> The Alaska Highway. Starting in 1942, 1,700 miles of highway were built over the course of 234 days, connecting eastern British Columbia with Fairbanks, Alaska.
I grew up in Dawson Creek. Our claim to fame: "Mile Zero of the Alaska Highway starts here!"
The Alaska Highway as built in 1942 is nothing like a highway that most people would envision. It wasn't paved, for example. It was good enough for military vehicles with crews of soldiers who could make ad hoc repairs to the road as needed while they transited through, who had extensive survival skills, and could literally radio for assistance if needed.
It wasn't opened to the public until 1948 -- so perhaps 6 years, not 6 months, is a better estimate of its time.
It's also shrunk by almost 20% as it has been continually rebuilt to make it passable by passenger vehicles.
It's an accomplishment, no doubt, but I feel like omitting these significant caveats is meaningful.
Hm, thank you. I checked a few sources when adding this one but will go back and double-check the specifics/qualifiers.
Wikipedia, incidentally, seems to support the shorter timeline:
“The official start of construction took place on March 8, 1942, after hundreds of pieces of construction equipment were moved on priority trains by the Northern Alberta Railways to the northeastern part of British Columbia near Mile 0 at Dawson Creek. Construction accelerated through the spring as the winter weather faded away and crews were able to work from both the northern and southern ends; they were spurred on after reports of the Japanese invasion of Kiska Island and Attu Island in the Aleutians. During construction the road was nicknamed the "oil can highway" by the work crews due to the large number of discarded oil cans and fuel drums that marked the road's progress.[9] On September 24, 1942, crews from both directions met at Mile 588 at what became named Contact Creek,[10] at the British Columbia-Yukon border at the 60th parallel; the entire route was completed October 28, 1942, with the northern linkup at Mile 1202, Beaver Creek, and the highway was dedicated on November 20, 1942, at Soldier's Summit.”
The road existed after 9 months, but it was unpaved and only open to military convoys, as it often needed repairs and was single track in places. This was fine for its purpose (a land route to get materiel to Alaska in case the Japanese attacked), but isn't what most people think of when they think of a highway.
It's a huge achievement - something like 10,000 soldiers were involved in its construction.
But I worry that when most people think highway they think of the interstate system - straight lines, gentle grades, paved, multiple lanes.
This is more akin to the army setting up camp and building accomodations for thousands - we wouldn't say they built 1000 houses, they put up tents. Still an impressive logistical feat, but... different.
Yes, that the name may be misleading is a fair point. Even though it was (I think) called the Alaska Highway from the beginning, I changed the text to say "military roadway" -- hopefully this helps to clarify. Thanks for pointing this out!
You make a fair point. However, what was built would have been considered a highway at the time, so while the distinction is worth keeping in mind, it is more an evolution in the meaning of the word highway.
No, it's specifically meant to illustrate that the inflation in infrastructure costs has been wildly faster than the inflation in the rest of the economy. The US can't build infrastructure for any kind of reasonable cost anymore and it is a big fucking problem.
Most of the route is through extremely rugged mountainous terrain, not tundra, and developing a roadbed that can survive -70 F winters is a lot more involved than "piling dirt into a shape"
I traveled it before was paved, and it was, and is, a true engineering marvel.
Yes, I both unravelled my vast illiteracy on road engineering in general and understated a bit on purpose. My point was that there is not much existing infrastructure to take into account up there. Your point taken though.
Do you have examples? I feel like things have gotten 10x more expensive and they're built with half the quality. In San Francisco we have a leaning skyscraper, the Bay Bridge which some people are still unsure is safe despite being several years late and costing 5x as much as predicted, and the Transit Center which was however many months/years late and is continuously shut down for safety reasons.
I realize San Francisco is a microcosm, but in my experience the focus has been on extracting money via corruption and building the bare minimum you can without getting sued.
You could argue the money is going towards greater safety for workers, but I see no plausible argument that it's resulting in higher quality work.
“We chopped down enough trees and leveled enough ground that you drive supply trucks through...sort of.”
I rode it on a motorcycle in 2015. The days of a “I survived the AlCan” being a meaningful t-shirt are well behind us. It’s a little rough, mostly paved, but not what they built in 243 days.
If you extend it to the whole Pan-American Highway, the timeline gets stretched to 3 or 4 decades from first proposals through execution, but the scope is pretty amazing. https://en.wikipedia.org/wiki/Pan-American_Highway
I've always found it kind of surprising that the Darien Gap has never been closed. I can only assume that if Panama and Colombia were rich countries rather than poor countries that it would have happened. People solved the engineering challenges required to build the Channel Tunnel and the Øresund Bridge, after all.
(Note that I'm not saying that it should happen or that it would be a good thing necessarily.)
It’s not just money, it’s a matter of government not having very good control over those northern areas of Columbia. Even if they built the road, it wouldn’t be safe to travel due to the density of smugglers and cartel activity in general.
Building (not to mention maintaining) a long road through a sparsely inhabited tropical marsh/swamp is pretty expensive and complicated, especially if you want to minimize environmental immpact.
I think this example along with what I just randomly know about the other examples, gives the impression that this list is essentially meaningless.
I recall the "Marin Ships" being indeed constructed quickly but of very poor quality and in no way equivalent to constructing real navel ships quickly, just "we can put X amount of metal into the water and hope it doesn't sink too quickly".
And so forth, for significant number of these examples. Some of the examples, that I know nothing of, might be more "real" or more significant examples time-efficiency, but the list doesn't seem a credit to them. Lots of these are "yeah, we can put a stake in the ground quickly" and that's it.
I had to look this up. Marinships were the west coast Liberty Ships, built in Sausalito. Later they build fleet tankers & oilers there, probably to the same "get it done now" quality standard.
My dad went to the European theater on a Liberty. He had the top bunk in the hold and spent the voyage looking at the truly awful welding that had been done in the shipyard. But they didn't have to last longer than a few years, so lots of corners were cut.
Sounds a lot like the Dalton highway , we had to rent a special vehicle which had a radio to do 2 way communication with the traffic and the highway wasn't paved.
A CB radio is not necessary on the Dalton highway, though vehicles that transit it frequently are usually equiped with one to communicate with the semis that make up much of the traffic (especially to coordinate passing on dusty sections). The Dalton has several unpaved sections but is fairly highly maintained and certainly does not require vehicles transiting to perform road work themselves. You certainly do not need survival skills (at least during the summer) to drive it as there is a consistent amount of semi traffic that can assist you if something goes wrong as well as periodic pump stations for the pipline. If anything, the Dalton would be more akin to the Alaska highway some time after it was opened to the public.
I drive the freeways just to go a few miles most days, and tens of thousands in a few years. Carbon be damned, I can't thank highways enough for the life they enabled. When teleportation comes along I'll firmly recant and hope this comment has disappeared, but until then I'm in debt to the national arteries I benefit from today.
I can't punish freeways enough for the life they've enabled. The closest thing you will get to teleportation in your lifetime is moving to a city. Maybe you cannot get from one side of the world to another in a minute, but you certainly can walk down the street to get anything you need in a minute or two in good major cities.
You're right and I wish I could afford it. Real Estate prices are validation that cities are a better way to live. I bought a car 8 months ago and I've only driven 1500 miles but I'd love to drive less. We have inherited a legacy of feeble foresight and racism, built into our cities and roads, but we're going to build something better in our time.
Most of the emperical studies I've seen suggest people who travel little by car generally do less trips vs. many shorter trips; i.e. they're not driving 6 miles a day, they're driving 20 miles twice a week.
Same here. I used to live in Jacksonville Florida and routinely used the I-95 to get around the city. What's funny is I'd always envision "warping" when getting on the highway, because inner city traffic speed is ~45 mph, and then it'd instantly jump to 80+ mph on the highway.
In Nevada and I believe other states before the 55 MPH speed limit was instituted, "reasonable and proper" was interpreted pretty broadly and might have exceeded 80 under good conditions.
There's no question in my mind that North America has gotten a lot worse at completing great civic projects.
Sometimes, that's for good reasons. We don't have slave labor, we have environmental review, and we allow affected communities to protest and block approval. In the old days when the city fathers had a bright idea they'd just plow through the neighborhood of $LEAST_POWERFUL_ETHNIC_GROUP.
We might have overcorrected! But I also don't know what model there might exist for moving quickly with a multi-stakeholder process.
Time and time again, organizations prove that product design/project management by consensus leads to budget overruns and under delivery.
Quality at speed requires dictatorial like organisation with high risk to all. In every example given people died, more were maimed and even more had mental breakdowns. Seems like most people aren't willing to be a part of something like that today and there doesn't seem to be any popular cause that would motivate such projects.
> In every example given people died, more were maimed and even more had mental breakdowns. Seems like most people aren't willing to be a part of something like that today and there doesn't seem to be any popular cause that would motivate such projects.
People are still totally willing to be part of something like that; people are as haphazard and risky as they ever were. The government won't let them because it doesn't trust people to assess the risks properly (there is evidence that the government is right; people are very bad at assessing risk).
However improving safety probably isn't going to cause the speed reductions, safety and speed don't usually conflict. The fastest way is often also a very safe way because it involves less unnecessary exposure of people to hazards. The issue is the government deciding that work flat-out can't be done (eg, can't open a mine, can't build a highway, can't build a building, can't hire/fire someone, etc, etc). There is usually a good reason but at the end of the day building infrastructure is a break-eggs-make-omelette situation. We don't know how to build infrastructure at scale without collateral damage, as it were. If we first gain consensus that something is a good idea, infrastructure will not be built.
> People are still totally willing to be part of something like that; people are as haphazard and risky as they ever were.
I think people were much more willing to risk their lives (and risk other people’s lives) in the past than they are today. We have elevated the role of the individual. Younger generations are all “me” generations. (I’m not saying that’s a bad thing, it’s true for me as well.)
Case in point is the Western expectation of acceptable losses due to warfare. If we needed to pull off another D Day invasion, I don’t think we could. Generation Z is not signing up to die in tens of thousands and I don’t blame them. The casualty rates that were once seen as the normal cost of war and industry are totally unacceptable today.
I think you're partly right here. The ingredients for people to risk their lives is not right. As history has shown us though, that can change very quickly. I hope it doesn't for all our sake.
Prior to Pearl Harbor, most Americans wanted no part in the war. After the attack and growing threat of Hilter, people saw the need to bet our enemies at all costs by taking crazy risks.
People change based on circumstances. While newer generations might be considered "soft" and all about "me" compared to previous ones, I'd say they'd likely rise to the occasion if needed.
This is a somewhat misleading argument. The US did away with the draft so that most people have not had to make this choice. The troops that did go to the Middle East wars have suffered many casualties (6-7K dead, many more wounded, many US contractors dead/wounded, PTSD/suicides/mental health issues afterwards, etc.). [1] All those affected were volunteers.
It's also clear that there are times when virtually everyone in the US has agreed on the necessity of war, the 9/11 attacks being the latest case in point. Conversely, many wars that in retrospect might seem to have been widely supported were in fact quite controversial. The Civil War is a case in point. [2]
> The troops that did go to the Middle East wars have suffered many casualties (6-7K dead, many more wounded, many US contractors dead/wounded, PTSD/suicides/mental health issues afterwards, etc.).
That is objectively not "many casualties". Which was also his point (which you apparently did not understand, somehow).
And I find your implication that as long as people volunteer it's fine if they die quite disturbing, although funnily enough a perfect example of the "me"-generation the poster you were replying to was talking about! Who cares what my country is doing as long as I'm not affected right!
It seems you inferred something quite different from what I was trying to say.
Having a small fraction of US society bear the costs of prolonged wars is a terrible policy that brought unnecessary pain to 10s of thousands of American families and many more in countries like Iraq. It also allowed American involvement in those wars to continue unchecked. If we want them to stop the simplest solution is to bring back the draft.
And if we don't get consensus we'll pay the consequences for decades or centuries to come. E.g. mining operations in many places polluted the local waterways and underground water sources. Fracking today is another good example - if you ask the locals no way it will happen.
> People are still totally willing to be part of something like that; people are as haphazard and risky as they ever were. The government won't let them because it doesn't trust people to assess the risks properly (there is evidence that the government is right; people are very bad at assessing risk).
The government doesn't let people partake in hazardous work? Which government is that? For example, the U.S. government will not stop you from working in mines, electronics industry, construction, or in automotive production. Corners cut when it comes to safety to save on cost still happens everywhere.
I've heard that SpaceX is a bit like that, and is advertised as such. People sign up. I do think people are willing to plow into things that truly are that inspiring, but we just don't do much of that stuff. It's usually not profitable.
It's also not always pleasant. A guy who went through special forces told me once: "I'm glad I did it, but I would never do it again."
This might be an unpopular comment, but I think at least part of the reason you don't see people plow into inspiring things these days is less about lack of profitability and more about an overabundance of diversity.
Inspiration looks different to different people. If people don't have a shared vision, they aren't going to be able to collaborate on shared projects at the same level as people that do have a shared vision.
Then there's Apple under Steve Jobs and the covert 5-year-long project that lead to the creation of the iPhone. I don't think it was that fast compared to what companies achieve nowadays, but it certainly was accomplished due in part to Jobs' quasi-dictatorial leadership.
Painful and discouraging virtual casualties which make it difficult to justify more ambitious or even equally functional efforts under 21st century conditions.
Agreed. These days it's impossible to select a site and begin demolishing the neighborhood within the same month.
> [T]he Sausalito site selected on 3 March... construction began on 28 March.... Construction start was delayed two weeks to allow the 42 families living on Pine Point, which was scheduled to be demolished to build the shipyard, to move.
People certainly died in the construction of the Empire State Building. I think I calculated that the risk of death for a worker on that was higher than the risk for a soldier deployed in Afghanistan.
I think it might also be a reflection of general polarization. Collective action is hard (certainly easier under a dictatorial regime, e.g. China), but in earlier days in the US we had more common purpose -- whether it was a sense that we are building a nation or WWII, those kinds of things -- which to some degree compensated for the inherently inertial nature of collective action within a liberal democracy.
It was also a time when it was more socially and politically acceptable to steamroll those who didn't share or was part of that common purpose. Not only in the US, but all around the world.
Same in Europe. Brussels city in the 1900s moved 20.000 families to build a single street (now the main pedestrian area). Mussolini paved over dozens of ancient Roman sites to build a big street through the city centre. Paris went through an entire planned effort to reshape the whole city, which created all the famous avenues you know today and at the same time created all the social problems (poor 'banlieue' areas around the city) we still know and love today.
> what model there might exist for moving quickly with a multi-stakeholder process.
There is no model for that.
The OP's list of "fast" was more of a compilation of cherry-picked projects that, with the right context and in the right hands, achieved an impressive outcome.
Moreover, each of them AFAIK had long prerequisites to get to the point where they could reach their endpoints. The path to great things invariably includes many failures, false starts, and uncertainty.
The author, perhaps, has watched too many films with breezy montages that end with triumph. It doesn't happen that way, except in the wet dreams of project managers.
> The author, perhaps, has watched too many films with breezy montages that end with triumph.
In fairness, the author co-founded Stripe when he was 21. So, perhaps it's fair that his views are colored by the fact that at a very young age he built the highest valued company in the history of the world's most successful startup incubator, but saying "it doesn't happen that way, except in the wet dreams of project managers" is, well, I'll just say a bit funny given who the author is.
To the high performance operator, inefficient bureaucracy has been a recognized thing for millennia.
The associated deadlines, proven over millennia to never be realistic.
Regardless, this is the mainstream, and this is how institutional projects are built and at institutional speed, i.e. sluggish. You know, like a snail without a shell.
Alternatively, for an equally long time now, in a certain way nature has favored those few individuals who have the instinctual need for non-sluggish, dare I say "fast" myself.
These are so hard to find that most time-sensitive projects simply start without them and substitute traditional deadlines and bureaucracy in order to measure long-term progress which would otherwise not be apparent.
When what is really needed is leadership from the top down building the team exclusively hand-picked for the very instinct which has always outperformed mainstream operation. But it takes one to know one.
When everybody on the job has always had the knack for outperforming anything deadlines have to offer, you don't have to remind them you needed everything "the day before yesterday".
No need for (lack of) progress reports either, when the goal is moving visibly closer continuously and it's turning out not to be a long-term project after all.
I’m curious about your last statement. You wrote “the author, perhaps, has watched too many films with breezy montages that end with triumph.” Then you made an even more disparaging comparison to wet dreams that is too distasteful to quote.
How many billion dollar companies have you started?
Japan seems to be completing civil engineering projects pretty fast compared to almost any other nation. Maybe there's something to be learned from there?
The lead engineer who completed the Toyota Camry Hybrid in record time dropped dead shortly afterward. Japan keeps its development speed high by burning through employees aggressively.
Since 1990 Japan has indeed only added $6tr of Money (M2), but as their GDP hasn't really grown that's an additionall 80% of GDP in terms of Money (the same meausure for the US is up 15%)
China on the other hand, money is up 150% of GDP, which clocks in at $26 Trillion dollars (a bit more than 100% of US GDP).
I think one person / body in charge but willing to listen sympathetically to the stakeholders before making a decision. I'm having a job thinking of an example but Hong Kong under British rule did pretty well. Run by UK civil servants tasked with doing well for the place and built a lot of infrastructure and housing.
Why is it incredible simply to do something fast? China is not a model to emulate. They throw a massive number of people on projects, pays them next to nothing (neither when they are maimed, nor to their families in case of death), and allow them hardly any time off. Wherever they're employed, it's not uncommon for Chinese men, and women, to only see their family 1-2 times a year.
The fundamental model that we forget today is that impressive things are built by oppressing millions of people.
From the pyramids, to Notre Dame, Neuschwanstein Castle, Great Wall of China, Panama Canal, Chichen Itza, Machu Picchu, Cancun hotel zone, etc every single technologically unbelievable achievement was built by exploiting large masses of the population.
The reason we cannot build the same today is that the pool of people that can be exploited has been much reduced.
Hallelujah. This is a passage about St. Petersburg, sometimes referred to (among other cities) as "The Venice of the North":
Starting with the construction of the Peter and Paul Fortress, Peter dragooned thousands of conscripts, convicts and prisoners of war to erect the city from scratch in a place where snow can fall as early as September and as late as May. Tree trunks had to be sunk into the swampy ground before it could support structures.
Living in ramshackle quarters and working with inadequate tools – often digging by hand and carrying the dirt in the front of their shirts – these involuntary labourers died in their thousands, carried off by disease or frequent flooding. As a result, St Petersburg became known as the “city built on bones”.
It is a massive undertaking (also including the resort chains down Riviera Maya), gigantic opulent sprawling palaces, built-in about two-three decades, in a remote place, jungles for hundreds of miles in every direction, in an area with no pre-existing infrastructure,
It was built super cheaply on the back of millions of Mayans - it is on the list because IMO is the most recent and perhaps last example of building "pyramids".
I think the "aesthetics" is quite culturally indoctrinated - I find the interior of baroque catholic churches garish and unbearably cheesy.
The resort construction in Mexico wants to deliver a grandiose feeling to tourists - and it does do that perfectly. I am repeatedly amazed at the architectural variety and talent that went into shaping and designing each resort. There are very few that are similar to each other or a "traditional" hotel.
I will also say that many of the resorts in Riviera Maya are anything but cheap quality - heavy marble floors, thick noise absorbing walls, high ceilings. You can almost sense the Mayan pyramid building ancestry in each.
We also used to have stronger public institutions that could maintain long-term vision, weren’t subject to constant political upheaval or budget cuts, could reliably staff planning and execution phases, didn’t have to put everything out to bid, negotiate with overdue/overbudget vendors, set up new processes for each project, etc.
Bureaucracy involves some waste in overstaffing, but also saves on the chaos that results from our starved, politicized institutions with little internal execution/planning capability.
IMHO, if there is one thing a communist government can execute on, it’s infrastructure. China now has 2/3 of the worlds high speed rail, despite being an emerging economy, and only starting 20/25ish years ago.
But unlike back then, we have power tools and vast mechanization. One Diesel powered machine can do the work of thousands of men. If we wanted to build a canal today we ought to be able to do it much faster regardless of carelessness level.
It surely a machine that can do the work of 1000 people more than compensated for extra safety.
Even if you have 1000 people you’re willing to sacrifice, if just one machine can do all their work then carelessness can’t be the main reason we’re so slow now.
Or maybe they use the machines and the same amount of people and just do more? It they did less it would slow things down....oh wait, that was the original point.
Umm the Irish Channel is just a neighborhood name. It's not a canal. Yes, thousands of Irish died from yellow fever and malaria while digging canals in New Orleans (Basin St and those going to the lake).
They were considered expendable because they were Catholic (unlike the Scotch Irish who were protestant). They were also lied to and told that New Orleans was close to the Irish communities in New England.
> Neighbors previously complained about the replacement of historic streetlight poles for Van Ness BRT. Replacement poles made to look like the originals ballooned the project cost by $6.5 million, according to city documents
This is a perfect microcosm of San Francisco governance. A small number of highly-motivated landowners unashamedly hijacking city policy for their own interests. How do we stand up to this?
I don't know, the charm of San Francisco is all of its unique elements that aren't found anywhere else. The hills, the cablecars, the Golden Gate Bridge. Why shouldn't the transit dept be responsible enough to "leave no trace" after it completes a project? What if the transit project tore up Fisherman's Wharf and they decided to replace it with a generic shopping mall? Or took out all the cable car tracks? What if they flattened Lombard Street to make it more "pedestrian friendly"?
Vote! Fight for candidates who will bring sanity to the system. Donate money: it's amazing how powerless tech is in a city where it dominates economically.
Aggressively tax landowners on the value they did not create and let them keep 100% of the value they did create.
They didn’t (solely) pay for the $6.5MM light poles, so they shouldn’t (solely) get the real estate value upside from having them there. Look up the Land Value Tax and start pushing for it.
I for one appreciate when a modern project takes care and expense to recreate certain historic elements that tie the project into its setting more seamlessly than a lowest possible cost execution.
I appreciate that whether it’s directly in my backyard or across town.
You forgot Banting and Best[1]. They went from coming up with the idea for synthetic insulin, developing the treatment, testing it on animals, then themselves, then testing it on patients, then developing and releasing a commercial product and winning the nobel prize in medicine in 3 years! This was of course long before the FDA was around and if they tried to do the same thing at that pace today we'd throw them in prison.
Genentech (dave Goeddel, herb heyneker, Dennis Kleid, herb boyer) cloned human insulin in a year (1977-8 I think) and got it approved by FDA ~4 years later
Early on in the Apollo program there was the plan to build a series of spacecrafts which would progressively solve intermediate problems on the path to the Moon landing. For that, a series of Apollo spacecrafts were planned, and production was running for years - Apollos flew also to Skylab and to 1975 docking with Soyuz.
Specific Apollo 8 was chosen to fly to the Moon because LEM wasn't ready - it was later tested on Earth orbit with Apollo 9 and flew to the Moon with Apollo 10 first time. So, Apollo 8 flying to the Moon isn't a separate decision, but change of existing plans - with a couple of missions swapped in time; fast, but not nearly as fast as creation from scratch of a spaceship or a program would assume.
True, but still a dizzying feat by any standard. World's in-a-class-of-its-own largest rocket, so far only tested in two unmanned launches, one of them a near disaster, stability problems now theoretically fixed, but fixes never tested in actual flight. Sure, let's improvise a new kind of missio in four months time, strap people on top of that thing, and shoot them to the Moon for the first time ever.
I don't believe that's quite how things are done these days. Wildly successful things, mind you.
I'd like to see the opposite: slow projects that turned out awesome.
Doing things really fast is more often than not the result of external circumstances. Maybe some new tech has just been made available, maybe it was a life-or-death situation, where dying at work is considered acceptable (ex: at war), maybe it had an unreasonable budget, maybe the "fast" project was just the tip of the iceberg, made visible by decades unnoticed of ground work, or maybe it massively prioritized speed over quality.
These situations are unlikely to relate to your projects, so yeah, things are going to be slow for you. Buy that's better than expecting some miracle, or considering human lives as disposable.
What would be more inspirational would be a list of slow projects, where people are trapped for years in development hell, with cost cutting, bad technical choices, mistakes and mismanagement. Projects where everyone involved say it is doomed and yet, at the end, something awesome came out.
If you are working on a building, what's the best thing to say? "That building took 1 year to make, you are at 2 and you didn't finish, you suck" or "That building almost crumbled, contractors defaulted, it took 10 years, but now, it is the pride of the city"? In the second example, you can see the mistakes and how they fixed them, the first one most likely involves a lot of luck and just sets unrealistic expectations.
The development of Git May have started in the back of his brain earlier than that. Code being written is kind of the very final stage of software dev.
>So I’d like to stress that while it really came together in just about ten days or so (at which point I did my first kernel commit using git), it wasn’t like it was some kind of mad dash of coding. The actual amount of that early code is actually fairly small, it all depended on getting the basic ideas right. And that I had been mulling over for a while before the whole project started. I’d seen the problems others had. I’d seen what I wanted to avoid doing.
But probably, the design of git, especially it's datastructures, started much earlier in Linus' head. I think he must have elaborted over this for quite some time.
Empire State Building has always boggled my mind, with its speed. Built faster than most houses today, biggest building in the city by far, in the middle of the city, amazing endurance. I need to read more about the process. I wish governments everywhere learned a few lessons from it.
I also remember that the construction site did not have a stellar safety record. It was typical for the industry back then, though. In the middle of the Great Depression, a builder would likely not be too picky and reject a chance to work on such a high-profile project when unemployment was high.
A relatively recent example of rapid infrastructure building (in California no less!) can be found with C.C Meyers.
"Proclaimed a “Local Hero,” Clinton “C.C.” Myers was lauded on the cover of the July 1995 issue of Comstock’s magazine for “working miracles in heavy construction.” Three big projects by C.C. Myers Inc. cemented his legacy: rebuilding two bridges in Watsonville after the 1989 Loma Prieta earthquake; reconstructing the Santa Monica Freeway after the 1994 Northridge earthquake; and rebuilding a stretch of Interstate 5, including a bridge, near Coalinga after the 1995 floods. He was equally famous for the efficiency of these projects — 55 days in Watsonville, 66 days in Santa Monica and 21 days in Coalinga — and the huge bonuses he earned for finishing projects early."
Constraints foster creativity (if the right people are in the team). When JFK announced (in 1961) USA's intention to put a man on the moon before the end of the decade, he set a time constraint that could not be moved without considerable global embarassment. In less than 9 years, it was accomplished. Perhaps the formula is: a deadline + right people + leadership support.
I wasn't just money, it was changing the rules for spending the money. PhD students were funded. Science & math teachers were well paid and respected. People wanted to work on the space program, even if on the periphery.
It would still be great to work at SpaceX, even if Elon is diet Howard Hughes.
Even adjusted for inflation, the U.S. federal budget is far, far larger than it was in the 1960s. So much so that even though NASA's budget is now less than 0.5% of the federal budget, NASA still spends roughly enough money in a decade to pay for the Apollo program. I've seen various figures for the cost of the Apollo program, ranging from $150 to $290 billion (adjusted to current dollars). In the last decade, NASA spent over $200 B.
I wonder how quickly we'd solve fusion with this level of focus. I bet it'd be less than 5 years to "energy too cheap to meter", unlimited carbon capture, etc, etc. Instead we've been bombing people in the Middle East for the last 20 years.
I think I’ll trade not living under the threat of nuclear holocaust and desire to beat an ideological enemy for getting to the moon slower. A feat which while fast possibly slowed the natural progress of getting and staying there, and which has as its main legacy intercontinental nuclear capabilities.
Well fast aircraft sort of lost their purpose once satellites got good, so it wouldn’t surprise me to find out they stopped caring about manned, fast aircraft with the SR-71.
Look at the defense allocations for basic research into hypersonic capable materials to prove that they are still years away the most basic deployment and once the research is done, there is no supply base to make the materials (FA8650-20-S-5003)
Pyramid at Giza, 20 years -- laying one 2.5-ton stone block every 2 minutes, of every daylight hour, for the entire period. Some people say 10 years, one per minute. Or, construction for half of each of 20 years, one per minute.
You could imagine getting pretty good at it, five years in.
A lot of these feats were the first _successful_ version, however the people behind them had careers of doing the same thing over and over again.
It's not like Brendan Eich went from not knowing how to write a language to shipping a language in 10 days however this article presents it like this.
It's not like Boeing went from nothing to 747 in whatever amount of time. Building an airplane get much easier when you already have a factory for building airplanes.
Like I can work on something for a decade and day before shipping, give it a new name and say "hey, I shipped it in a day".
It might take a chef an hour to make a meal. But he might have been optimizing that hour his whole life.
The tale of the Chinese emperor and the picture of the rooster
Once upon a time, there was an emperor who was extremely fond of roosters, and so one day he commissioned the most famous painter in all his realm to paint him a picture of a rooster. The painter replied that it would take him three years to accomplish this. The emperor was secretly rather annoyed, but in the end he agreed.
When the three years came to an end, the emperor summoned the painter, but on seeing that he was empty-handed impatiently demanded: "What about my painting of a rooster?" The painter remained perfectly calm, took up a piece of paper right where he stood, and started to paint. The brush flew across the page, the ink danced. With apparent ease, he produced the lifelike image of a rooster, capturing its very essence. It took him less than three minutes to complete. On seeing this, the emperor was furious and could stand it no longer: "Have you been deliberately deceiving your king? Is this some act of rebellion? It took you just three minutes to paint that picture; why did you make me wait three whole years?" To which the painter replied: "Sire, first please calm your fury; and when you have done so, follow me and see for yourself."
The painter led the emperor to a large house and opened the door, and on looking inside, the emperor realised that the house was filled to the roof with sketches of roosters. Then the painter spoke: "These are my efforts of the last three whole years. Without those three years of labour how could I possibly have produced that perfect rooster for you in less than three minutes?"
Not sure why this is getting down votes. This needs to be kept in the back of the mind that overnight success is a fallacy. Not sure why pc chose not to put the idea that each of those fast achievements had years of works behind them.
Why not add "Usain bolt took only 9.86 seconds to run a 100 m"
Is he claiming that the people involved weren't experts? It doesn't seem that way to me.
I think it's just the opposite – he's demonstrating how much rapid progress can be made when you cut the red tape and get the right people working at full capacity.
He then contrasts this with the municipal government of San Francisco, which operates under a different set of principles. Namely: maximize red tape, use unqualified people, and exert minimal effort, all while lighting money on fire because you have an enormous annual budget and no accountability.
I agree that's the point, but for all we know there could be years of education, planning, failed trials, preparation, and/or red tape omitted from these timelines.
was having the same thought as i was going through the examples. and when the unix example came up, it dawn on me that the author is clearly omitting the background of these people to make a more compelling point.
yes they did great feats but most have a history of doing similar and learning from it in the past. fast-forward to present days, and our issues with not going fast enough, maybe the conditions and situations are different? maybe we are in a learning phase?
also for a better account of what happened in unix case, i recommend kernighan's latest: unix: a history and a memoir.
Regardless, there are lots of stones to be thrown at our inability to build infrastructure and do projects that are not novel in any way on-time and on-budget.
Poor Brendan. Wanting to write Scheme for the browser and ending up writing JavaScript. Sure, today JS is a proper language, but it atillmlacks the elegance and adherence to principle of scheme. Maybe that's why it's so successful :)
Almost every item on the list is from before 1960. After 1960 there is not a single construction project in the US. The later projects happened either in unregulated areas (JavaScript, Amazon Prime) or in China (Shenzhen, Luckin Coffee), with a single European outlier - the TGV.
> The war was under way in Europe, so the Army's need was urgent and demanding. Bids were to be received by July 22, a span of just eleven days. Manufacturers were given 49 days to submit their first prototype and 75 days for completion of 70 test vehicles.
> Probst turned down Bantam initially, but agreed to work without pay after an Army request and began work on July 17, 1940.[39]
> Probst laid out full design drawings for the Bantam prototype, known as the Bantam Reconnaissance Car, or BRC, in just two days, and worked up a cost estimate the next day. Bantam's bid was submitted, complete with blueprints, on July 22.
A lot of the construction projects came at the cost of bad safety standards and general poor treatment of workers, including plenty of workers deaths. Something to keep mind when bemoaning the slower speeds nowadays.
The 1000 dollar human genome [1] is one of the most staggering examples of this.
"In dozens of presentations over the past few years, scientists have compared the slope of Moore's law with the swiftly dropping costs of DNA sequencing. For a while they kept pace, but since about 2007, it has not even been close."
To go from a cost of 2.7billion and one decade to where we are today (people are now talking seriously about the zero dollar genome, aka the price of sequencing can be trivially offset by selling some meta information to some company or other) can be seen as a feat of scientific infrastructure-making that blows any Alaskan highway out of the swamps.
The Manhattan Project moved along remarkably quickly. From the time 56,000 acres were authorized for purchase at Oak Ridge, to the Trinity test, was under three years (September 1942, July 1945).
In 5.x years id Software built and released: Wolfenstein 3D, Doom, Doom 2, Quake, Quake 2.
It’s always interesting to see what kinds of things make it to the front page because of who wrote them, rather than the content itself (which in this case is just a few cherry picked examples with no context). I’m not exactly sure what anyone is supposed to learn from this aside from the facts themselves.
When we pick a set of facts to highlight, we are editorializing by what we choose to include or not include in such a list. This is the core of every good documentary I have ever seen: ultimately, they are compelling editorials because they construct a narrative by telling or showing you things that actually happened.
Whether or not those things are a representative sample, or plausibly support the narrative in the wider context of human society, is left as an exercise to the reader.
I recall reading once that Thiel said that most or all of the critically important events in history were accomplished by small, highly motivated teams. (Pardon me if I am misremembering the quote exactly.)
Perhaps pc is simply reminding us that “move fast and break things” isn’t always a bad thing? Just speculating, though.
Nothing wrong in itself. But it is appropriate to wonder about the takeaways. Without proper context and history, the lessons you draw might not be helpful ones.
I just went to Dubai for the first time ... one thing about Dubai is they can still build fast. All of Downtown Dubai was built in about a decade and it is really something. Tallest building in the world, largest mall in the world, multiple huge public spaces and many luxury hotels and residential buildings. All built in the 2000s in the shadow of the financial crisis.
And what they're doing for the upcoming 2020 Dubai Expo is equally mind blowing.
Before the downvotes come: I know it's an autocracy and the building labor practices are questionable. But many achievements on the list are subject to those caveats so it deserves mentioning.
San Francinsco's city government is completely incompetent and the city itself has devolved into a disgusting cesspool.
But if you live in the city, you don't mention this. Instead, you loudly proclaim at every opportunity that San Francisco is best city in the world, and you express your pity for the poor schmucks who are forced to live anywhere else.
The crime, the drug addicts, the homelessness, the mental illness, the broken car windows, the absurd cost of living, the blighted neighborhoods, the pervasive human misery, the garbage, the heroin needles, the human feces -- this is to be considered part of San Francisco's urban charm. Anyone who acknowledges how bad the situation has become should be ostracized, downvoted, or told to move to Walnut Creek.
You especially shouldn't be critical if you run a tech startup in the city, since you need a continual supply of young, impressionable employees to buy into the hype and move here.
But once you move just outside the city perimeter, then you can afford to be a obliquely critical of the city's slow bus lane construction.
Finally moving out of SF was the best decision I ever made. It only became clear once I was removed from SF for enough time. Fuck that place. Yeah, it's the home of technology, but it's slowly rotting and dying and it makes its resident miserable terrible people.
I'm happy to be living somewhere I love now (HK). And if things ever change, I'll hope not to make the same mistake and stick around so long.
Congrats on getting out! I started a new job here a few months ago, but I'll be requesting to transfer to a remote team as soon as I hit my 1-year mark.
I think both things can be true at the same time. San Francisco is an amazing place to work in technology (in my view, nowhere else comes close). But the municipal government is also a complete cesspool of inefficiency & corruption.
Cheap, fast, and good are three interesting measures. It’s often said you can pick any two, and from experience I would have to agree.
I’d be curious to hear about projects spanning this gamut – something done fast n’ good but costing dearly; fast and cheap but horribly executed (I’d bet a large slice of projects fit in quite well here), or the amusing slow, expensive, and ultimately shitty. Perhaps most interesting would be story or two about projects which miraculously attained the trifecta... that is, if any of these stories actually happen to exist.
He's rightfully dunking on the Van Ness bus line for some reasons, but it's basically a major utilities infrastructure project disguised as a bus lane. They're overhead power and reconfiguring the street, but they're also replacing ~4 miles of 1800s water main, putting in a new earthquake-resistant sewer system, and redoing the fire hydrant water feeds.
I feel like it's more of a branding failure honestly. People would be much more understanding - if you heard "we're replacing 1800s utilities with 21st century tech built for the earthquake zone in SF and you get better transit along with that", I think the conversation would be much different than "it's taken a decade to build a new lane".
That reminds me a bit of the California High Speed rail project. A bunch of the work in the central valley is grade separation for existing freight rail lines.
But then we'd wonder why the transit needs to be conditioned on the utilities. This is political malpractice undermining support for bus lanes citywide.
News article says [T]he $189 million project, which is part of a larger complimentary (sic) sewer and street light replacement project totalled at $316 million.[0]
The new Berlin airport is a classic example, and one that really undermines the stereotype of German efficiency and competence when it comes to infrastructure projects.
Across BER airport, the Elbphilharmonie and Stuttgart 21, the stereotype of German efficiency is dead in the water at least as far as infrastructure projects are concerned.
That goes for all cathedrals though. Just two random examples, Notre Dame took about 100 years for the main structure, Cologne Cathedral took 600 years.
Crazy Horse Memorial has been dragging out too. Got the impression that what exists now is functional enough to serve as a tourist trap so there's less incentive to complete it.
Hum, I think things should be broadened beyond just a fast/slow axis, and inquire what some of these projects actually serve. Or at the least what they've costed in the long-run in terms of resource use.
Definitely! That would be very cool to see. I bet there are a lot of great things that have moved slowly with great success. Berkshire Hathaway is probably a good example, although I don't know how it compares to other companies during a similar window of time.
>And one of the most fundamental errors made
in advanced developments is to stick to a design
even when it is very complicated,
and to not strive to delete parts and processes.
I do think these kinds of things are worth remembering, but looking them over it just reminded me of the quote, "it took me 10 years to become an overnight success"
From The Everything Store, A book by Brad Stone -- what I got to know was that Amazon had the pieces in place that eventually helped create Prime service such as efficient warehouse operation that can deliver packages in 2 days, quick payment processing, rich customer data etc. So to think that suddenly Amazon Prime was invented and implemented in just 6 weeks seems wrong.
This is all so ridiculous. Yes so much can be done quickly if you take chances, ignore worker protections, cut corners, and do things like they used to be done back in the day.
Let's take this Marinship as one small example of why almost everything on the page is probably not relevant today:
> Physical construction began on 28 March. Construction start was delayed two weeks to allow the 42 families living on Pine Point, which was scheduled to be demolished to build the shipyard, to move."
So 'delayed two weeks to allow 42 families to move'. Can you imagine what would happen now or even 20 years ago? It would be tied up in courts (including the court of public opinion) for years even decades. Lawsuits filed, legal challenges and so on. (see 'Blue Route' in Philadelphia) [1]
As a programmer, I often stumble upon the performance arguments and pitching technology Y over currently used X "because of better performance, i.e. it's fast".
Sure, the impressive feats of engineering accompanied by a short time to build, sounds good. But how much money and time went into maintaining all the flawed decisions?
The stories of Xerox Alto's first GUI-oriented computer are usually accompanied by how Apple stole it and made it better. To me, this sounds like their speed did not help them ship a great product to the end-users.
JavaScript was prototyped in 10 days, which is impressive. But there are also many jokes using this same fact as a punchline. This quick development was paid many times over by having to deal with poor language decisions. It took another 20 years for the language to start moving and developing in a more developer-friendly direction. And some of it will never be better because it has to be backwards compatible for all browsers that ever existed.
Comparing a "highway" (later edited to the military roadway) from 1942 to construction project in one of the larger cities in the world is not fair. I agree the time it's taking, and the cost is absurd. But the cost of work has drastically changed since, and the 1942 number most likely doesn't account for the fact that it was built by soldiers who are already on the payroll, so there were probably little additional worker expenses. And the standards now are more strict than they used to be.
Calling the Eiffel tower a building is a bit like calling a bridge a building. It was built, certainly, and it even has a small portion that offers "rooms", but it's really just a structural skeleton. And in that light, 2 years 2 months is really not all that surprising, or even fast.
>TGV. On April 30 1976, the French government approved a plan to build a high-speed rail link between Paris and Lyon, the first high-speed rail line in Europe. This line was to use completely new electric locomotives, also to be developed in France as part of the project. The ensuing line opened on September 26 1981, 1,975 days later. On September 24 1996, the California High-Speed Rail Authority was formed. The completion of the first phase of California's high-speed rail project, a line connecting San Francisco and Anaheim, is currently estimated to happen in 2033, 37 years (i.e. around 13,000 days) after the authority was formed. Source: On the Fast Track.
Wow, what the hell is wrong with the TGV in California? Are they still working on it? 33 years!
This isn't necessarily a good thing. The earliest versions of Javascript were insane. Bad isn't the word. Buggy isn't the word. You've never seen a language as ridiculous as early Javascript. It took a good twenty years to fix it.
Subways drastically vary depending on the material the ground is made of and what is there already. For example London is relatively easy to build deep tunnels in, but on of the big things that has increased the complexity of Crossrail is that central London has a lot of stuff underground that isn't necessarily documented correctly. There are water mains, sewers, electrical cables, ventilation shafts, other subway tunnels, stations, then abandoned instances of all of those and plenty of archeological materials (London as a city has been around for over 2000 years).
The TGV entry skips over the 20 years of R&D on high speed rail that preceded the first major construction, much of which was wasted on gas turbine designs.
While I do get the "spirit" of this, I feel it's highly inaccurate and/or misrepresentative in some cases.
An example would be Javascript. Sure he implemented a prototype in 10 days, but the whole implementation was probably a fraction of the test code for V8, a modern implementation of the beast that is Javascript today. I'd even argue that the VM engines from 5 year ago would still hold the same property.
Things were built quickly in that era. Most skyscrapers wenr up in a year or so, like the Empire State Building. The whole Hoover Dam project took less than five years.
> Japan, an island empire totally dependent on maintaining open sea lanes to ensure her raw material imports, managed to build just sixty-three DDs (some twenty or so of which would have been classified by the Allies as DEs) and an unspecified (and by my unofficial count, relatively small) number of 'escort' vessels. In the same time span, the US put some eight hundred forty-seven antisubmarine capable craft in the water!
> The United States built more merchant shipping in the first four and a half months of 1943 than Japan put in the water in seven years.
> Again, the United States had to devote a lot of the merchant shipping it built to replace the losses inflicted by the German U-Boats. But it is no joke to say that we were literally building ships faster than anybody could sink them, and still have enough left over to carry mountains of material to the most God-forsaken, desolate stretches of the Pacific. Those Polynesian cargo cults didn't start for no reason, and it was American merchant vessels in their thousands which delivered the majority of this seemingly divinely profligate largesse to backwaters which had probably never seen so much as a can opener before.
Agreed, it is nuts. The Liberty ships being built in just days. Planes and tanks and trucks streaming off the assembly lines in the tens of thousands. Millions of tons of high explosives. Truly incredible.
Nowadays, it takes us the better part of a decade to get a minor bridge project done, and building a four-story apartment building is a multi-year endeavor.
Madrid Baracas Terminal 4 roughly equivalent to building the whole of Heathrow airport i.e including 2 runways and a metro line extension. 7 years from conception to first passengers.
There is the B-29. The US spent more money on that than the Manhattan project. I think the US spent even more money on radio and radar development during WWII.
I think it's easy to forget that although these things were achieved in a very short space of time, it's not without years of toil that lead to it.
For example, you can't write a programming language in 10 days, without having spent a long time thinking about how you'd go about doing that prior to doing it.
2019: San Francisco Web Developer, also opposed to the new bus lane, is still configuring webpack and package.json. Their intent to build a blog was announced in early 2015 via a Twitter post containing way too many emoji.
Covered quite well in other comments but: health and safety, working around existing infrastructure, political constraints on throwing money at problems[2], lack of skilled workers[1], etc.
[2] Although if you can convince politicians that infrastructure helps fight terrorists/blocks immigration/lowers taxes for the rich, you'll probably be drowning in money pretty quickly.
lol, here in Germany things somehow never get done on time.
We are still waiting for an airport in Berlin and a trainstation in Stuttgart. While the last one still has about 3 years to be "on time" it will probably be late.
When you see examples of waste like this, which are not uncommon, it really doesn't take much critical thinking to become skeptical of those in favor of increasing taxes to pay for solutions managed by any Government agencies.
>>San Francisco proposed a new bus lane on Van Ness in 2001. Its opening was recently delayed to 2021, yielding a project duration of around 7,300 days. “The project has been delayed due to an increase of wet weather since the project started,” said Paul Rose, a San Francisco Municipal Transportation Agency spokesperson. The project will cost $310 million, i.e. $100,000 per meter. The Alaska Highway, mentioned above, constructed across remote tundra, cost $793 per meter in 2019 dollars.
One of the parts of the answer may be that organizations are now designed for more creativity, being better places to work, and maybe even throughput at the expense of latency.
FYI This page doesn't display very well on my mobile, and Firefox does not seem able to put it into reader view, which seems odd for a page of plain text.
I just started following Patrick Collison's career and have read 2 books from his reading list. he's one of those people that you feel lucky to be living at the same as. Going through his website is fascinating and he does this in his spare time while also building one of the most admired startups in existence. I'm excited to see what he does next. What a fascinating mind
Bear this in mind when people express skepticism about the existence of the 10x developer.
My own productivity can vary over approximately 2 orders of magnitude day to day so I don't have any problem with believing in individuals differing by a single order over time.
Here's another one that I think belongs on this list:
NYC spent 6+ years and $13 million trying to reopen the Wollman skating rink in Central Park. In 1986, Trump negotiated a deal with the city for 2½ million, promising completion in 6 months, by Dec 15, in time to salvage most of the winter season. He and his team opened the rink 2 months ahead of schedule, and ¾ million under budget.
This isn't about economic progress, it's about humans getting things done in a short amount of time and with fewer resources that what we currently hear about. If you want to expand his list to include other things why don't you send your examples to him?
> Should we really be worshipping economic progress?
Yes.
The genie is out of the bottle, there's no going back to some imaginary idylic pre-industrial world. The best way to preserve what there is and free up more space for preservation is to pursue technology and efficiency ruthlessly. Do more with less.
We can have development with constraint, we can have development without the cult of growth, we could go back to the more inclusive, constrained form of economic progress seen in the immediate post war. We could even try to tune the system in new ways to limit the clear problems of unconstrained globalised capitalism.
We do not need to go back to a Medieval agrarian feudalism, return to mercantilism or unwind the industrial revolution. It happened, we should learn from it and make the best of it...
I grew up in Dawson Creek. Our claim to fame: "Mile Zero of the Alaska Highway starts here!"
The Alaska Highway as built in 1942 is nothing like a highway that most people would envision. It wasn't paved, for example. It was good enough for military vehicles with crews of soldiers who could make ad hoc repairs to the road as needed while they transited through, who had extensive survival skills, and could literally radio for assistance if needed.
It wasn't opened to the public until 1948 -- so perhaps 6 years, not 6 months, is a better estimate of its time.
It's also shrunk by almost 20% as it has been continually rebuilt to make it passable by passenger vehicles.
It's an accomplishment, no doubt, but I feel like omitting these significant caveats is meaningful.