding industry, which builds nuclear submarines, started with some extremely modest boats. The path we followed to get to where we are was precisely starting with little fishing boats and evolving to where we are now. There is obviously a path from small boats to large boats. To suggest otherwise is absurd. As is the idea that small companies can't become big companies or move to make more sophisticated products.
You might be about to make the argument that government funding is necessary or something, but the big problem here is unusually simple - the US is not globally competitive at manufacturing, a startup would be expected to fail if left to compete in the market and that is why nobody is going to try (or succeed). They've already been beaten out of the market.
> You're putting words in my mouth because that is probably not how people would start out.
I'm making concrete your vague handwaving, feel free to provide a more complete explanation of how you think this goes.
> But the literal incorrectness of that statement is so wrong it is puzzling. The shipbuilding industry, which builds nuclear submarines, started with some extremely modest boats. The path we followed to get to where we are was precisely starting with little fishing boats and evolving to where we are now. There is obviously a path from small boats to large boats. To suggest otherwise is absurd.
No it didn't. And certainly not the military shipyards discussed in this article. While a couple (Newport News, Brooklyn Navy Yard) have origins as more modest shipbuilding sites, the vast majority were purpose-built sites for military construction built with US government dollars. Like the Manhattan Project or the Hoover Dam, there wasn't a smaller initiative that gradually developed into a massive industry. Massive industry was the go word.
These types of projects, continental railroads, interstate highways systems, and military naval shipyards, they can have private-public partnerships but they do not grow organically from more modest roots without public dollars.
> You might be about to make the argument that government funding is necessary or something, but the big problem here is unusually simple - the US is not globally competitive at manufacturing, a startup would be expected to fail if left to compete in the market and that is why nobody is going to try (or succeed).
This is irrelevant, pick your foreign military shipyard and you will find that the story is the same. Bohai, Zelenodolsk, Cherbourg, any of them. The government recognized a need for a new, larger, more powerful class of ships and either built or expanded the logistics necessary to construct the vessels.
> I'm making concrete your vague handwaving, feel free to provide a more concrete explanation of how you think this goes.
Imagining counterfactuals is fundamentally vague. Someone would need see an opportunity to move in to shipbuilding at the low end and take it. But that is the low end, ie, probably not starting at 50kt tankers. You're imagining scenarios where it doesn't work. That is really easy, most business ideas don't work. But successful companies come from the unusual situations where ideas do work, so the fact that you are good at imaging how things won't work isn't really material. You're one of the approximately 8 billion humans that isn't capable of founding a shipping startup. I am too. But nonetheless we can be confident that the best positioned people in the US could do it.
The issue is not the path, the issue is that US companies can't compete. The existing, established, US shipbuilders got whipped by the Asians. The question here isn't whether startups work, the issue is the market that they'd inherit if they got big doesn't support US shipbuilders.
> No it didn't.
You are, and I'm speaking quite literally here, absurdly wrong. Modern ships didn't pop into existence from nothing. Nothing does. The whole modern economy is just iterations of smaller simpler designs starting with wood fires and spears. It is just a basic fact that there is a path from any industrial activity to any other.
And companies tread it fairly regularly. It isn't unusual for people to move from one vertical into another.
> Like the Manhattan Project or the Hoover Dam, there wasn't a smaller initiative that gradually developed into a massive industry.
The Manhattan project is within startup budget. OpenAI is talking trillions to develop AI, that is like 50 Manhatten projects. Nuclear science turns out not to be that hard once people know that splitting atoms is a very energy-plush operation.
And the dam example is silly. Are you suggesting the Hoover Dam was the first dam? No, the US had a huge dam building industry. It was a big project, but it was done by existing companies.
> Modern ships didn't pop into existence from nothing. Nothing does.
They didn't evolve from some minor capital investment. The problem is not one of ideas it is one of scale. In that sense, yes, they popped into existence from nothing.
The Panama canal is the result of centuries of technological development in the technology of canal construction generally, but fundamentally it is not the result of a startup or even a hundred startups building tiny canals until they had enough capital and expertise doing so to build a big canal.
It is a result of governments (plural, the first attempt failed), in a one-off immense expenditure of capital and manpower unable to be mustered by the entirety of the global shipping industry (for everyone with eyes and a map could see the potential), constructing something that was not there before. "From nothing".
What will the startup do that's so much better than the existing shipbuilders? China has much cheaper labor, and produces a huge amount of steel domestically. They build big ships and smaller ships. What does your hypothetical startup do that will let it compete against existing incumbents that have established incumbents with much more experience?
With drug trafficking, it need for new manufacturing is because they had an illegal use case. And those subs usually don't last more than one journey. It's not really comparable to the shipping industry at large.
Who were the existing providers in the space/rocket industry? NASA and their foreign counterparts?
Shipbuilding is much larger. Shipping is a huge industry with many parts. It is not necessarily impossible for a new company to come in to any single area, but they wouldn't be able to break into something like military ship building.
A possible avenue for 'disruptive startup' is the USV area, see Anduril. But even then that required immense amounts of capital because you are competing against the big dogs (Raytheon, Northrop Lockheed, Boeing, BAE etc.)
The existing providers were NASA and United Launch Alliance, the latter of which is a joint venture between Boeing and Lockheed. They merged their space launch businesses in 2006, because neither could individually sustain their launch services given 1) the expense of their systems, and 2) the low number of launches in the late-90s/early-00s. They were the sole provider to the USG from 2006-2015, only because SpaceX won its lawsuit against the Department of the Air Force in 2015, which had been blocking SpaceX from competing for DOD/NRO launches.
Maybe a startup can focus on making maritime drones, just like the Ukrainians do. I bet there's probably a lot of demand nowadays for that kind of thing. I checked and Anduril appears to be making something they call AUVs (autonomous underwater vehicles).
Workers are unionized in a lot of overseas industries that are more productive than the US. Don't know about Korea but they are in Japan and certainly in Europe. Part of it is have better union structures than we do; sectoral bargaining means any employers and unions spend less time fighting each other.
A big problem for US shipbuilding is the Jones Act, which is so protectionist it's easier for the industry to flee the country entirely than deal with it. It's really harmful to our overseas areas like Hawaii and PR too.
I am from South Korea. Hyundai Heavy Industries' Labor Union, the union of the largest and probably the most productive shipyard in the world in Ulsan, is likely one of the strongest union in the world.
https://www.youtube.com/watch?v=-uXqkbpy-6s is an official music video of The Song of Hyundai Heavy Industries' Labor Union. Lyrics is like "Comrades, remember, the spring of '89, we will fight with our lives" etc, referring to one of the most legendary event in South Korean labor history, 128 days strike at HHI of 1989.
The union is a sectoral union, a chapter under Korean Metal Worker's Union, but it is so large it has a special carveout solely reserved for it. Korean Metal Worker's Union otherwise forbids company-specific union.
Since this was well received, here is a full translation of The Song of Hyundai Heavy Industries' Labor Union, done by ChatGPT and reviewed by me:
The Song of Hyundai Heavy Industries Union
Brothers of Hyundai, standing tall,
Piercing violence and suppression's call!
Forward! Forward! Let us rush on through!
Towards the liberation of workers, true!
Even if the power's violence is widespread,
We'll fight with our lives, ahead we'll tread!
Liberation! For liberation we fight!
Struggle! For struggle we unite!
At the vanguard of the workers' liberation call,
Hyundai Heavy Industries Union, standing tall!
The Sun is rising over the Mipo bay,
The hot tears of our mothers, they say
Forward! Forward! Let us rush on through!
Towards the liberation of people, true!
Comrades! Don't you remember the spring of '89?
We'll fight with our lives, our destinies entwine!
Liberation! For liberation we fight!
Struggle! For struggle we unite!
At the vanguard of the people's liberation call,
Hyundai Heavy Industries Union, above all!
My point was that you will not solve this with a startup, because of how a startup must operate due to its inherent nature (ruthlessly capital efficient, constrained runway). You must operate this as a long term, sustainable operation (perhaps a public private partnership), like you would build a nuclear reactor over years or a decade. To not do this is to ignore the muscle memory needed to retain the core component: teams of skilled labor with options. It’s a flywheel you bring up to speed with capital, a pipeline/book of work far into the future, and domain experienced management.
Boeing is learning this the hard way currently, for example, and is in talks to buy the subcontractor (Spirit AeroSystems) it spun out: https://news.ycombinator.com/item?id=39571819
> A big problem for US shipbuilding is the Jones Act, which is so protectionist it's easier for the industry to flee the country entirely than deal with it. It's really harmful to our overseas areas like Hawaii and PR too.
In theory, protectionism should help the industry. The government has opened a niche for a US shipper and a US shipbuilder (or maybe a vertically integrated shipper and builder) to get mainland goods to Hawaii, PR and Alaska.
Doesn't seem like a wide enough niche to get much traction though. It's easier to simply not serve that shipping market.
And shipyards are hard to start. They've almost certainly got to be in environmentally sensitive areas, that are often pretty expensive real estate, and then you've got all the machinery and what not. Hard to get that done these days.
They'll tolerate it if appropriately compensated, e.g. if you structured the venture as a workers' cooperative where they share the upside as well as the downside.
Unions in the United States are entrenched with a particular government-backed structure that puts the high-ups far from the working guys... they're not all necessarily leftist or willing to accept models other than management-and-bargaining.
A lot of people believe the sanctioned union structure was a compromise with employers who did not want fully functional collective bargaining counterparties but who knew they couldn't eliminate unions entirely, and they could be right as far as I know.
An argument in favor of this viewpoint is the treatment of sectoral/sympathy strikes (and other important solidarity tactics like sectoral boycotts) under the Taft-Hartley act: https://en.m.wikipedia.org/wiki/Taft%E2%80%93Hartley_Act
People think of the NLRA as a legacy of the New Deal and therefore inherently pro-labor, but conveniently forget that the deal was changed by a bipartisan alliance of corporatists over Truman's veto.
Sectoral/sympathy strikes were historically the backbone of union power in the US and remain so today in jurisdictions with strong unions. Legally protecting ineffectual strikes but not effective strikes traps labor in a local maximum, where there's just enough to lose to disincentivize going for more substantial gains.
I wonder if instead of disrupting U.S. shipbuilding, it would be more cost effective to figure out containerization of weaponry. Figure out how to land a drone in a shipping container and you might be able to turn any standard civilian container ship into an aircraft carrier with 1000 drones, 10x the air wing of a modern fleet carrier. You also trivialize logistics for this - the armaments, fuel, spare drones, sensors, etc. also go in containers, and resupply could be done at any container port and carried by any other container ship.
There's a long history of this in both world wars, eg. auxiliary cruisers and escort carriers, and they were fairly effective. See eg. the Battle of Samar, where 18 escort carriers and their destroyer escorts took on most of the Japanese battle fleet and won.
Similar ideas have been tried in the past, they always end the same way. The converted civilian ship has zero survivability and is sunk by the first passing warship to the embarrassment of all the planners involved.
Today such vessels are so susceptible to anti-ship missiles and submarines that they're laughed out of any serious military planning meeting where the participants aren't already desperate for tonnage.
Right, but I wonder if developments in the rest of weaponry have changed the calculus around that. It's pretty likely that a fleet carrier or other purpose-built warship also has zero survivability in the face of a barrage of anti-ship cruise missiles too. That's what recent wargames have shown, and also has been born out by Russian experience in Ukraine. The strategy for survivability is "don't let hostiles get within your air defense bubble, and shoot as many missiles down with your CIWS as you can if they do."
An escort drone carrier could follow that strategy as well, and much more cheaply. The point is to make them expendable and cheap. Yes, if they get hit they're gone - that was always the case in WW2 as well - but there are now so many more options for blowing everything else out of the water that might get close.
> Right, but I wonder if developments in the rest of weaponry have changed the calculus around that. It's pretty likely that a fleet carrier or other purpose-built warship also has zero survivability in the face of a barrage of anti-ship cruise missiles too
This is incorrect.
> been born out by Russian experience in Ukraine
Defending your assumption with the Russian fleet is certainly a thing you are doing right now
The solution is similar to WW2. Build disposable ships faster than the enemy can sink them. This would work against most enemies except China can build them even faster
There's no such thing as a blue water autonomous torpedo, there will always need to be a human-occupied platform to project power and manage military naval operations from.
This is without getting into the obvious problems of "lol how do you plan to control your USV? Pacific ocean spanning tether?"
As long as there are blue water navies, there will be military shipyards servicing them.
Looks like Boeing is finally delivering an "Orca" prototype XtraLargeUUV (85 t, 26 m) - more sub than torpedo. Seems to be pier deployment with 6k nm diesel/electric range.[1]
Already a thing (mostly remote controlled drone ships). But you'll always need mitigating controls for unexpected or low frequency events, and it's not always feasible to fly in a crew at which point the crew needs to be on board. This effectively is the same as the crew that's already there right now, so it doesn't actually impact all that much.
The biggest benefit of remote, drone or automated piloting is the way you could make better use of a person's time, and maybe save a tiny amount of wages, but it's mostly the resulting functionality/features/business processes that is the benefit, not the headcount cost.
Container ships already have tiny crews relative to the amount of goods they transport. The systems are already highly automated and you need some level of crew anyway for unexpected events, maintenance, etc. The idea that "automation" is this magical thing when the labor component is already pretty small is just silly.
In spite of the intro, those seem to pretty much be all people-mover/metro systems. They're not long distance transport in potentially unpredictable conditions.
And none of them are without support crew; The crew just doesn't need to ride on the train itself, but rather waits in stations or in depots along the tracks.
I’m not sure why container ships couldn’t be piloted remote if drones are. It’s just that it is cheap to hire a bunch of sailors from the Philippines so automation isn’t really necessary.
Right. You don't always need to have a driver potentially doing nothing 99% of the time. But you probably have remote and station staff a lot of the time.
Fully autonomous ships are pointless. But automation improvements across the US fleet while retaining reliability are a significant factor behind America's dominance of the seas.
Activity-hours aren't fungible and have constraints. If I need to eat by 6pm and get home from work at 5:30 pm, I generally can't just choose to swap a 9 - 10 pm hour of streaming with an hour of preparing food.
While some people can level up and perform food-prep activities during the 9-10 pm hour so that the following 5:30 - 6 pm slot can be filled with quick prep, many people don't know how to do so, nor see the ROI for such activity.
I am not saying you have to do fresh pasta every day (nor should you eat pasta every day!). I like to do it, same as bread and other things, but I obviously don't always do everything.
But parts of the pleasure of many things comes from both the anticipation and the satisfaction of making something good and taking the time to cook is part of it. Which sometimes involves organizing your day so that you can do it.
My kids have a hyphenated last name; my wife's last name starts with A, mine starts with L, and the kids' last name is A-L to give them "alphabetical privilege". Also, she carried the kids.
Wow it’s crazy how he didn’t add a disclaimer to his original post.
EDIT: Did he really change his mind? Still seems like he thinks that book is pseudoscience, just that sleep is more important than he originally gave it credit for.
A formula 1 race car is _much_ faster than a freight truck. However a freight truck can deliver a lot more goods across the country than an F1 car. Which is your big organization optimizing for?
My hunch is that this is exactly what he was expecting. There is a lot of hype around ChatGPT passing the medical exam and this exercise is a counter point to that.
That's not true, ChatGPT is a model. Quote from the ChatGPT announcement post:
> ChatGPT is a sibling model to InstructGPT, which is trained to follow an instruction in a prompt and provide a detailed response.
The medical exam has very specific questions which define what you're expected to include in the answer. The question asked in this case was nowhere near that detailed, do I don't think they're comparable. To really evaluate something beyond the "random generic user" level, you need to be familiar with the tech as well.
The article really tells us more about the experience of someone with no chatgpt knowledge checking their own symptoms rather than its usability for emergency diagnosis.
any example that doesn't use the current sota isn't a very good counter point to be honest. 3 barely passed. 4 aced it. For all we know, GPT-4 erases most of his concerns (not saying it would).
Not an FDIC auction, but if you’re interested in what this sort of bidding looks like the book Barbarians at the Gate details the LBO auction of RJR Nabisco. That auction was a lot of bankers squired away in conference rooms on separate floors of an office building while the auctioneers walked bids between the various groups.
