Hacker News new | past | comments | ask | show | jobs | submit login
Why does so much government tech investment deliver so little? (airstreet.com)
60 points by sebg on Nov 10, 2023 | hide | past | favorite | 102 comments



Government tech investment delivered the Internet, the single largest advancement to humanity, even bigger than the moon landing.

It's also like he doesn't know what investments are: a gamble that you think the proposed company/solution/whatever is going to succeed, and that the few successes will be far more valuable than what was spent on all the investments, successes and failures combined.

This is like if I wrote an unhinged rant about how Y Combinator wastes money and is stupid because most of the picks have turned out to be shit choices, and ignoring the fact that the few that survived, found their market fit, and succeed were highly profitable and helped fund future rounds to continue the process.


These famous US government investments are more precisely described as military investments, which they were. Almost all effective government investment was closely connected to a specific military objective, and measured against delivery on that objective. The US has been unusually effective at keeping these projects from turning into pure jobs programs, which is the default outcome in virtually every country.

It is an interesting question as to why the US defense institutions are so consistently effective at producing advanced R&D when that is not the outcome of almost any government program in the world, including most other parts of the US government. Even most modern trauma medicine was pioneered by the US military. These programs usually manufacture sinecures in most countries. The US has somehow, in this one narrow area, largely avoided that trap. It would be interesting to know if it is cultural and/or an artifact of how the programs are structured.


It feels like the flow of useful research out of US military spending has slowed down considerably. You notice people are still pointing at the internet here.

The main advantage the US has had historically is just a much larger budget. If you can throw ten times as many darts, you'll hit the bulls eye ten times as often.


The Mansfield amendment in the 70s banned military funding for research that was not tied to specific military stuff, that might be a factor here


DARPA still has a budget though, even if it is much smaller than NSF's. And there is still a lot of DoD funded research.


Almost all government R&D spending is via small(ish) grants to academia for entirely unfocused "blue sky" research. Once you accept the premise that research should have no specific application in mind, it becomes impossible to determine if it's actually doing anything legitimae or is just fraud, which is why there's so much non-replicable crap coming out of universities.

Military R&D is done with a specific, achievable goal in mind. It's not blue sky, it's more like "we need a better missile, how do we do that". Most of the real impactful innovations out there came as a side effect of trying to achieve some other goal, with only rare exceptions (ChatGPT?).

As for why other countries don't, their military spending is too small to catch up and then exceed the state of the art. All the money gets spent on buying stuff the US developed.


For projects I am familiar with, usually you have to propose actual deliverables, and there are project reviews and reports that have to be delivered to program managers.

But given the nature of research, it is kind of unpredictable. One trick is to propose research that you've already completed or at least gotten into a prototype phase.


My brother in Christ, do you not think the government invests in any sort of infrastructure or medical R&D? This was a single infrastructure project that pioneered new techniques like ground freezing to let tunnels be dug bigger than before or through less stable soil[1] or any sort of grant money given to university researchers on various drugs that then get spun out into products or brand new companies.

For all its faults the US government is still the greatest producer of scientific research and advancements as far as we know in the history of the universe and that includes far more than just military related projects

[1] https://appel.nasa.gov/2020/07/23/the-big-dig-learning-from-...


I think “investment” in regards to government is meant less as “wild bet that will get us all rich” and more of “expenditure that will save us money 5 years down the road”.

Conflating the two is just disingebuous, and provides cover for all sort of unaccountable boondoggles like healthcare.gov (initial launch, not the hastily rebuilt version) and California EDD’s latest redesign apparently costing a shiny $1.2 bln https://calmatters.org/economy/2023/11/california-unemployme...


There's a big difference between government contract work that results in lackluster results and cost overruns, and research grants.

I think the article is focusing on the former, but the title is admittedly vague.


The internet, GPS, and NASA in general are gov-tech home runs for the ages. But the first two were military projects, and NASA was meant for winning the space race part of the cold war.

I wonder if they went well because there was a clear path forward in each case. When the local town puts up a site, they are probably not as inspired by a clear vision.


Don't leave out NOAA/NWS which provides radar and satellite coverage for most of the country and supercomputer modeling to deliver predictictive results unimaginable 100 years ago.


The internet was only one of many telecommunications digital protocols. It was inevitable that computers would be connected together, with one protocol or another.


> the single largest advancement to humanity, even bigger than the moon landing

Such an incredibly American thing to say, with a healthy portion of recency bias sprinkled on top.

This is only beaten by that one time someone online referred to the HMS Titanic as USS Titanic. Because it was groundbreaking for its time, it must have been US American, of course.


Claiming the Titanic would seem to be a peculiar kind of exceptionalism, assuming it wasn't a joke or a typo.

And wasn't the web (which is how people commonly experience the internet) invented by Tim Berners-Lee at CERN?


There was another government that tried to create an Internet: France's Minitel. It eventually failed miserably in front of the assault of private sector pushing the US Internet, even if it seemed as a good thing for a while.


I don't know a lot about Minitel, but it sounds like it was impressively useful years before the web.

In the US I'm sure many AOL users didn't think they needed the web. Apple even collaborated with AOL to create their own online service, eWorld, just as the web was taking off (they shut it down after a couple of years and migrated users to AOL.)

There is an interesting effect where early technology adopters get left behind once a replacement technology takes off and goes mainstream.

I'm sure many people who had Blackberry, Palm or i-mode devices didn't think they needed the iPhone.


>Government tech investment delivered the Internet

Hmm ... nah. The internet would have happened anyway. Also, we owe its growth to plenty of ISPs everywhere and quite possibly ... piracy and porn.

Even if it was true and the government single handedly created the whole www, that shouldn't be a carte blanche to just waste money that could otherwise be invested in a better way.

Also your bit about YC, YC is private money, they could throw all of it into the ocean and I wouldn't bat an eye. The government, however, takes almost half of what I make, for sure I'd like to know, with unprecedented detail, where and why my money is going to.

You couldn't be more wrong if you tried to, lol.


> The internet would have happened anyway.

Maybe "mega-AOL" might've happened, but an open network of networks where there's no parasitic mega-corp acting as gatekeeper and tollbooth operator? Citation needed.


The bulletin board networks.

People with two or more computers were always finding ways to connect them before the internet. Networking protocols were a dime a dozen in the 70s and 80s. The idea that people would never hit on the idea of a protocol over telephone wires is not particularly compelling.

It's like saying if the Wrights had not invented controlled, powered flight in 1903 we'd still not have airplanes. (It would probably have been accomplished by other people by 1908 or so, 1910 at the latest.)

The internet was just one step in a long line of digital telecommunications networks, starting with telegraphy. See "The Victorian Internet" by Tom Standage

https://www.amazon.com/Victorian-Internet-Remarkable-Ninetee...


AOL was made far after that and still exploded in popularity. Facebook today is approaching the level of mega AOL.

There is a constant pull between open and closed systems. Nothing the government did ensured networks would be connected together. TCP/IP is such a low level building block it’s easy to create closed networks with it.


Happened anyway? You have any idea of the value that was created by having it happen many years earlier? Modern integrated computers let alone the Internet were brought into being because of defense investment by the government. These things rarely happen entirely organically in free markets. You think SpaceX would exist without heavy early investment by NASA, or later money from the commercial crew program? Unlikely.


>Happened anyway?

Absolutely, zero doubt about it!

>You have any idea of the value that was created by having it happen many years earlier?

There's no early or late, it happened when it happened.

Also www means "world wide web" which, by definition, implies a collaborative effort of the whole ... world? Sure, props to the US for lighting the first spark, but after that everyone did their part, even us, its users.


Yes, an effort of the whole world's public institutions. Tim Berners-Lee didn't work for the CERN Corporation when he set up the Web.


Tim Berners-Lee’s contribution was not novel. Hyperlinking things had many parallel competing implementations.

Had he waited a couple of months we would have essentially the exact same thing with a different name.

TCP/IP and lower level protocols that connected everything together is really the boon. The protocols for how we exchange data on top were inevitable.


TCP/IP and protocols which originated from ARPANET?


There is a looong way from a University Communication protocol to my ISP giving me gigabit connection for 50 bucks a month to connect to Pornhub.

Private sector delivered the Internet. Academia researched and established standards. What the US government did was having the foresight to understand this need and grant research funds and contracts to the sectors able to deliver.

It's not a negligible contribution by any means but it's pretty small in the grand scheme of things.


Yes, tcp/ip was a foundational protocol that came from DoD money. My point is that CERN’s contribution was negligible to the Internet, not that government investment is irrelevant.


Some effort from everybody in the world, even us, its users.

I'd estimate the "GDP" and traffic related to public institutions as a single-digit percentage of the whole web.


Private interest created the use cases and further innovation but public investment, as they say, took the initiative in creating the Internet.


Private interest created the internet you use today.

It’s like saying Pythagoras created all of modern math, including calculus.


No one here is claiming the Internet was created by one single party or another. Taking the initiative in creating something is a different thing from creating all of it.


Piracy is routinely found to be illegal when it meets the judicial system, and the distributors of porn profited greatly from the distribution system before the internet. Why would they be the drivers of this?

edit: replied to the wrong comment, but it's still relevant to the assertion that the internet would have been invented without public funding


Funny how much these invisible hand types rely on unfalsifiable claims to make their point. “Would have happened anyway, 100% positive of this”. Oh for sure dude.

The burden of proof always seems to fall on people who live in reality.


"happened anyway" is always logic I see applied to government made advancements and not to private ones.

A good read on the topic is "The Entrepreneurial State" by M. Mazzucato.


Not always, not the case with me.


Would you say the Google search engine or the iPhone have happened anyway?


Definitely!

Google was just one of many search engines when it was created, arguably the best of its time, but still not the inventors of the search engine.

Same thing with Apple and the iPhone.


Sounds... strange.

> that's not a carte blanche to just waste money that could otherwise be invested in a better way.

What are some of the better ways to approach investment by the government today, compared to what we have?


In-Q-Tel seems to be doing well from the outside. Curious what people think. I’m also confused why the EU pursues broad megaprojects instead of adopting the model that produced their antecedents. Is there some historical reason?


Let the people discover what to invest in themselves.


I'd rather have a panel of experts in the field with clear criteria decide the merit of the grant vs a crowdfunding-like model where people vote for the most charismatic person with the most well-produced video.


What clear criteria do they use to decide the merits of grants?

> the most charismatic person with the most well-produced video.

Ability to actually communicate your idea and justification for being funded, effectively and in a professional manner, is something notably missing from many academic projects. It's not exactly zero signal.


Those “panels of experts” tend to be interested in what they are already trained in. An expert is the last person you want to push for new ideas.


> Let the people discover what to invest in themselves.

and then the people will be asking for faster horses.


Did the government invent automobiles?


> Government tech investment delivered the Internet,

Don't forget about microchips.


>> Government tech investment delivered the Internet,

> Don't forget about microchips.

The Romans are all bastards, they have bled us 'till we're white, they've taken everything we've got as if it was their right, and we've got nothing in return though they make so much fuss, what have the Romans ever done for us?

https://youtu.be/Qc7HmhrgTuQ


Why are all of these examples of successful government tech investment so old? ARPANET, moon landing etc. are all 50+ years old - before most HN readers were even born. How many billions/trillions have been spent since then without achieving anything nearly as spectacular?


The internet didn’t become the central piece of society and commerce until around 20 years ago, so partially it is explained by needing time to mature and adopt.

Investments from the 90s onward have been more public-private partnerships, but things like the human genome project or a number of cancer treatments and pharmaceuticals have been developed in this manner.

Third, I’m not sure I would regularly expect something as big as the internet to be invented every decade.

Not necessarily disagreeing with you because I feel similar to your comment, but I wanted to dig in a little to see if it’s more truthful or more biased.


By 2000 the military involvement in the internet was long since history. If you want to talk about the time when government investment was what mattered you have to go back ~35 years or more.


My point was that large investments, especially in infrastructure, need decades of time to bake. So investments made this century could just be cooking.

I also suspect the push by the uS right wing have impacted the number of great government innovations. The federal role nowadays seems to be to do the core research that is not directly profitable, then transfer control to companies to reap the rewards. (Because of the often incorrect “the government can’t do anything right” narrative pushed in the states)


This really seems to have worked out kind of OK on one hand, although sometimes I wish that the public had more ownership of the technology that we funded.


Because we've been gutting scientific grant spending since like 2005, leading to a cutthroat environment where for the most part only the most well-connected and well known get the big grants?


Investing in something not successful is not specific to governments though.

You can find recent examples too. Also take into account the survivor bias.

The latest technology that got appreciated on HN I could think of it the Mistral LLM, trained on a government funded EuroHPC. It’s a few weeks old.


> Why are all of these examples of successful government tech investment so old?

There are indeed newer examples such as GPS, RF-CMOS, Lidar, Siri, mRNA vaccines that DARPA invested in one way or another. Not saying they are as spectacular as the internet.


Not to mention that perhaps the low-hanging fruits are all reaped in the earlier years.

The newer tech might just not be as impactful as the yester-decade's, because those impact is already felt, and creating more impact in the face of today is getting harder and harder. Exponential growth is not possible to achieve in practise.


NSF has funded all sorts of major advancements and inventions we depend on: cloud computing, multicore processors, wireless networks, fiber optics, computer graphics, bar codes, MRI, PCR, CRISPR, improved LASIK, doppler radar, modern internet and web, cryptography, AI/ML, etc.

Also: Google (pageRank, data centers), Nvidia (CUDA), VMware (virtualization), Qualcomm (CDMA) - many tech companies have leveraged research (and people) funded by NSF.

On the NIH funded/medical side we can look at things like mRNA vaccines, which may be transformative in fighting many diseases.

On the DOE-funded side we seem to be seeing major advancements in nuclear fusion.

On the DARPA side modern self-driving cars (and associated businesses) seem to derive from DARPA's grand challenge in the 2000s.


Rna vaccines might be that but it is still too early to tell.


NSF research funding has paid huge dividends in tech.

Poorly paid graduate students are the secret of US research productivity, and they also graduate and work in companies to produce all kinds of technology.


military and government in the US should be distinguished because it makes a stark and pointed difference in results


Having been through this two major reasons.

they never talk to users. Seriously it’s crazy - you think it’s bad in industry - govt is wild. A bunch of folks 10 levels away decide on things. They will literally never even do one click in a demo - it’s all paperwork based - so when in 10 seconds users can tell it’s hot garbage (ultra slow, ancient Java version required and all) it’s too late.

Sane developers are smart enough not to contract with the government. McBride principles compliance? Virgin redwood compliance? Race info on all staff. Ethnicity for all staff? I once escalated to city attorney a compliance requirement that we hand out an IRS form for a program that no longer existed (advance earned tax credits I think) and they said form still had to be handed out (because some political body had required it be) but we could tell employees who filled out the form it was for a program that was no longer in operation and we would not process the form. Multiply by 20. So you get compliance and other middlemen who exacerbate the first issue - the folks developing are super removed from users.


It's fascinating how Americans assume 1.) their experience in America is exact the same true for whole world. 2.) everyone in the word understands specific American terms.


The answer to questions of the type "Why are governments so bad/inefficient/wasteful/slow/etc.?" is always the same: Because they have no incentive to do better.

When you are literally the one who makes the rules by which the game is played, it's insane for others to expect you to make a genuine effort. Apart from the (infrequent) elections, which provide at best a very weak control mechanism, governments are free to do as they please, with almost zero external pressures, while also being able to tell all other players what to do. The results are completely predictable, and indeed unavoidable unless the framework were altered radically.


>with almost zero external pressures

a lot of government action happens in the context of competition between nation states with stakes that can be much more intense than in the private market. That's how you got the Meiji Restoration, The Prussian education system that's now everywhere, lots of fancy tech during the Cold War, and much of China's gains during the last 30 years.

Much off the innovation in the US is government driven, military technological superiority. The Silicon Valley far from being some sort of private market Mecca has its roots in the defense sector.


When a government is a regular client (even if one with a very fat wallet and lots of requirements), it is the private free market that researches and develops the solutions and products, not the client.


We also don't allow your government to develop competencies. All those examples of government failures in tech are failures of private sector consultants because we're allergic to allowing the government to develop in-house expertise.


Nothing stops the government from developing competencies. The problem is we systematically don’t reward competence or punish incompetence.

There are many bright people in government but they spin their wheels waiting in line for purely seniority based promotions and pay raises. Then they eventually report to an idiot who doesn’t want them rocking the boat and they die inside or leave.


If you're paid better working in the private sector selling your experience back that sounds like a big factor in why we wouldn't keep experience in house.


Here in Norway our IRS has long had very good web pages, including ability to edit you tax return (pre-filled), unlike most other agencies.

Turns out they do most development in-house. Most other use consultants, both for project planning, management and development. Go figure.


Why hasn't Soviet technology dominated? There were no private sector consultants there mucking up the works.


For one, not doing everything through a web of public private partnerships isn't the same as central planning.


It's more than that, it's the cover-your-ass mentality that develops when every small part of your job is dictated by reams and reams of bureaucracy and policy. Building successful products requires direct communication with users and organizational tolerance for making mistakes. Government replaces direct communication with requirements documents, and tolerance with "accountability", lest taxpayer funds be "wasted".


I think the view is a bit myopic. Certainly, government projects have inertia & cost overrun. But over the last six decades, here are some of the initiatives on my mind, which couldn't have been done without government investment in technology:

* Internet, as many people pointed out already.

* Nuclear energy & commitments to nuclear projects e.g. nuclear medicine, nuclear chemistry, medical research, particle physics & major instrumentation grants

* Space technology. Ability to understand the cosmos, origins and the unanswered questions about the eventual fate of our world & mankind. Including some flattering moments such as Apollo program.

* Meteorology & mapping: This is usually the underdog. We owe a lot to meteorological studies and USGS. Humanity (not just the US) has managed to stave off famines, several natural disasters or be prepared to face the next simply because how well we have mapped earth and it's resources by remote sensing technologies, many of which are government developed or sponsored in large.

I think what is rather true is government projects are endurance races. They aren't as flashy as SpaceX or ChatGPT, but without them the next glitzy tech will be arduous to get. As with all races, one can't win everything & the slow nature of it makes us question its worth sometimes.


Also, Fauci helped fund gain of function experiments in Wuhan labs. Those had to be done by government because a private person would catch criminal charges.

https://nypost.com/2021/10/21/faucis-agency-admits-it-funded...


Disagree on Internet. The demand for fast information exchange was already apparent with phone networks, telegraphs, and the mail.

Once we had general computing, there was going to be something to fill the void if DARPA didn’t fund TCP/IP. We would probably be dealing with AT&T addressing schemes instead…


There was going to be something to fill the void, the questions there are "when?" and "for how much?", no private company would deploy novel inter-networks communication for cheap since it was an unproven technology. No private company would want to make it open for competitors to use the network, monopolistic tendencies don't naturally go away in the private sector, much the opposite.


SpaceX development was a PPP


When I want to look at my localhost and do "host ::1" or "dig -x ::1", my machine spits back something about ARPA. What is ARPA?

It's how the US government paid for decades for the R&D to create the Internet. Just like how the Air Force paid Fairchild to make semiconductors. Or the CIA paid Oracle to create relational databases.

Technology was created in the US via the funding of the YS taxpayer, sent to the Pentagon and then sloshing around. Years or decades later it is ready to be a product and given over to private profit.



This article is higher education research erasure. The private sector gets absurd amounts of benefits from university research technology transfer. That's government tech investment that's handed to private companies on a silver platter.


The EU research frameworks are first and foremost aimed at building cross-national relations, second at industry subsidising under an accepted pretense, and third at forging academic industry transfers. Getting actually market results from the stated research objective is a far fourth.

Think about it this way: if you have a true strategic research objective as a company, would your approach be 'I am going to share this idea with in a not public but not private either research proposal after I have built out a an international consortium and parcelled out different workitems for the realization of this idea to a host of 'partners' whos stratgic innovation ideas allign with my own, the hope for a 5% chance of receiving approval and a partial funding grant 2 years after I got the idea appoved internall, that imposes an extremely heavy reporting and management process on me to monitor and report on a series of prenegotiated 'deliverables' and evaluations over the next 3 years after which we can maybe exploit this resulting shared IP?

The unfortunate reality is that as a result most research departments, both industry and academic, spent nearly all their effort on grant aquisition and grant management, and only what is left goes to producing contractual agreements and passing reviews with no further leeway for actual research

There are exceptions. I would fully support a strategic EU tech 'decollonization' effort as it is badly needed, but the current programs are never going to succeed in that as they are not set up for this, and getting a true successful effort of the ground would probably meet with substatial political detraction.

P.S. Despite the above I have loved working in the EU research space. The International work environment is truely enriching, and most people working in this arena are realy bright and at the top of their game.


I think most industry research departments don't care about EU or government grants at all, unless you work in a very different industry to tech.


And yet tech is the most oversubscribed program of all. There are 20x to 30x the number of grant applications to the budget.


This seems like barely-massaged LLM content and/or just a bad marketing page.


I suspect far more private investment delivers little or nothing. Meta’s billions wasted on “the metaverse” seems on the same scale, if not more, than government tech R&D spending. WeWork, Most of Musk’s projects, Theranos, FTX and almost all of crypto/blockchain investment (speculation) delivered no innovation other than novel crimes.


Because tech is really process orchestration and automation

Most gov offices have individuals who run processes with no concept of why they are getting what they are getting, why they are changing it, and where it is going

Compartmentalization is great for blame management, but terrible for efficiency


> However, government attempts to spur business to invest in bad ideas distort existing markets, they lead to potentially productive investment being channeled into unproductive ends.

> Europe has incredible potential, considering its political stability, world-leading universities, and existing talent. But it cannot and should not throw centuries-old economic truths out of the window to conjure up a simulacrum of technological sovereignty. Superpower status will not come via death by a thousand grants or attempting to recreate Silicon Valley in Strasbourg. It will come from what markets do best: discipline, picking battles wisely, and allowing bad projects to fail.


One reason could be that, governments tend to invest in sectors that are not jet profitable. The moment a research area looks as if it is ripe for commercialization, private entities join and for various reasons are much faster.

Good examples are quantum computing and fusion. Both were purely government funded, now that the fields approach usability we see private companies enter.


Anyone who have experience working inside the Goverment and working in a small company / startup [1] will know. The objective of the tech investment were not to deliver. But to show they have done something.

[1] I avoid using the term private sector because lots of big companies have similar attitude.


Anything today with government has same characteristics, not only tech. The very simple reason is that just ask yourself have you ever even know how these investments are made, and have you ever heard any of these opportunities if you are just a normal guy.


This article feels very specific to Europe. I think the US does a really good job of funding tech innovation, especially given its size and how complex it operates.


Survivorship bias.

If a tech company (especially a small one) sucks, it will disappear rather quickly. So you only see the ones that made it work.

If the government IT sucks, it won't disappear.


The government is a very poor allocator of speculative investments. The people who select what gets invested in do not have their own money at stake, so they are able to choose based on priorities other than profit maximization.

The government should stick to enabling fair play in the private sector and stay away with gambling taxpayer dollars.


You are sending this message on the descendant of a government speculative investment…


Government invented it, private sector made it legendary.


Because taxpayers don’t hold govt officials accountable. We’re all too busy working.


This article is all over the place. Large organizations, corporate or government or other, eventually ossify and fail to adapt. Some find ways to reinvent themselves. But the framework is bizarre. Under the model that tech investment is taking a bet, these values are in contention. You simply don't hear about startups that did not pan out and ate all the capital invested. And these truths are not centuries old, we reinvent our view of capitalism every few decades. The defense industry in most western countries is a good example of this where it's still very Keynesian.


> You simply don't hear about startups that did not pan out and ate all the capital invested.

I recall reading about a big flush just a couple weeks ago.

"Convoy, valuing it at $3.8 billion just 18 months ago."

https://www.cnbc.com/2023/10/19/bezos-backed-freight-firm-co...


Becuase software development is hard and most people, even those who call themselves software engineers, cannot do it


software development is not hard at all. the reason everyone thinks it is hard is that there are too many software developers and 99% of them should have picked a different profession


if 99% of people cannot do something, then it is considered hard




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: