Hacker News Reader

I disagree. The reason why the most innovation has occurred in this sector is because we're some of the best-funded. Look at the arms industry and you will see a similar pace of advancement - mostly due to the bottomless and nigh-unchallengeable defense spending of Uncle Sam.

Why this unprecedented level of funding? IMO this really is the heart of Thiel's argument: internet companies are capable of general obscene, absurd, ludicrous profit margins.

Think about a company like Facebook - employs a few thousand engineers, but generates revenue and impact far in excess of what you might expect from a crowd of that size.

And this is really Thiel's argument, the way I understand it. He's not saying that the iPhone isn't a "technological" breakthrough, but rather it's not a "societal" breakthrough. In the old days, when someone introduces a product, it creates new jobs across the board.

When Ford introduced a new car, it didn't just mean jobs for the people in the boardroom and drafting tables. It created jobs in their factories, and indeed tens of thousands of jobs across a wide supply chain necessary to build the product. It created new jobs in dealerships, and auto shops. The job-creation effects of a product were incredibly far-reaching.

Now, fast forward to today. What jobs are Facebook creating besides the ones inside their walls? There is no employment ecosystem, no supply chain. And in this way, while Facebook is extracting record revenue from their business, the rest of America is no better off.

It used to be that success in capitalism is the tide that floats all boats, but this is increasingly not true, especially in our sector.

  Look at the arms industry and you will see a similar pace 
  of advancement - mostly due to the bottomless and nigh-
  unchallengeable defense spending of Uncle Sam.

But I don't think you're really disagreeing here.

The DoD is a lot less squeamish than civilians about experimental projects that result in death. Leave aside the issue of whether this is good or bad. For them, they are going to be seeing dead bodies as a natural course of doing business, and they want fewer ones, and so they're (much) more willing to pursue riskier projects that may cost lives.

Soldiers can be ordered to run into enemy fire, so they can also be ordered to test drive automatic cars (funded by DARPA). Autonomous helicopters that could slice someone in two if the programming goes bad nevertheless get the green light because there are already people being sliced in two, and DoD wants to reduce that number.

DoD is really pushing robotics and cybernetics forward in particular because those two result in less dead soldiers and (in the future) soldiers who can function even after losing limbs and eyes[0].

Perhaps the most concrete example of this thesis is what happens when the priorities of DoD and FDA collide. Take a look at the Pathways to Innovation program that FDA just greenlit[1].

  On a Tuesday afternoon call with reporters, Shuren said he 
  expects just one or two devices to be approved through the 
  new pathway each year.

  The first device up for review under the new pathway comes 
  from the Defense Advanced Research Projects Agency 
  (DARPA), the research arm of the Department of Defense.

  The brain-controlled prosthetic is designed to restore 
  near-natural arm, hand, and finger function to patients 
  suffering from spinal cord injury, stroke, or amputation. 
  An implanted microchip on the surface of the brain records 
  neuronal activity, decodes the signals, and activates 
  motor neurons that control the prosthesis.

  The first chip will be implanted in a patient within six 
  months, Geoffrey Ling, MD, PhD, program manager at DARPA, 
  told reporters.

There is a very good reason why no private company has yet gotten this kind of fast track, but the DoD has. From FDA's perspective, DoD is a (more powerful) fellow agency which deserves professional courtesy. Moreover, FDA doesn't suspect DoD of just being in this to make money. Thus a dot mil is allowed to play by fundamentally different rules than a dot com.

This is my point: when lives are on the line, people will be willing to risk lives to innovate. If not, then not.

[0] http://www.youtube.com/watch?v=TW78wbN-WuU

[1] http://www.medpagetoday.com/PublicHealthPolicy/FDAGeneral/24...

It might seem contradictory, but the DoD cares more about it's people than most organisations and at the same time is more than willing to risk lives for progress. Consider the initial proponent / researcher of car seat belts in the US was the Air Force which studied the issue in large part because car crashes where killing more pilots than airplane crashes (at the time). At the same time, they are vary willing to test new airplanes even though test pilots often die. And more than that they where willing to risk a pilot to recreate the conditions of a crash to find out why it happens.

If you run the numbers the DoD has directly saved more US lives since WWII in that same time period. Which is not to say it's worth the budget just more useful than you might think.

How is the iPhone (and FB and Twitter) not a societal breakthrough? Each has developed entire ecosystems around their respective platforms. Hundreds of companies exist and thrive because of these technologies. Their impact on the global economy, and on job-creation, is much larger than the sum of the engineers within their walls.

Sam Peltzman's book "Regulation of Pharmaceutical Innovation" makes a statistical case that the FDA's "safe and effective" mandate has cost more lives than it has saved.

From the back cover: "The 1962 Amendments by Sam Peltzman concludes that the 1962 ("proof of efficacy") amendments to the Food, Drug and Cosmetic Act of 1938 (1) have cut the number of new drugs (new chemical entities) brought to the market each year by half with no corresponding reduction in inefficacious drugs, (2) have doubled the costs of drug development, (3) have have increased drug prices by reducing competition among drugs (so that the sick pay $50 million more each year for drugs), and (4) have imposed substantial net social costs from death and illnesses that could have been prevented if the development of new drugs had not been hampered by the amendments. These conclusions are reached despite the fact that Professor Peltzman has assumed for the amendments benefits that cannot be proven and may not have taken place."

HIV has been cured for Simian monkeys for quite a while because we could invasively experiment on them; so, the FDA stopping us from invasively experimenting (including involuntary innoculation of disease free subjects) on humans may have "statistically" cost more lives than it saved, but you'd have to be a Nazi to advocate it.

That is not the thesis of Peltzman's book. Peltzman compares pharmaceutical innovation before the 1962 FDA amendments with after in the US, not with nazi germany.

Drug trials fail all the time though, and not only because of risk aversion.

If you look at the first 50 years or so of the pharmaceutical industry, there were wonder drugs being discovered every year. We may simply have exhausted a lot of the low hanging small molecule fruit.

Also, if risk aversion was slowing pharma innovation then we'd see a lot of potentially efficacious compounds failing safety trials. Instead what we see is that the cost to discover candidate compounds before they're ever tested on humans.

It's absolutely true that risk aversion is not the only thing holding back drug trials. But it's definitely part of the problem.

Think about insulin. In 1921 Banting and Best had the idea. It was in a patient's arm by 1922. And they won the Nobel Prize by 1923.

  http://www.nobelprize.org/educational/medicine/insulin/discovery-insulin.html

  Early in 1921, Banting took his idea to Professor John 
  Macleod at the University of Toronto, who was a leading 
  figure in the study of diabetes in Canada. Macleod didn't 
  think much of Banting's theories. Despite this, Banting 
  managed to convince him that his idea was worth trying. 
  Macleod gave Banting a laboratory with a minimum of 
  equipment and ten dogs. Banting also got an assistant, a 
  medical student by the name of Charles Best. The 
  experiment was set to start in the summer of 1921.

  ...

  In January 1922 in Toronto, Canada, a 14-year-old boy, 
  Leonard Thompson, was chosen as the first person with 
  diabetes to receive insulin. The test was a success. 
  Leonard, who before the insulin shots was near death, 
  rapidly regained his strength and appetite. The team now 
  expanded their testing to other volunteer diabetics, who 
  reacted just as positively as Leonard to the insulin 
  extract.

  The news of the successful treatment of diabetes with 
  insulin rapidly spread outside of Toronto, and in 1923 the 
  Nobel Committee decided to award Banting and Macleod the 
  Nobel Prize in Physiology or Medicine.

That was before the FDA gained the power of premarket approval, a time when pharma moved at the speed of software. Nowadays that past is characterized as a time of patent medicines and scam artists. But if you actually get into the history books you'll see that description is like characterizing the modern internet as a giant repository of spam and scams. In other words, absolutely that spam/scam stuff existed (and indeed crystals and faith healing persist to the present day), but real innovation in drugs also occurred at the same time.

The fundamental point there is that Banting and Best could do empirical human trials. You can't predict what a novel drug will do without humans who are willing to risk their lives. In computers these people are called early adopters.

  If risk aversion was slowing pharma innovation then we'd 
  see a lot of potentially efficacious compounds failing 
  safety trials. Instead what we see is that the cost to 
  discover candidate compounds before they're ever tested on 
  humans.

Yes, the thing is that drug companies face a significant penalty with the FDA if they fail a safety trial. It's not just a one off, it influences the FDA's perception of them as a "reputable" company. If the FDA thinks of you as disreputable or a cowboy, they will inspect the hell out of you and issue Form 483s [0] till the cows come home. And that's not even including the PR impact and the market sell-off should you ever actually have a TGN 1412 incident[1].

So there is a very significant economic risk to "failing fast" in pharma. Over and above the hit to brand, if the PR fallout is bad enough the FDA has a number of tools with which to criminalize failure. The Park Doctrine[2] is what they do for "off-label marketing", and they would escalate considerably if they felt public opinion demanded it.

Anyway, this is why drug companies are doing all these simulations and iterating primarily around drug molecules that are known to be safe. No one ever got fired for buying IBM, and no one ever got fired for trying molecules compliant with Lipinski's rule of five[3]. Sure, that's still a pretty wide class of molecules, but "radical innovation" in compounds is just not going to fly with the FDA.

After all, it took the FDA more than a decade to begin considering adaptive clinical trials[4] in lieu of the standard Phase I/II/III design. How long will it take them to agree that aging could be treated as a disease, or to allow the development of enhancing drugs?

These things probably won't happen until there's some sort of major conflict in which enhancing drugs become the difference between victory and defeat. Then the floodgates will open.

[0] http://blog.fdazilla.com/2011/09/fda-will-issue-10000-483s-i...

[1] http://en.wikipedia.org/wiki/TGN1412

[2] http://www.gatewayfda.com/fda-regulations/under-park-doctrin...

[3] http://en.wikipedia.org/wiki/Druglikeness

[4] http://jnci.oxfordjournals.org/content/102/16/1217.extract

> Nowadays that past is characterized as a time of patent medicines and scam artists.

What is more pertinent is that it was the time of the Tuskeegee syphilis trials.

I want to upvote your comment ten times, but there's only one of me. Thanks for the point of view, history, and references.

Why, do you think, are the pharma researchers not shifting cutting edge research to places where regulation is more pliable and inventing new treatments there. Surely Latin America or China (maybe even India, depending on bureaucracy) could provide the kind of market and test bed for new treatments that fast innovation would thrive in.

Good question. In theory this is possible, and there are a number of biotechs and medical device companies that are looking overseas.

Europe is the preferred place right now for traditional biotechs & medical device companies, as getting a CE Mark is a lot less onerous and much more predictable[0] than the increasingly random FDA process.

As for China (and India), they are very promising but will require entirely new business models, for a few reasons.

First, the current biotech industry is based around an artificial barrier to entry (namely FDA approval) and artificial scarcity (namely drug patents). Remove the former and you remove the rationale for the latter. This is their whole business model.

Second and more fundamentally, most existing pharma companies have now gone through almost 50 years with the FDA. The "fail fast" model was 2-3 human generations ago, and even really good scientists like Derek Lowe have a bit of Stockholm Syndrome. In the worldview of (most) modern pharma scientists, the FDA process is considered a law of nature, something on par with s and p orbitals.

Removing FDA approval and patents from the situation would suddenly make two of the most important departments in any pharma company (regulatory affairs and IP) completely useless.

A world without the FDA, where doctor, patient, and chemist made their own decisions without third party regulation would be painted as strange and worrisome. Like the TSA, the FDA has dozens of full time PR professionals[1,2] dedicated to convincing you that "your safety is their priority". Unlike the TSA, most people (even most in the industry) don't question whether the FDA is actually protecting people.

Without going into detail, then, successful business models for pharma/biotech/device companies in China/India would be based on validating products in a highly empirical fashion, getting them to market rapidly and protecting them from competitors via execution without any IP enforcement.

Think something more like the supplements market, where creatine, omega 3, and energy drinks rose to the fore over the last few decades...except with even smarter people doing the formulations.

No traditional pharma is culturally capable of doing this. Moreover, even if they were, the FDA would frown on this threat to their review authority. If this hypothetical rogue pharma had any desires of ever selling into the US market or not having their reputation destroyed by the FDA with a few press releases, they wouldn't try setting up this sort of low-regulation phenomenological branch overseas.

So, to really exploit the potential of true drug legalization, it'd have to be new companies.

[0] http://www.eurekalert.org/pub_releases/2011-05/iti-fpa052311...

[1] http://www.fda.gov/NewsEvents/Newsroom/MediaContacts/default...

[2] http://www.fda.gov/AboutFDA/ContactFDA/FindanOfficeorStaffMe...

Thank you for writing an informative reply to my question. It's a lot to think about.

According to Guns, Germs and Steel, necessity is not the mother of invention. In fact the opposite is true: inventions appear when the needs of the inventors are already met.

I really recommend the book, it's a great read.

Jared Diamond notwithstanding, WW2 led directly or indirectly to:

  1. The atomic bomb
  2. Jet aircraft
  3. The digital computer
  4. Modern rocketry

And the Cold War gave rise to

  1. The internet
  2. The space program
  3. The interstate highway system

These are the kinds of big swings that Thiel is talking about and that's just off the top of my head. You can go through previous wars as well. Wouldn't be surprised if WW1 was a real shot in the arm for aviation, for example.

Probably the biggest peacetime innovation since the end of the Cold War has been the sequencing of the human genome. But arguably that was also a computer science innovation, as it revolved around better assembly algorithms and required no test pilots, clinical trials, or atomic bomb detonations.

The first flying jet aircraft was before WW2, the Heinkel He178 on August 27, 1939.

The first freeway network was in Germany before WW2.

Before the internet gained popularity, about everyone with more than one computer tried to connect them together. There were quite a proliferation of networks before the internet subsumed them all (after all, even the term "inter" net was derived from connecting disparate networks together, not computers). There was BIX, FidoNet, Compuserve, Prodigy, MCINet, just to name a few off the top of my head. Some students at Caltech in the 70's built their own ad-hoc network when I was there.

While the other networks have all been forgotten today, to suggest that without the ARPAnet networks wouldn't have happened is without foundation.

... you do know that germany was heavily militarising for years before WWII started on 1 Sep 1939, including heavy spending on military R&D?

Germany didn't just say "hey, first ever flight by a jet aircraft! let's celebrate by, I dunno, invading Poland in four days' time!"

The jet powered Heinkel was built with Ernst Heinkel's private funds, not government funds.

And I'm sure that Heinkel was so puritanical about the jet engine not being used for military purposes that he refused to consider the substantial profits he might make from using them in the multitude of bombers he was already making for the Luftwaffe, nor that he might be able to create a competitive fighter design (given that he'd lost out on that side). I'm sure his research was not at all motivated by the possibility of future government coin.

Of course Heinkel wanted to sell the jet to the government. But the fact remains that jet engines and jet aircraft were not developed with government research & development contract money.

Neither the US, British, nor German governments wanted anything to do with funding jet research until they saw flying jet airplanes.

Jet aircraft are a poor example of foresighted government research.

How did we get onto government money? No-one was talking about government funding until your second-last comment. You were saying that the first jet engine was tested before WW2. I said 'sure, four days before WW2, but nevertheless was significantly influenced by the events directly leading up to WW2'. Not sure why it matters whether the money was public or private.

Hmmm. My point is that innovation speeds up dramatically during wars, in part because the involvement of dot.mil organizations can supersede the regulations imposed by dot.govs.

Taking your points in turn:

1. August 1939 in Nazi Germany wasn't exactly peacetime :) They were already planning to invade Poland. I believe it's pretty well established that aircraft development slowed between the wars but really got underway again with WW2 and then the Korean War.

http://www.ww2pacific.com/jethist.html

2. The Autobahn was arguably a peacetime development (though very useful for war), but it's inarguable that the US Interstate Highway System was developed for defense purposes.

3. Would computer networking have happened in some fashion? Probably, but DARPA saw it as a national security issue and cut through various kinds of red tape. Fiber could be laid across huge swaths of the country without environmental impact statements or FCC involvement. It's hard to say what would have happened without sponsorship from one part of the government.

1. Heinkel's jet airplane development was not funded by the government, as the Nazi government saw no military purpose to jet aircraft. They only got interested after it was flying.

2. Both the German autobahn system and the US interstates were partially justified by military use, and both were done in peacetime.

3. Computer networking did happen independently from the ARPAnet, and it did cover the entire country, both with dedicated lines and piggybacking over the phone lines. All the networks I mentioned were national (and even international) in reach. Everyone with more than one computer wanted to connect them together.

I should add that in Great Britain as well, the government did not begin funding of jet engines nor jet aircraft until after government officials saw flying jet aircraft.

In America during WW2, the government told Lockheed to halt their dev work on jet engines and concentrate on piston engines. Flying jet aircraft (from GB and Germany) again changed their mind.

Wow, what's with WWII and the Cold War lately? Everyone is trying to find silver linings in them. So I'll just whip out my previous comment about the broken window fallacy: http://news.ycombinator.com/item?id=3255480

The whole war = innovation thing never held much water for me.

First of all, there is no way to know what would have happened if there was no war. Some of those things might have been invented anyway. Maybe other more important things would have been invented.

Also, if you look at non-war years they too are full of innovation. The period just before WW1 brought us the light bulb, the telegraph, the phone, skyscrapers, the type writer, etc. etc. I mean, it was over a slightly longer period of time, but I don't think it would be outrageous to argue that the rate of innovation was just as high as during WW2.

Well, that's an entirely fair way of looking at it. Pshaw! A few years of war held less innovation than a hundred!

A working telegraph over 8 miles was invented in 1816.

typewriter 1829

telephone 1876

light bulb 1879

skyscrapers ~1890ish... but preceeded by buildings with similar number of stories from Roman times!

Yes, you're quite right. A hundred years (two thousand, if you include Roman precedent) does indeed hold more innovation than four.

And if you characterise ~98 years ('period just before WWI' with examples, 1816-1914) as 'slightly longer than WW2', which came in at 6 years (5 in earnest, 4 if you're American), perhaps you should pursue a career in archaeology, paleontology, or politics...

Don't know how to attribute it, but PCR's pretty amazing, too. Being able to sequence Neandertals...

Cold war wasn't a war.

Work on jet engine began long before WW2 (and it failed to accelerate during WW2).

So did the work on nuclear fission (if it wasn't for top scientists incl. Einstein convincing US President to fund it, there wouldn't be Manhattan Project at all. German scientists fortunately failed to persuade Nazis even though they were much closer to nuclear device.)

Here is where the first footnote[0] links to:

http://progressivefix.com/wp-content/uploads/2011/06/06.2011...

We've had tremendous innovation in the medical field - just look at life expectancies over the last few decades. Thiel's argument on transportation is also weak - cars/planes haven't gotten faster but they've gotten much more efficient.

The biggest innovation failure lies with the government. And governments can certainly test things out and 'fail' as we can in consumer Internet. In fact these days, almost every government policy seems to fail!

This is probably due to the cumbersome nature of the democratic process - whereas entrepreneurs can quickly disrupt inefficient systems and rely on the market to validate them.

I would say it is because a) there is a lot of money to be made and b) Moore's law forces one to move fast to corner a market because one can almost guarantee that someone will disrupt it within a few years.

Early customers were knowledgeable about this. Because of that, they accepted (partial) failure of products in exchange for earlier delivery and/or lower prices. Those customers, in turn 'educated' the masses to accept crashing products.