It's time to build, writes Marc Andreessen. The piece is framed in the context of the Covid pandemic. The first part of the essay is about preparation for pandemics,

Part of the problem is clearly foresight, a failure of imagination. But the other part of the problem is what we didn’t do in advance, and what we’re failing to do now. And that is a failure of action, and specifically our widespread inability to build.

There is one more that can be added, besides failing to plan or build: failing to maintain.

In 2006, California, under Gov. Schwarzenegger had a plan in place to address pandemics like the current one, and it wasn't just a plan, the equipment required was actually manufactured and subsequently purchased.

The state, flush with tax revenue, soon sunk more than $200 million into the mobile hospital program and a related Health Surge Capacity Initiative to stockpile medicines and medical gear for use in outbreaks of infectious disease, according to former emergency management officials and state budget records.

But then

But the ambitious effort, which would have been vital as the state confronts the new coronavirus today, hit a wall: a brutal recession, a free fall in state revenues – and in 2011, the administration of a fiscally minded Democratic governor, Jerry Brown, who came into office facing a $26 billion deficit.

And so, that year, the state cut off the money to store and maintain the stockpile of supplies and the mobile hospitals. The hospitals were defunded before they’d ever been used.

Much of the medical equipment – including the ventilators, critical life-saving tools that are in short supply in the current pandemic – was given to local hospitals and health agencies, former health officials said. But the equipment was donated without any funding to maintain them. The respirators were allowed to expire without being replaced.

Now one may wonder what can be done to avoid this entire class of mistakes. Sure enough, becoming a prepper society is not the most efficient thing there is, but if you are going to predictably have impredictable shocks (As to the when and what exactly) that affect everyone, you probably need those emergency plans. So the working assumption should be that there will be a next global pandemic for sure, what is unknown is what virus and where it will come from, but it will come. There will be plenty of time to plan and build and prepare, but those preparations have to be maintained over time.

Death by friction

Finance is another of the topics touched on by the article; the fact that the US government can be paid by everyone but not easily pay everyone probably shares something with the US tax hell; in at least the UK there is no need to think about taxes for most people the government knows everything they need to know about you anyway so they are paid for you. In the US it is the same case - that's why tax fraud is a thing, they are double checking your work for no good reason. In the UK I was once taxed in excess. As the meme goes, you won't believe what happened next; I just had to log-in in a -nicely- designed website, write my bank account details in a box and two days later I was refunded. Just like that. No need for complex forms, any paperwork, phone calls, or anything of the sort.

Venmo -and to a lesser extent, Plaid- are symptoms of failure too, they are patches to problems, not fixes to their root cause, they are -welcome- innovations around a broken system but not a fix for the system itself. The real need for their existence means that the underlying systems are insufficient to meet the demand for fast payments or easily accessible APIs. In the UK the new wave of banks like Monzo or Starling, and probably all the other ones at this stage, provide the same functionality as Venmo, only that the payment is real (It lands in a real bank account instantly). Now why can't you do instantaneous payments in the US? For banks, the story won't surprise you: Turns out that starting a new bank is not as easy as your cookie-cutter SaaS business. Without going to fully free banking you can do better; again the UK is the leader at fintech for a good reason; not only they crafted favorable regulations but also issued good guidance and explainers of the steps one needs to follow to start a bank or set up a regulatory sandbox to help banks get started without having to go through onerous approval processes1. When you reduce friction to go from idea to product you get more products, you can either increase the desire to build or you can make building easier; but changing desires is harder than incentives.

Skyscrapers, fast and slow

You don’t just see this smug complacency, this satisfaction with the status quo and the unwillingness to build, in the pandemic, or in healthcare generally. You see it throughout Western life, and specifically throughout American life.

You see it in housing and the physical footprint of our cities. We can’t build nearly enough housing in our cities with surging economic potential — which results in crazily skyrocketing housing prices in places like San Francisco, making it nearly impossible for regular people to move in and take the jobs of the future. We also can’t build the cities themselves anymore. When the producers of HBO’s “Westworld” wanted to portray the American city of the future, they didn’t film in Seattle or Los Angeles or Austin — they went to Singapore. We should have gleaming skyscrapers and spectacular living environments in all our best cities at levels way beyond what we have now; where are they?

I wouldn't say it is smug complacency. There is some complacency, to be sure, though I don't know how smug it is. There is no demand for smartphones unless Apple shows up; or there is no demand for fast email apps until Superhuman shows up. Once you see 90 Hz refresh rates on screens you didn't know you wanted it. There is no demand for the future until the future is here.

In housing it's not complacency either. Take San Francisco. Why isn't San Francisco outcompeting Singapore at skypscraper building? It's illegal to build. Okay, but why? After all San franciscoans do want more housing, it's not just the tech bros who rant on twitter about it. Once it is possible to build those gleaming skyscrapers, the US does not do bad compared to other countries, the need for national flagellation here is rather moderate. So here it's not that the real state developers are lazy and complacent, it's that the political structures of, in this case, San Francisco, are captured by a politically organized minority who does not want density and skyscrapers. That's not a problem of not building, but of not allowing building in the first place: politics and regulation. When politics is less of an issue you get nice neighbourhoods as a service. In fact looking more broadly the list of the world's tallest buildings, sorted by year has a respectable number of entries for the US in 2019 and 2020. What one can then say though is that the US skyscrapers look bland and uninspiring compared to their Asian counterparts. Why is that? The US designs are probably cheaper to build, but they don't provide the same aesthetic experience. There is no demand for cool buildings until they are in your city but by then it is too late.

At first a reaction I can imagine to the piece may be that focus on skyscrapers may be misguided; while they have some intrinsic appeal as unique works of art - a reason skyscrapers are so expensive is that each one tends to be an artisanal, one of a kind piece, not a mass-market product - they are demanded because they enable economies of scale in cities but in turn those exist because we are not pushing hard enough for remote work. The right ideal would thus be not skyscrapers but nice walkable cities and work from home, which while they appeal less to the futuristically inclined, is what people actually would prefer. However that critique would be misplaced: With skyscrapers and faster transportation, life in the city becomes compatible with life in a spacious house with an ample garden, so you can fluidly decide where you want to live. With progress in transportation technologies tradeoffs become less of an issue.


You see it in education. We have top-end universities, yes, but with the capacity to teach only a microscopic percentage of the 4 million new 18 year olds in the U.S. each year, or the 120 million new 18 year olds in the world each year. Why not educate every 18 year old? Isn’t that the most important thing we can possibly do? Why not build a far larger number of universities, or scale the ones we have way up? The last major innovation in K-12 education was Montessori, which traces back to the 1960s; we’ve been doing education research that’s never reached practical deployment for 50 years since; why not build a lot more great K-12 schools using everything we now know? We know one-to-one tutoring can reliably increase education outcomes by two standard deviations (the Bloom two-sigma effect); we have the internet; why haven’t we built systems to match every young learner with an older tutor to dramatically improve student success?

One answer to why not do all the things we have learned is because very little works. I wrote a longer research essay on Bloom's two sigma and education research in general and the conclusion one gets from it is that researching better ways to learn is hard. Bloom's two sigma is more like Bloom's one sigma2, at best, though there is promise in software-based methods to complement or even replace teachers. It remains a puzzle why large public education systems have not deployed these systems at scale; for them developing software tutoring is a one-off cost - the content of what is taught changes very little -, so presumably this may have to do with either lack of knowledge that these systems are possible, or opposition from within the system itself (again, politics).

The emphasis on more universities is also misplaced; those top end universities are to a large extent accreditation systems for those already skilled. Sure you can scale up Harvard to educate 1000x more people but I doubt that will yield anywhere near 1000x more patents, companies, or research coming out of the expanded set of alumni.

Atoms, not bits

You see it in manufacturing. Contrary to conventional wisdom, American manufacturing output is higher than ever, but why has so much manufacturing been offshored to places with cheaper manual labor? We know how to build highly automated factories. We know the enormous number of higher paying jobs we would create to design and build and operate those factories. We know — and we’re experiencing right now! — the strategic problem of relying on offshore manufacturing of key goods. Why aren’t we building Elon Musk’s “alien dreadnoughts” — giant, gleaming, state of the art factories producing every conceivable kind of product, at the highest possible quality and lowest possible cost — all throughout our country?

A reason is that even when it's possible to build highly automated factories, so can the Taiwanese in Vietnam, leveraging the cheaper workforce too; software is eating the world, not just the US, and technology diffuses across borders. Plus you don't always want more automation, relative to a given level of technological development; right now it is more efficient to employ human workers in many industrial tasks, hence the cost of the workforce continues to play a role in location decisions. And then again; the costs and benefits of forcing reshoring of manufacturing are there; suppose you have a national pandemic stockpile, ready to be deployed when needed. That buys you enough time for international supply chains to sort themselves out, without having to own the manufacturing capability the entire time. 3

You see it in transportation. Where are the supersonic aircraft? Where are the millions of delivery drones? Where are the high speed trains, the soaring monorails, the hyperloops, and yes, the flying cars?

Supersonic aircrafts are slowly getting here. First by means of business-class expensive Boom Aero planes and hopefully in the future SABRE powered hypersonic economy class airliners. The same can be said about everything else. The hyperloops and the flying cars are being built, and not in Asia, if one specifically worries about the West. More than a unqualified call for building; as I see it there is more of a need for building faster.

In fact, I think building is how we reboot the American dream. The things we build in huge quantities, like computers and TVs, drop rapidly in price. The things we don’t, like housing, schools, and hospitals, skyrocket in price. What’s the American dream? The opportunity to have a home of your own, and a family you can provide for. We need to break the rapidly escalating price curves for housing, education, and healthcare, to make sure that every American can realize the dream, and the only way to do that is to build.

This is a key insight, and a point I make in my skyscrapers essay but one shouldn't be that optimistic: It is true that one can bring down the cost and time requirements of manufacturing by means of mass-manufacturing, and this can be applied to housing too but if you also want the designs to be gleaming and unique, forget about returns to scale and learning by doing. The other items are equally problematic; building the physical spaces themselves is not the problem (Unless it is a hospital); the problem is the affordable provision of high quality services (education, and healthcare). The US has probably the best healthcare system in the world by risk-adjusted outcomes (ie. if you have cancer you have better odds of recovering in the US than in the UK), and the top universities are equally the US. It is of course a notoriously unaffordable system. Regardless of what plan one has to solve it is not going to have software at its core but politics. Can one build an affordable healthcare system in the US given the current incentives and regulatory landscape? Probably not, though the better if I'm wrong. Can you just even open a hospital and provide super cheap healthcare?

The P-word

I expect this essay to be the target of criticism. Here’s a modest proposal to my critics. Instead of attacking my ideas of what to build, conceive your own! What do you think we should build? There’s an excellent chance I’ll agree with you.

While I don't intend this to be a critique of Andreessen's essay; in fact I agree with its spirit; I think the more interesting question is Why aren't thing being built? Why aren't things being built faster than they are? What are the mechanistic forces at play in each sector?. My thesis is that the mentality of improvement - to put it in Anton Howes' terms- has not been lost; rather the frictions in going from zero to curing cancer have increased.

But what should be built. Look at any organization that has been around for a long time. Is it a law of nature that, as they age, organizations degrade in function and stop making the most efficient use of resources, making their replacement a necessity4? Take the entire world's space launch sector. For decades we launched rockets and threw them away; sure there were proof of concepts that showed that avoiding this engineering form of economic masochism was achievable but the idea of reusing rockets didn't really became a thing until SpaceX came in. It wasn't the insiders in the industry, or it wasn't anyone in particular demanding reusable rockets, it was an unthinked possibility -like the smartphone- until some random techbro made it real. What's like the space launch industry? Most airplanes are made by Boeing and Airbus and it doesn't look like they are getting replaced. Most major drug companies are over a century old.

There was some recent discussion about an article that called for cutting off funding to the WHO. My response is that maybe we need to get rid of it and make another one. That sentence sounds unthinkable. It's THE WORLD's Health Organization, right in the title. But it doesn't have to be that specific organization providing worldwide coordination for health. The same is true for everything else, just because it has been there forever doesn't mean it will or should stay there forever, let alone the reason for it being there is that is is proper, good, and efficient.

Which brings us to the politics issue. Naturally if one buys the idea that just like Javascript frameworks organizations need constant replacement one can't help but notice that governments are not disrupted at all, save the occasional revolution or war which are virtually nonexistant in civilised areas of the globe. So it's no surprise that frustration with and refusal to engage in, politics, seems part of Silicon Valley's ethos, which leads to attempts to seek solutions that work around it like seasteading or charter cities (Hopefully those will exemplify of what's possible and spark change in larger polities!). But just like an individual philantrophist funding scientist on their own is less effective than affecting the way the NIH allocates its own vastly greater funding, working around politics is less effective than playing the game. That said as a young smart individual what will you do: Rest and vest at Google, or play the slow moving lowly paid think tank game? It's a no brainer. But it's also a prisoners dilemma: You and your similarly situated techbros will end up poorer long term. ¯\(ツ)

What better regulation could look like

If you live in Europe5 you have probably seen the CE logo . But do you know what it takes to get it on a product? One might think that it involves a lengthy process involving sending a government agency samples of the product that have to conform to absurdly exact requirements about how rather than what6. But not necessarily, it involves in the simplest case just asserting that you conform to a published standard (Or you can make up your own standard if you are willing to defend that it is as safe as the official one), slapping the marking on your product, and that's it. Afterwards, governments will be sampling the market to see if the products actually conform and if they don't they will force a costly recall. This mechanism couples both low friction (You can just sell if you want) with incentives to conform to the standards (If you get it wrong, you do pay a hefty price). This is all good for riskless products, but even for more complex ones, involving, say, medical equipment equipment, you can have a private certification agency known as a notified body -among various competing ones- do the testing and assessment for you. In effects this introduces regulatory competition instead of having a single regulatory agency. While this system does not apply in the automotive, aerospace or pharmaceutical sectors -but it does for medical devices in the EU-, imagine if you had, in the US, FDA competitors (regulated by a Meta-FDA); one could for example scrap the Phase I/II/III trial system and have relaxed requirements on a case-by-case basis. Any proposal on accelerating drug-to-market could be tried. One would have to design the meta-regulations carefully so as to line up incentives to balance safety and innovation; for example you could make these regulators liable if they certify an unsafe drug, while forcing them to have an insurance policy in proportion to the volume of drugs that gets certified. You could start a predictions-market-based agency that uses the market to judge the effectiveness of what is considering for approval, you name it.

Hence what urgently needs building is systems (institutions, regulations) that enable and encourage builders. The quote goes "Build and they will come", but we may say "Regulate wisely and they will build" as well.

Appropriately, one of the Collison questions is

How do you ensure an adequate replacement rate in systems that have no natural way to die?

Schumpeter claimed that the problem of capitalism is not how capital is allocated to existing structures but how structures are created and destroyed. Systems and institutions inevitably get stale or become less effective but also work very hard to survive. (See also: the institutional imperative, Berliner's commentary on the "invisible foot", and Herbert Kaufman's short book.) Bankruptcy takes care of senescent businesses. But how do we get a sufficient replacement rate in systems and institutions that aren't naturally subject to extinction processes?


The linked reports also says *On the other hand, start-ups whose founders did not have previous FS experience – for example those whose leaders were technology experts – had to invest significant time and money in securing external legal and compliance support. This shows that, if a FinTech start-up is considering performing regulated activities, making an appropriate investment in compliance and legal capability should be a key component of its overall strategy.


Thus you end up with company-provided food, snacks, gyms and the like conveniently located near your workplace. At least if you work for a big company.


It may get to two sigma for disadvantaged kids; and also it seemed to be less than one for very bright kids. So while better education methods can get you some gains for the average kid, the case is more difficult for those deep into the right tail of the cognitive ability distribution. It is an experiment worth trying, all things considered. What happens when you use high quality software in a large sample of students?


Admittedly this is part of a larger debate I don't do justice to here


Unless they have what value investors call moats, competitive advantages that can keep a business afloat even with incompetent management - to some extent of course-, like network effects, or high entry costs.


Like the longstanding prohibition in the US of overland supersonic flight instead of a more sensible prohibition of flying if the aircraft generates noise above a set threshold, which would automatically enable quiet supersonic planes as soon as they enter in service.


In the US there is a more limited arrangement in the form of the FCC declaration of conformity and the concept of the Nationally Recognized Testing Laboratory (Like Underwriter Labs); it is more limited because the former covers only electromagnetic testing and the latter covers workplace equipment. The EU system covers many more things under a single umbrella and allows for self-certification in some cases