The Apps=>Infrastructure=>Apps=>Infrastructure Cycle

My view has been, and is, that we are in the “infrastructure phase” of the crypto market development cycle.

To elaborate, I believe that we need better infrastructure (e.g. better base chains, better interchain interoperability, better clients, wallets and browsers) before we can see a robust application development environment and so I have stated many times that right now is a time to focus on building (and investing in) that infrastructure. That view has been the prevailing wisdom inside of USV for quite a while now.

Well a couple of our colleagues at USV decided to poke holes in that argument and spent a few weeks doing research and then writing this post.

The post is called “The Myth Of The Infrastructure Phase” and it was researched and written by Dani and Nick.

I have a feeling that this post may be headed to similar territory as Joel‘s now famous Fat Protocols post because, like that one, it takes a conventional wisdom and turns it on its head.

Dani and Nick argue that there are no distinct phases but in fact, a virtuous cycle of apps>infrastructure>apps>infrastructure that brings a new market/technology into its own.

Read the post, as this argument is well researched and well made.

However, as much as I agree with their arguments, I continue to believe that for investors, the best bets right now are infrastructure bets. It remains too hard, too expensive, and too frustrating, to build decentralized apps and the big value unlock will come when that changes. I think the returns on investment on infrastructure will be higher in the phase we are in right now. There will come a time when apps development will have a better ROI, but I do not think we are there right now.

USV has made investments in decentralized apps, like OB1 and CryptoKitties, and we will continue to do that. But our primary focus is on infrastructure right now.


Comments (Archived):

  1. jason wright

    automobiles required an integrated road network to ‘take off’, but aeroplanes only short runways. i wonder if there’s a technical ‘short’ solution to this chicken and egg dilemma to hasten progress, a ‘hack’?

    1. falicon

      In crypto, I think it’s the combination of mobile and “3rd world” populations…

      1. jason wright

        I do wonder if an alternative approach is being overlooked in the rush?

  2. LIAD

    I love how are you guys have the intellectual honesty to publish the pieces on the firms blog despite them not having unanimous agreement within the firm.1. Current prevailing wisdom is that fat protocol theory has been at least somewhat debunked. Time will tell whether 2014 or 2018 conventional wisdom was right.2. The infrastructure begeting app logic passes the historic smell test however a virtuous repetitive cycle a flywheel of sorts does seem more plausible in the world of software based infrastructure as opposed to the old hardware physical world stuff.3. In crypto’s case perhaps both are wrong. It’s not infra begets apps and it’s not a flywheel. Perhaps it’s about desire being a first cause and until then everything else is irrelevant. Right now there is limited mass use case desire for current crypto enabled emergent functionality. And until that changes all is is academic.

    1. Nick Grossman

      I actually do suspect that in crypto there are some different rules, vs previous platforms — esp given the open, interoperable nature of everything. many dapps are platforms/infrastructure by nature

  3. Popa Bogdan

    How would you rate kin? As an app or an infrastructure project? I say that because i feel it doesn’t have a clear identity atm. On one side it is lauched by an app(kik), but on the other side it is a project to build on as a developer…and it’s also pieced together by combining the best of current infrastructure (eth liquidity and stellar tx speed)

    1. falicon

      Infrastructure. Or more accurately, a tool (not an app). Because you can’t do anything with it on its own…it lets/helps you build/do something else.

    2. fredwilson

      I think of Kin as infrastructure for mobile app developers to use token incentives and economics as their primary business model

  4. William Mougayar

    Not much to argue about this, except that infrastructure is being painted with a broad brush, because it really includes middleware which is an important part of it. Most App developers want to interact with easy to use middleware that insulates them from the bottom layers of the infrastructure, just as Web/Java/App developers don’t really care or need to know too much about the underlying Internet infrastructure.

    1. jason wright


      1. William Mougayar

        We need more. Solidity is not for the average programmer.

        1. jason wright

          Solidity feels more of a friction layer. I wonder if Gavin Wood is giving online tuition from his Swiss residence?

          1. William Mougayar

            he is in Germany.

          2. jason wright

            His Wiki states his residence is Switzerland.Online tuition from Germany?

    2. falicon

      Boom.Apps->tools->infrastructure->tools->apps->tools->infrastructure etc. etc. etc.

      1. William Mougayar

        Exactly. You know it well 😉

      2. Chris Lu

        Seems almost like we need something to bridge the knowledge/learning gap between existing tools/infra to the tools and infra in the new space. For example, machine learning libraries written in Python or TensorFlow bridged the gap for programmers/hackers and pure ML academics.

        1. falicon

 was a small but good start

  5. falicon

    Two things:1. Investors rarely make a market. The first “apps” phase in these cycles are generally not funded…i think we are tech. in the second rotation of “apps” right now, but it’s still really hard, and likely bad investment (as Fred states)…but also getting some wins at the app level is the only *real* thing that will push the cycle forward from here.2.

    1. fredwilson

      Bitcoin was not funded. So you are right

  6. obarthelemy

    Interesting, thank you.

  7. kidmercury

    I think you can see this playing out with ethereum and EOS. Ethereum is infrastructure, some apps on it but they kinda suck, EOS comes along, better infrastructure for apps and those apps on it are better and gettinmg more engagement than ethereum apps but still not good enough. To Fred’s point probably better to invest in EOS than dapps on it, if those are the options and one is required.

    1. jason wright

      Are you taking your own advice?

      1. kidmercury

        i only own a small amount of cryptocurrencies through coinbase, if they offer EOS soon i will probably sell the ethereum i have with them and buy EOS. i don’t think EOS is good enough and expect it to fail in the not too distant future (but i still view it as a significant improvement to ethereum, and i doubt ethereum’s roadmap will enable it to catch up), but i can see a scenario where EOS rises sharply before being abandoned for the next wave in the infrastructure/app loop. in general most of my money is in safer boring stuff, with a small high risk portion dedicated towards entrepreneurial endeavors. so in sum i don’t have much skin in the game at this time.

        1. jason wright

          EOS feels forced and inauthentic, and very ‘Hollywood’. I don’t want to give support to that.They go around other crypto communities trying to buy important members to strength their project and weaken the competition.

          1. kidmercury

            IMHO EOS is a deeply flawed system but a step forward in where things need to go, due to being a better platform for users and developers (faster transactions with no cost). EOS lacks sufficient measures to protect against fascism (takeover of governance by corporate interests), which IMHO is why it will fail. but everything devolves to fascism in the long run anyway 🙂 IMHO competition between blockchains will increasingly resemble competition between nation-states, in which currencies are weaponized as a form of economic warfare and propaganda is used to destroy consensus on competing chains while strengthening it on the same chain. the saving grace is that we’ll all be dual citizens (more than 2 probably), which i think will help keep the system honest and competing in the users’ interest. at least for a while….

          2. jason wright

            “EOS lacks sufficient measures to protect against fascism (takeover of governance by corporate interests), which IMHO is why it will fail.”That’s the DNA of EOS shining through. Sharks give birth to baby sharks.

          3. kidmercury

            i agree….i love the EOS investors and block producers on twitter, they’re too caught up in how much money they will make to see the flaws in the system 🙂 the good news is that future chains will learn from the good and bad that EOS has to offer.

          4. jason wright

            i’ve never seen a technology with more potential to support the rise of grassroots social, economic, and political movements than crypto.If Marx and Engels were alive today they’d both be learning how to code.

          5. Girish Mehta

            “But everything devolves to fascism in the long run anyway…”Ha. Thats an interesting statement you slipped in there quietly 🙂

          6. jason wright

            ‘climat sans frontieres’ – nationalists are unable to recognise climate change, and climate change does not travel with a passport.Perverse as it may seem to be, what we really need is a sharp but short acceleration in climate change effect to put just enough land under water to destroy the credibility of the nationalist climate change denial mantra forever. If the coastal margins of parts of England, Ireland, The Netherlands, Denmark, Germany, and Italy started to get soggy the political elites of those countries would force through immediate and radical reform on the world stage.

  8. PhilipSugar

    Here is the way I see the world:Timing is everything.Everything that sits between the end customer the person parting with the money and business they do that with becomes commoditized.’Yes for periods of time the business “that they do that with” will pay dearly to be able to get the tools to “do that” And there is nothing wrong making money during that time.Let’s take, Intel, Dell, Cisco, and Netscape as examples. All made tons of money as the internet needed to get built out.But do you are I know what browser I am using? How about what computers Disqus is running on? How about their datacenters networking equipment?So once it becomes a commodity then it becomes the growth of the last person getting paid.But look at Apple! Apple is the last person getting paid. Do you know any of the suppliers of the parts in your iPhone? Now how about Android? Well the reality is that Google is trying to commoditize the OS layer, but Apple has said I will control the consumer and I will also control the hardware.So that is why I so strenuously objected to that talk where he said a protocol developer is worth 10,000 more than an app developer. Maybe for this brief moment in time he is right.

    1. JamesHRH

      Apple for killed at the start of the cycle, when computing was a land rush. Gates killed them.When computing became an experience play, Bill retired and Steve killed.Experience = last guy paid ( although, it’s first guy to me, customer hands you the $ ).Bill’s customer was a developer, that’s how he covered ground. Steve was natural an endnote of cycle guy, seeing only the end user as the customer ( not how it works early on ).

    2. cavepainting

      You nailed it! This is very true. Only thing I would say is that protocols will be valued more only if apps are incentivized or see a definite advantage in using the native token of the protocol. That is a big if. But when true, more value accrues to the protocol because it has scale and leverage across many apps that find product market fit.Now in what cases would that happen? That is the big question. But literally hundreds of base layer protocols are being funded based on this premise.

  9. Frank W. Miller

    There’s so much to say about this post.1) How did it get done?The great Fred Wilson is running around (probably) inside and outside USV trumpeting crypto to high heaven for years at a time. There are two possibilities I can see. First, you’ve got some other folks in the firm with some balls to put something together like this in direct opposition to your advertised notion of things. Assuming they all keep their jobs, I laud you for a system that allows for this.Second, its a marketing play. Not too happy with the way things are going in crypto these days. Made some bad bets. Need a way to keep those LPs happy in the face of losses. How bout an internal struggle with potential scapegoats (i.e. those that wrote the post, I doubt you’ll be scapegoated) to paper it over.2) They’re rightApps, and in particular the “killer app” whatever that is at the time are the forcing functions. While BTC and Eth are perceived as the apps here, they are fundamentally flawed as I’ve said. They are neither currency nor asset. They are a new form of pure gambling where the only thing you can do is look for bluffs and try to beat the other guy from selling before you do. Online gambling is big, but this fad (and thats what I think it is) isn’t even good enough to hang with them.3) They forgot one thing that also has an effect.They did forget one thing about this particular technology that should be factored in. We’re talking about fully distributed systems. In theory the infrastructure are the computing devices that the end user owns. There’s no need to build out anything unless you’re trying to do one thing, centralize the distribution. China builds giant data centers to mine. Various companies build clearinghouses and wallet servers, etc. to try to scoop up the various elements of the distributed system in central locations. If stuff gets centralized, it flies in the face of the basic tenant of these systems, full distribution. That contradiction is fatal imho.

    1. jason wright

      It’s not fatal. It’s a bridging phase. we will get beyond it…eventually.

    2. Richard

      Building the infrastructure that supports illegal activity, including drugs and prostitution. Hear no evil, see no evil, speak no evil. 3 Monkeys VC.

    3. LE

      They are neither currency nor asset.They are marketing and PR that draw attention (and investment) to the ‘space’ at this point. But to the slant that I will make elsewhere, it was not planned that way. It just happened.Without those coins there would be nominal attention to the blockchain. They are the killer app even if (to your point) it is gambling.

    4. LE

      Online gambling is big, but this fad (and thats what I think it is) isn’t even good enough to hang with them.One reason for that is that with other types of gambling the intermittent reinforcement is programed in and fined tuned to perfection. Same with other addictions like online games (fortnite) or even game shows on tv (the easy question juxtaposed to the difficult question).

    5. LE

      to put something together like this in direct opposition to your advertised notion of things. Assuming they all keep their jobs, I laud you for a system that allows for this.No it’s the opposite. In fact there is a history in tech, dating back to at least Bill Gates early days of Microsoft, [1] that he liked when people would give him a hard time and challenge and not fear losing their jobs. It was a positive attribute not a negative one.Will note also that this is par for any small business or family run operation. The boss likes it when people lay it on the line and does not like yes men. All they care about is ‘will this put money in my pocket am I the fool at the table?’Need a way to keep those LPs happy in the face of losses. How bout an internal struggle with potential scapegoatsOnce again that’s a corporate or organizational way of thinking about it. Forgetting whether they are LP’s or your family accountant it doesn’t matter why money was lost or a mistake was made. Baring force majeure there are no excuses.[1] And perhaps it makes sense given your comment that is where it started.

      1. PhilipSugar

        People like me love getting called out. It is my advantage.

    6. sigmaalgebra

      Approach 1: Find a wave and ride it.Approach 2: Create a wave and ride it.Approach 3: Most waves are created by moving air. So, to create a wave, move a lot of air!

      1. Frank W. Miller

        The reason that apps drive everything is they are solving something real that humans need. Computers are useless in and of themselves. They are tools like hammers or screwdrivers. There has to be something for them to do to to make them useful. BTC and Eth are not useful in and of themselves, they’re just gambling. If someone ever hits on a set of features on top of the blockchain that addresses that human need, e.g. Western Union on the blockchain, then it will solve a human problem and the use and infrastructure will follow.

        1. jason wright

          Western Union on the blockchain?

          1. PhilipSugar

            It already is.Being on a college campus with a huge Chinese population. If your parent in China wants to get you money over here you do it with bitcoin. Turn it in a minute and you have no value risk. That is why there are so many houses and cars owned by young students. Get money out of China.If you want to buy drugs you need to wire your drug dealer money so they can buy it from their supplier.

          2. jason wright

            Yes. I think Satoshi had offenders like Western Union (and the banking system in general) in mind when s/he came up with bitcoin. bitcoin is the human need Frank is searching for, but in a guise he seems not to recognise.

          3. JamesHRH

            Totes legit mainstream use case.

        2. Michael Elling

          You see we have this golden thread we call infrastructure for the emperor’s new clothes.

    7. Frank W. Miller

      BTW, I just read the FAT protocols post and I don’t agree with it at all. I think the blockchain based ecosystem (if it actually ever comes into existence) will be similar to the original diagram, lots of value in the apps riding on top of the relatively mundane and standardized protocol layer.

      1. Michael Elling

        Over the past week I attended a cybersecurity conference put on by SAP and Hasso-Plattner Institute in NYC, a lecture by Google’s Chief Privacy Officer at Princeton and a NewAmerica book lecture for LikeWar: The Weaponization of Social Media.If one puts them all together and thinks about the history of radio and what happened in the form of WWII and the printing press with the religious wars, one walks away with a very dark picture of where we might be in 10-20 years. In fact every great technological/network leap resulted in some pretty horrible conflagrations. Only now, at digital scale globally, it’s probably going to be really, really bad.Yet no one is seemingly willing to put forth a new data/information paradigm to alter course from what is looking inevitable.So let me ask one thing: when are we going to get away from this simply dreadful TCP-IP protocol layer(s) (after 20 years we’re at less than 40% penetration of IPv6) and do something about building real incentives east-west and north-south in the informational stack to counter the enormous threats and inefficiencies that very few fully comprehend? In the process we can rebalance risk and address the value/cost disparity that has developed in digital networked ecosystems everywhere over the past 50 years.

        1. Frank W. Miller

          I’d separate TCP from IP I’ll talk about each.TCP is a technological marvel. If you look at the history of the development of that protocol, mountains of serious research and engineering development by both industry and academia yielded a transport protocol that works equally well over 9600 baud modems and 100 Gbps Ethernet. The sliding windows and backoff/restart algorithms are truly dazzling. The Internet highway is littered with the corpses of other transport protocols that have been proposed to try to supplant it and none come close to its performance and flexibility. If you need reliable, in-order delivery, I don’t think anything will beat it out anytime soon.Security is generally implemented on top of TCP and I don’t think that is going to change anytime soon either. There used to be something called the ISO/OSI stack. Security was generally considered a layer 5 (or Session) protocol and thats where you see it now in the TCP/IP stack in things like SSL that happen after the TCP connection is established.Wrt to IP, I was around for a lot of the transition work for v6 (I used to go to IETF to attend the SIP sessions and the v6 stuff was going on down the hall). v4 is really good and really ubiquitous but as we all know, it has a limited number of addresses based on the 32-bit src and dst address fields.Thus the thought was that a replacement was needed and IETF went off came up with v6. There are really two fundamental problems with v6 imho that have limited its use. First, the design basically just sucks. Its not backward compatible with v4 which was a huge mistake. They went from 32-bit to 128-bit addresses which is overkill [1], i.e. 64-bit would have been a more logical jump. And it was designed by committee. The last point means that it has all sorts of kitchen sink features in it like discovery et. Al. that were already solved by other things. The result was that its a hard jump operationally from a purely logistical point of view.Second, NAT was developed about the same time. NAT basically solves the address problem and adds a bit of network security to boot. Given the poor design of v6 and the ability of NAT to handle the address problem, v6 just hasn’t caught on.I’m sure this is flamewar stuff but that’s my opinion.[1] I used to tell my students. The 128-bit address field is wide enough to give a separate IP address to every sub-atomic particle in the known universe times 6. Or something like that 😉

    8. falicon

      If USV LP’s are unhappy, ain’t no LPs anywhere happy…they hit too many game winners to worry about a strike or two…

    9. Michael Elling

      Don’t forget about connectivity and “real infrastructure”. It’s every actor for themselves and the developers/platforms against all!

  10. JLM

    .In the military they tell you if you’re looking to get into the fight, “Run to the sound of the gunfire.”In business, the corollary is to “Run to the sound of the applause.” Go where the customer is applauding with their money. [Do not run to the sound of the laughter, because they may be laughing at you and not with you.]There is a reason why it is a continuous cycle of infrastructure >>> application >>>infrastructure >>> application – because the customer will pay more for increasing capabilities and increasing capabilities require better and improved infrastructure.Whether the chicken preceded the egg is not important as long as the omelet is tasty. People pay for omelets, not chickens, not eggs.JLMwww.themusingsofthebigredca…

    1. LE

      People pay for omelets, not chickens, not eggs.And to add to that ‘and only when they are hungry’.In the military they tell you if you’re looking to get into the fight, “Run to the sound of the gunfire.”I remember that as a theme in ‘Full Metal Jacket’. The character played by Matthew Modine wanted to ‘get into the shit’ or something similar.This also dovetails with conventional wisdom as far as where you want to be for career or mating. You want to be or relocate to a ‘target rich environment’.

      1. JLM

        .In the military, this one trait – looking to get into a fight is a huge discriminator amongst officers. Many times, you are given an order to simply, “Take a patrol out there and see what the fuck is going on, Lieutenant. Report back. Feel free to start some shit.”This is why the aggressiveness which is bred into paratroopers and Rangers is so important for the officer corps. In my day, every regular officer had to go to Airborne and Ranger Schools.It worked.The other funny thing is who really did move to the sound of the gunfight. Sometimes it’s a guy who is little, but has huge heart and balls, even though he weighs one thirty nothing when wet.I think this bravado is characteristic of successful entrepreneurs. They are not afraid to mix it up. You can’t win if you can’t fight.One of the bits of wisdom my father imparted to me when I was headed overseas was, “When you get into a fight, remember to fight. Always remember your men are looking at you. Whatever you do, they will do.”At VMI, everybody had to take boxing. Not only did you learn to knock the crap out of someone, but you also learned to take a punch. I got knocked cold a couple of times, but I always got my licks in.In the Army, I always used to have my units extensively trained in knife fighting and hand-to-hand combat. They used to get in a lot of trouble on payday. No knives, but lots of fights which they never lost. It builds confidence.JLMwww.themusingsofthebigredca…

        1. LE

          Exactly. And this, in a way, is like (what I call) ‘secret service agents and prostitutes’. (Not denigrating them btw). As in ‘what type of guy would agree to take a bullet and put himself in harms way’ as his job? I can tell you that guy is not me. I would be afraid of catching some STD from a prostitute. I don’t even like to shake hands with people. Firefighter running into burning building and risking life? Takes a special person to do that. I am not even close to that special person.So you have the peanut gallery taking aim at people who have a host of risk taking qualities that honestly don’t come close to ordinary people. I mean shit my Dad wouldn’t even let me ride a mini bike! And the closest he ever came to bombs was at the end of the war when they were marching the prisoners around from camp to camp and the sound of bombs meant ‘we are getting rescued’.

        2. Andrew Cashion

          Used to play a drinking game in high school called “disarm Rob” The rules were to use a medium sized kitchen knife -and the winner was to wrestle the knife out of Rob’s hand.

          1. JLM

            .In Ranger School you learn all about knife fighting – killing someone with a knife and disarming a man with a knife.Every time you start a class, they tell you to always carry a pistol.I have disarmed a lot of people with knives, but usually rubber knives. Only once ever have I ever taken a real knife from someone.JLMwww.themusingsofthebigredca…

          2. Andrew Cashion

            Not sure I could have passed the that kind of school. Would have enjoyed fighting the fighting though.…dium=social&utm_content=062316-news&utm_campaign=benshapiro

          3. Lawrence Brass

            did you ever got killed by a rubber knife?

          4. JLM

            .Few hundred times. The other guy was trained by the same instructors to kill me. Sometimes he won. Sometimes I won.JLMwww.themusingsofthebigredca…

  11. Twain Twain

    @fredwilson:disqus — Does USV have a view on Sir TBL and MIT’s SOLID infrastructure?https://uploads.disquscdn.c

    1. Lawrence Brass

      This IS the real web/3.

      1. Twain Twain

        They opened up a dev community in July so, of course, I was in the first 300 and have been following their progress.It’s a bit buggy. For example, GET calls to schema don’t work so some apps like profile info for Contacts don’t render.Still, I like that Sir TBL is thinking of “upending the Web”:*

        1. Lawrence Brass

          Are you posting about it elsewhere? to follow..

  12. Girish Mehta

    Re: “I have a feeling that this post may be headed to similar territory as Joel‘s now famous Fat Protocols post..”Serious question: Is there any evidence from the time that the Fat protocols post has been published in 2016 till today to show that it likely to be correct ?Secondly, no surprises here and agree with the general direction of Dani and Nick’s post. And their chart showing the apps-infra cycle since 1970 is great.But it glosses over the question of crypto-skeptics (like me) which is – Is Bitcoin really a killer app ? Are the other apps listed in Web 3.0 portion of the chart like Cryptokitties meaningful sustainable apps ? Or tokens ?Compare the breakout app for the Web 3.0 cycle – bitcoin with the breakout app shown in the first cycle – Email.The post is right in the argument it makes against the myth of the infrastructure phase but I did not realize till now that it – “…what we mean by the myth of the infrastructure phase — if we think about an “infrastructure phase” divorced from the apps that will use it” was a serious argument.The fundamental question that so many people have asked here so many times in simple terms remains – Who really needs and wants crypto outside a limited use case ? What is the legal, large scale end-user driven (not investor driven) demand for crypto ?p.s. In all matters crypto, I readily conceed to the “add yet at the end of the sentence” argument. There is no argument against that argument.

    1. sigmaalgebra

      > Is there any evidence …To paraphrase…”Evidence? We ain’t got no evidence. We don’t need no evidence. I don’t have to show you no stink’n evidence.”That is essentially the current position of the Democrats on the Senate Judiciary Committee.Now there is a big pop culture norm that evidence is so out of date, so pedestrian, restrictive, inconvenient. So, sure, if want to convince a lot of people of total nonsense, e.g., the NYT and human activity causing significant global warming, then work to lower the accepted standards of evidence!!!!!NYT? Sure, their top climate scientist Tom Friedman explained to us ignorant, unwashed masses in the streets that CO2 causes global warming because it absorbs sunlight.Well, Tom, let’s see: Exhale and see if you can see your CO2. Look at the gasses that come from carbonated drinks — see any gasses? If CO2 absorbed sunlight, then it would be visible, e.g., a dim cloud.Tom, to believe that CO2 absorbs sunlight, it would help to be smoking funny stuff. So, THAT’S what you saw when you exhaled!!!!!Thus, on lowering the standards of evidence, the NYT has done astoundingly well, now has their newspaper less meaningful or useful than Charmin, uh, right, used Charmin!!!

    2. Michael Elling

      Repeat after me, “nowhere in nature are purely distributed networks sustainable.” For that matter neither are purely centralized. This “anti-platform” madness misses the central mark that the internet simply imbalances risk by being settlement free. Until that issue is addressed (and the fact that everything still rides on TCP/IP) things will continue to get progressively worse. See my note elsewhere about north-south settlements.

      1. Frank W. Miller

        Human beings themselves are a fully distributed network…

  13. Steve Monk

    I noted the article mentioned a USV book club. I assume that’s internal to USV, but I’m curious about how that is run. I’d be very interested to hear how that is organized and how the discussion is facilitated.I’m interested in having such a book club at my company. Actually, I’ve started one, we’re on our second book, and I’m thinking about how to tweak and improve it.

  14. LE

    I think the returns on investment on infrastructure will be higher in the phase we are in right now. There will come a time when apps development will have a better ROI, but I do not think we are there right now.You are by and large an engineering and a numbers guy so it makes sense that you would see it this way. But business at the core is analog – but perfected and tuned with numbers. Sure ROI could be better. If the thing you are betting on works out. One thing that I never see here, that may exist but that you don’t typically blog about, is the reason and how you feel things will play out. What is the roadmap. I could tell you the roadmap for any business that I ever got involved in even if they are small ideas. By ‘roadmap’ I don’t mean ‘business plan’. I never did one of those.I think the last time I saw that was with twitter where you figured (w/o ever probably thinking about ROI) that it was a good bet because of your own personal experience with blogging.In a similar way Jeff’s idea was to have a place that had everything for sale. As such you’d end up going to that place all the time and it would become a habit. The ROI was so unimportant in the initial years that the company lost drastic amounts of money until it achieved it’s goal and became profitable. The core of that success is analog and plenty of tools were hired in order to make the numbers work to perfection.ROI might work with some type of real estate investment because that is a mature industry with no secrets at this point. (Of course the analog part is knowing where to get the bargains or having a theory to beat the other buy before the market gets hot or some other advantage).

  15. LE

    Dani and Nick’s post was much better than I expected actually. One thing though is that it studies things that worked rather than (because not possible) the entire set of things that happened which includes failures taking the same path. (And that assumes the analysis is even correct). That is the general problem I have with most books or articles to begin with. Look at GE which was a darling for so many years. Now they want to say ‘eh it’s because of this or because of that’. But the truth is they never want to do the same when things are working. When things are working they jump to identify reasons w/o knowing the pool of similar companies that the same concepts are not working for…. https://uploads.disquscdn.c

    1. jason wright

      blue horseshoe loves anacott steel

    2. JLM

      .Where is Jack Welch when we need him?The decline of GE is a travesty.Another 10% and it will be a huge buy. ATT also.JLMwww.themusingsofthebigredca…

      1. LE

        Ok let’s memorialize this and see what happens. I don’t do stocks (an advisor does some for me – mainly so I can avoid FOMO) but my ‘bet’ would be based on what you are saying it’s a good buy today. Reason is focus on whether something is the right decision not if it’s cheap or at a low point.So if I bought stocks I would buy it TODAY just based on what you said ‘huge buy if 10% lower’. (I don’t know anything else it’s a dart at a board and you are helping me point for that matter).Let’s see what happens….. https://uploads.disquscdn.c

      2. Lawrence Brass

        The devaluation of brick and mortar I can understand but high tech non-compute equipment? There is something wrong.Magic happens to me when I hear a GEnx or Trent machine revving up. More so if I am on board. Don’t tell me it is something that is losing its value.. please.

      3. Michael Elling

        But shouldn’t he share some of the blame? Was it not a lot of financial conjuring that was not sustainable? As evidenced by the fact that many of those businesses became financial albatrosses, only evident as they were unwound? The decisions of senior management last well beyond their tenure.

        1. JLM

          .Sure, theoretically anything is possible.The financial businesses – which were initially based on their ability to fund them with commercial paper which was created by their manufacturing heritage – were all subject to bigger and deeper cycles than they expected, so the strategy might be suspect.I do have to say that if you look at the stock price, Jack Welch delivered a consistent upward flow of earnings. In the end, that is all that Wall Street grades on.It is a long story, but I had first hand knowledge of their ability to match inherent, embedded losses with future gains to blunt the impact on the numbers.I saw a very skillful financial management team who kept them from having to write down mortgages (GAAP “mark to market”) by cleverly re-structuring their deals into JVs and partnerships (GAAP “cost basis investment in partnership”).Their only financial sin was getting caught in a declining market with a lot of loans with people who had other fatal problems and could not work their deals.JLMwww.themusingsofthebigredca…

  16. sigmaalgebra

    For yourThere will come a time when apps development will have a better ROI, but I do not think we are there right now.That’s like a comment on craft beer: If are considering craft beer, as customer, investor, whatever, then to heck with evaluations of the whole industry of craft beers. Instead, evaluate craft beers one at a time. Craft beers have enormous variety, so much that evaluating the whole industry is nearly meaningless. If you find a craft beer you like, then fine. If not, then, okay.Broadly the situation is similar for investment themes: Investments have to be in one company at a time, and so do the evaluations. A theme and a dime won’t, or at least shouldn’t, cover a 10 cent cup of coffee. Sure, have to consider the size of the market, would be competitors, new technology, etc., but, still, a theme is silly.

  17. Terry J Leach

    I totally agree there is a cycle and not a distinct phase of infrastructure build out in cypto. Apps point the way to what is needed in the infrastructure and the cycle repeats itself. I believe all innovation that will ever exist already exist, so I look for clues about the future of clues in technology that already exist like Webassembly, Serverless functions and other technologies.Developers will not adopt any wholesale new technology, but if cypto uses emerging technologies for browsers, virtualization it’s more likely to gain mainsteam adoption.

  18. Phuoc Do

    The first wave of crypto infra produced one type of applications: token trading. Like sea shells, or gold, when people first acquired these tokens, they probably didn’t use them as money. It looks like we’re in the 2nd crypto infra building stage.

  19. cavepainting

    No bet on apps makes sense until it suddenly does in retrospectMost bets on infrastructure will make sense until they don’tNot all infrastructure is the sameNot all apps are the sameEvery blockchain startup or idea needs to be evaluated on its ownMost will flatter only to deceive and crashWhile some might stay dormant and bloom suddenlyThere is no guiding theme or rhymeExcept to reason probabilities from first principlesAnd having the humility to knowThat You will Still be WrongMost of the Time.

  20. Twain Twain

    @fredwilson:disqus – I commented on @nickgrossman:disqus’s post that Google was an app that became infrastructure with acquisitions like DoubleClick and MetaWeb.

  21. Michael Elling

    Gertrude Stein must have said “infrastructure is infrastructure is infrastructure” at some point. Or at least she would have if confronted with this article.The metaphors using lightbulbs/grids and planes/airports are true, but then the authors fail to state, or consider, the “real” infrastructure that all of this data connectivity and exchange rides on. In the 1970s-2010s examples they miss fiber, distributed computing (Wintel), dial-up (modems) leading to broadband, and finally mobile (4 generations of wireless in a little over 20 years). These were all, IMO, pulled by market demand for easier and less expensive access to information.But now we are seeing very large cracks developing in that progression. Digital divides are growing, not declining. But this community is doing little if anything to address it.In the other historical examples there were north-south settlements between the layers that facilitated the resulting push-pull in a coordinated (scaled) fashion. Show me this north-south connection since it was broken with the 4 layer TCP/IP stack.In the real infrastructure story today (layers 1-2) it’s every end-user and network access provider for themselves and the developers and platforms against all.

  22. Matt A. Myers

    Except for private blockchains offering a centralized service between non-blind trusted parties, the only infrastructure profit seems to be coming from miners selling their Bitcoin (et al) in the fluctuations in the Ponzi-Pyramid structures of these crypto-“currencies.”It’s so incredibly short-sighted to think it’s a good idea to enable “blind trust” in systems allowing bad actors to flourish, and incentivizing them to flourish – too.Obviously the army of HODLers is growing – perhaps the most adamantly “for” Bitcoin (et al) are the ones with the biggest losses.”Fool me once, shame on you. Fool me twice, shame on me.”Or perhaps Stephen King’s version: “Fool me once, shame on you. Fool me twice, shame on me. Fool me three times, shame on both of us.”