Why Decentralization Matters

As you can see, I am not writing this presidents day weekend. I explained why on friday.

Instead I’ve been posting things that I found worth sharing with all of you.

Today, I want to share a blog post written by my friend Chris Dixon about decentralization.

This is a portion of the post that I like very much.

Decentralization is a commonly misunderstood concept. For example, it is sometimes said that the reason cryptonetwork advocates favor decentralization is to resist government censorship, or because of libertarian political views. These are not the main reasons decentralization is important.

Let’s look at the problems with centralized platforms. Centralized platforms follow a predictable life cycle. When they start out, they do everything they can to recruit users and 3rd-party complements like developers, businesses, and media organizations. They do this to make their services more valuable, as platforms (by definition) are systems with multi-sided network effects. As platforms move up the adoption S-curve, their power over users and 3rd parties steadily grows.


When they hit the top of the S-curve, their relationships with network participants change from positive-sum to zero-sum. The easiest way to continue growing lies in extracting data from users and competing with complements over audiences and profits. Historical examples of this are Microsoft vs Netscape, Google vs Yelp, Facebook vs Zynga, and Twitter vs its 3rd-party clients. Operating systems like iOS and Android have behaved better, although still take a healthy 30% tax, reject apps for seemingly arbitrary reasons, and subsume the functionality of 3rd-party apps at will.

For 3rd parties, this transition from cooperation to competition feels like a bait-and-switch. Over time, the best entrepreneurs, developers, and investors have become wary of building on top of centralized platforms. We now have decades of evidence that doing so will end in disappointment. In addition, users give up privacy, control of their data, and become vulnerable to security breaches. These problems with centralized platforms will likely become even more pronounced in the future.

Click here to read the entire thing.


Comments (Archived):

  1. TeddyBeingTeddy

    Chris Dixon is incredible. Every post he does is fascinating. (you too,of course, Fred!)

  2. TeddyBeingTeddy

    The fact that CD writes in a manner that all can easily understand means that he has 100% understanding of the topic. Impressive.

  3. William Mougayar

    And we are just getting started on this trend.First, technical infrastructure gets decentralized (Blockchain related), then new protocols/Middleware emerge, then Apps that empower users start to flourish.Attract and Extract will be an old game, soon.

    1. Frank W. Miller

      I will differ with you here. The nature of distributed apps is such that they require a full stack from infrastructure to app on day one. Unlike an operating system kernel that might have a network protocol stack implemented within it that is not used by all of the apps running on that system, all the distributed apps need their blockchain and there is no ephemeral blockchain out there just running that just exists in the ether with no app on top of it that I can just start building on. This is why I keep hitting the app part so hard. You need to know what app knobs you are going to turn on top of the blockchain before you pick up your favorite open source blockchain implementation.

      1. William Mougayar

        The stack can also consist of decentralized technologies, eg IPFS. Check out for eg the web3 architecture from polkadot.The modularity of architectural pieces is increasing with decentralization. The vertical stack is getting further fragmented as it gets decentralized too. I’m not sure I understood where the disagreement lies exactly.

  4. JamesHRH

    Anyone who has lasted more than 10 years at a large company can tell you that there is a rhythm to it – one CEO centralizes, the next distributes.Its chocolate and vanilla, with the exception that no CEO turns the company into a chaordic enterprise. Its varying degrees of centralized control.The question remains – a centralized network enforces standards by exerting control over the operation of the network while a decentralized network……….

    1. kidmercury

      Part of me thinks the decentralized wave in crypto has come and peaked with this latest run in all the currencies, and the next wave is the run to centralized crypto

    2. cavepainting

      Great article, but it does not quite get to to the heart of why this maters to an average user.From a user perspective, decentralization means agency or self-sovereignty.What does that mean?Have supreme authority and freedom over one’s physical property and personal data on four key dimensions: Ownership, privacy, security, and control.Ownership: I can prove that I own something without ambiguity and the proof is widely accepted.Privacy: I can conduct my business with varying levels of anonymity if I choose to, without affecting my credibility. Security: My physical propertyand virtual data is in my control and cannot be hacked or stolen.Control: I own my data, can share what I want, and with whoever and when on my terms. When I have agency, everything happens with consent and all trade-offs are conscious choices.I think we need to evolve more solid criteria for judging apps that claim decentralization; not by tech jargon (like zero knowledge proofs or Snarks), but by what it means for users on these four specific dimensions and in specific contrast to centralized alternatives. The import is of course different depending on specific verticals and use cases.

      1. JamesHRH

        When I first started in startups, i had peripheral exposure to a phenomenal private equity investment ‘ circle ‘. Families that called on one another with opportunities. The circle had high standards, low friction and frequent successful syndications.It was both decentralized and centralized, in that there was no formal physical structure that held all together. However, access to the circle was very difficult, required both a level of access and of assets that was hard to come by, while ongoing access was deeply trust based and banishment from the circle ( which I never heard of or witnessed ) was based on behaviour.No one set or monitored these standards, but the benefit to being in the circle was so great, that behaviour was never the issue.Can you think of an online network – of any kind – that you know of or can imagine?No, me neither.Blockchain my be the next wave of development platform, but only because a new generation of developers wants their own super cool flavour of ice cream.There is nothing inherent to the invention that provides actual value to actual users…….because trust is what BlockChain says it will provide and no one wants to trust each other when they can have a 3P do the job for fractional fees.If you are missing the point, it took the highest quality of behaviour to make that investment circle function. Most of us are not at that level.The crypto currencies will be a store of value, for people who want to be black market or grey market. No one mainstream will touch the stuff.

        1. cavepainting

          There is a bearish case for blockchain and it is similar to your arguments.But there is also the possibility that developers figure out how to build decentralized apps that actually make user value propositions better not worse.All human progress has occurred because of crazy people with weird ideas that seemed unworthy at the outset.These things are probabilistic, not deterministic and none of us know how exactly things will shape up over time. We are just making our best guess and taking a leap of faith. While the chances of high value-add DApps might seem remote now, that may not be the case when infrastructure and platforms get more evolved.

          1. JamesHRH

            But, the probability of web apps having wide appeal caught wide spread attention when lots of Normals or Normal Bridgers got exposed to them.That is not the case w DApps – its techies and finance people, no Normals or Normal Bridgers.Supernova folding upon itself, within the next 18 months.

          2. cavepainting

            Don’t disagree that a huge correction is in the offing. But I disagree that the entire sector is a dud beyond dark-net and store of value use cases.

          3. Lawrence Brass

            What idea do you get from the crypto sector bouncing in-sync with the equities market? Supernovas don’t do that. Crypto is a part of the system now.Feel the Force, James. 😉

          4. JamesHRH

            I have to admit that my greatest concern is, that despite a healthy belief in the complete amorality of Wall St, they will just add the crypto hocus pocus to their portfolio of Caveat Emptor offerings.:-0

          5. Girish Mehta


          6. sigmaalgebra

            > All human progress has occurred because of crazy people with weird ideas that seemed unworthy at the outset.These things are probabilistic, not deterministic and none of us know how exactly things will shape up over time.Maybe reality can look this way, but if just look at some historical data then see that lots of big things happened from being planned and, then, the work done according to the plan resulting in the intended success.(1) When the Wright brothers went to Kitty Hawk, a long trip from their bicycle shop in Ohio, they had done some good engineering and experiments and believed that they would be successful and were; they were right and for the right reasons.(2) Kelly Johnson’s P-38: Just as Johnson imagined it, it flew and was a good candidate for the best fighter of WWII. There were some really good V-12 engines from Rolls-Royce and then Packard and Allison. The P-38 used two of them and was quite light otherwise.(3) Kelly Johnson’s SR-71: The Pratt and Whitney Florida group had done good work on a turbo jet engine that could at about Mach 2 switch over to a ram jet engine and then go to Mach 3+. And he learned how to work with titanium. Long the highest performing airplane, and never lost to enemy action.(4) The Manhattan Project: Szilard, Teller, Fermi, Wigner, Einstein, Oppenheimer were all right, right from the beginning, for the right reasons. They had two approaches, and both worked just as planned.(5) Transistors: In the 1930, AT&T knew that vacuum tubes were too hot, noisy, and unreliable. They had solid state rectifiers and wondered if they could get a solid state amplifier and get rid of vacuum tubes. They started a project. WWII was an interruption, but soon after the war they had success — Shockley, Brattain and Bardeen were right and for the right reasons.(6) Satellite navigation: With the first missile firing submarines, the US Navy wanted some better navigation. Some guys at the Johns Hopkins University Applied Physics Lab (JLHU/APL) did some back of the envelope calculations, and soon the Navy had their satellite navigation, just as planned. The receiver on the roof reliably knew its position within one foot. Later the USAF wanted more accuracy, especially vertically, and got GPS, as planned.Lesson: It is possible to have big successes just as planned.

          7. cavepainting

            Fair enough. But you are basically selecting and showcasing the successss. For every one of these, there likely were hundreds that failed. Also I am not saying you cannot do castle building, just that there is no way to be absolutely deterministic about this.If you can guess the probabilities better than others because you are closer to the problem and understand something others don’t, you have a unique advantage. In many cases, necessity and demand acts as an impetus like no other(like the manhattan project).

          8. sigmaalgebra

            > For every one of these, there likely were hundreds that failed.In my experience, not for projects done by the people and organizations I mentioned. Their ‘batting average’ was really high. For them, it’s still more amazing that the main project go-no go decisions were made from project proposals just on paper.Yes, for them and their projects, there was not much concern about “product-market fit”, that is, will the final product be accepted and applied as intended in practice. But, for transistors, vacuum tubes were already very popular in practice, and transistors were so much better that transistor based replacement products were nearly guaranteed success. Sure, there remained the issue of commercial competition, whether Zenith, RCA, GE, or whatever, transistors were going to replace nearly all vacuum tubes.Lesson: For the commercial market acceptance, pick a problem where the first good or a much better solution will be nearly a “must have” for enough people and revenue per person to make a successful business, and then invent and deploy that solution, hopefully with some barriers to entry.For a barrier to entry, AT&T knew the world shaking value of what they had and arranged the patent situation so that they were giving the transistor to the world for free.For more from AT&T, R. Hamming noticed that when data was sent, it could arrive corrupted. So, he invented error correcting coding (ECC). For this he went into what was then some quite obscure abstract algebra, really linear algebra but where the numbers were from finite fields, e.g., integers modulo a prime. Cute. Now Microsoft’s SQL Server nearly insists on ECC.Once Hamming did his work, there wasn’t much doubt about the value; it was fully clear that there would be valuable applications.He picked an important problem and found an excellent solution, and the first good solution. That the work would be widely used commercially was beyond doubt.

        2. kidmercury

          crypto stuff will enable greater scalability. right now AWS hosts a huge portion of the internet, google is trying to crawl every web page/mobile app screen…..it’s not sustianable. much like how internet decentralization unleashed a wave of contnet abundance that could never happen in the material world, so too will blockchain stuff take digital aggregation and syndication to an unpredecented level. they will still need, though, a central network operator to manage the whole network, principally through monetary policy that aligns incentives of all participants to achieve maximum scalability.

          1. JamesHRH

            This would be a version of a BlockChain that does not allow anonymity?And, its really just a new network management configuration, not a new financial system configuration.Big, but not as big as advertised?

          2. kidmercury

            there’s probably a market for anonymity blockchains somewhere, but IMHO that is not where the real action will be.it is a new financial system configuration, because tokens will drive all the incentives. when “blockchain youtube” is born, a whole bunch of computers are going to have to agree to share resources to deliver videos. who shares what, how frequently, at what price, etc will be managed on the chain, and the laborers will be compensated in youtube coins, or whatever they’ll be called. the value of those youtube coins may trade on the free market, and they may also be exchangeable inside the network by service providers authorized to sell in an on chain marketplace. this will have enormous financial consequences.agree it’s not as big as advertised. it’ll be bigger.the problem is with the decentralization argument, what is really going to get the magic going and harness decentralization will be a centralized network operator that may seem draconian at times, especially after the honeymoon phase is over. but i think the honeymoon phase may last a very long time, much longer than the honeymoon with google and facebook.

          3. PhilipSugar

            I have thought about that….SETI but for profit: https://setiathome.berkeley…Now it might be like Uber, where you manage to find people that don’t really understand the true costs, plus you wring out all of the other costs, like management, etc. Oh will the ISP’s hate and try to throttle.The issue will be though security. Because like all things anonymous you will get bad behavior. Really bad behavior. I know first hand somebody who got their streaming server hacked, wasn’t checking and was serving up child porn.Not fun and costs a ton of money and time when the police raid your house put you face down cuff you, take every piece of electronics you miss work and your wife who is a elementary school teacher has one hell of a shit ton of explaining than you do which was very uncomfortable.

      2. William Healy

        I like the categorizations. Simple and precise.

    3. PhilipSugar

      I think here is the thing: Managing the top of the S curve is tough for any Enterprise that is not being controlled and owned by the founder. If you are a CEO that was brought in on the steep part of the S curve and being paid handsomely, it is hard to say, well growth is over. If you went public, the market is going to punish you brutally when growth is over. Too high on the way up, too low on the way down. If you get brought at the top of the S curve, you are expected to get growth going.So now you as the CEO who knows you are making too much money is faced with the three envelope problem. You got three envelopes from your predecessor to open in order if you run into problems1st: Blame everything on predecessor2nd: Re-org3rd: Prepare three envelopesMy example of opposite is the company SAS. James Goodnight celebrates a 1.3% YoY growth: https://www.sas.com/en_us/n…Why? Because he can.

      1. JamesHRH

        Agree totally.The central theme is control.The central theme to Blockchain is no central control.Who is putting big $ into that?Gamblers.

        1. PhilipSugar

          There is a saying in Gambling. If you look around the table and don’t KNOW who the sucker is……..It is you.If you I and two of my buddies are playing team poker against the rich fifth person. They will end up with nothing, heck we might let them win a couple of times and then pull their underwear over their head, hit them with a taser, kick them in a private part, and take all of their money.I am not debating Chris’s assertion about the S Curves. Great point. But an unregulated game???If you go to a Casino, there are some rules, and some controls. Not a lot and frankly really they just want to make their percentage, so are they going to look the other way? Yes. But not blatant.You into an unregulated card room?? Good luck.I’m not one saying no regulation!!In my utopian world each public shareholder (and that is even if you own a mutual fund or are in a percentage) You get to vote your approval for all pay over $1mm per year and the yea or nay means an untaxed dividend of the money. Fund managers. No vote. Allow you to bulk all your shares together and have one vote per year, for all your holdings.

          1. JamesHRH

            But you are then saying that BlockChain is just new plumbing.You are also stripping it of the gambling attributes – no central clearinghouse, no anonymity. Its a tool and a revolutionary one…but not one that upsets the social order.I can get down with that.

          2. PhilipSugar

            I know what I do not know and I don’t know enough to say.What I do know is when I see tons of people that don’t have a clue and are looking to make a quick buck because they don’t have one, gambling with money they can’t afford to lose

  5. Steve Zengel

    Thanks for taking time to share this. Hope the curves on the slopes were as awesome as those in this post!

  6. Grace Schroeder

    What does this mean for the proliferation of point solutions?

    1. SubstrateUndertow

      Isn’t the core mojo-magic of a network-effect economy that it finally offers tools for facilitating an organizing gradient that pushes point-solutions towards more integrative organic/organismic interdependency-solutions ?

      1. Grace Schroeder

        That’s been my view. More function has been offloaded from big platforms to newer technologies. My since is that the integration application’s job now is to orchestrate data across connected tools – and ultimately parse the relevant functions and data to the blockchain. (FULL DISCLOSURE – SMOKING MY OWN CRACK)

        1. sigmaalgebra

          What’s the role, purpose, utility, upside of the blockchain?

          1. Grace Schroeder

            Mind you – my worldview is a work in progress. My sense of blockchain technology in its current form is that it recognizes that there is efficiency in recognizing longitudinal transaction data that must be shared among multiple parties as a separate class with unique lifecycle requirements. You can get on the train when you need it via integrated solutions, and the train rides on for newcomers, new cars, and new cargo.

          2. sigmaalgebra

            When I was a teenage American male, I was into cars and worked on them, e.g., the old one I shared with my brother during high school and college. So I had tools. I still have tools. When I go to an auto parts store for, say, a headlight, air filter, automatic transmission fluid, etc. I see a huge collection of tools I don’t have in my toolbox. E.g., I have only one size of the Torx screwdriver and none of the several other sizes. I have no tool for expanding the split rings common in automatic transmissions. I got the one Torx when I needed it, e.g., for changing a headlight or installing a new side rear view mirror.Now I’m in computing much more than auto maintenance. Again, I have lots of tools, want nearly all of them; but again there are lots of tools out there I don’t have or want.More generally I have lots of tools, mechanical, mathematical, computational, etc., but there are a lot of tools I don’t have and that because I see no need.So, in principle there can be> longitudinal transaction data that must be shared among multiple partiesFine. But:(1) Are there important uses now?(2) If there are such uses, is blockchain technology clearly the only or best “tool”.That is, is there a problem that needs a new tool, and is blockchain the best or a good such tool?Or, nearly no one just accumulates just every possible tool that might be useful eventually.E.g., when I got a Torx screwdriver, I knew just what I was going to do with it. For blockchains, I don’t see much that I or anyone else will do with it that is not on or over the line of what is legal.The situation of too many tools still looking for important uses, is, as I suggested by example, quite general.An old critical remark on a lot of academic research seminars is solutions looking for problems. Often the criticism is justified.

          3. Grace Schroeder

            Important use now – off the top of my head, I think of the background check industry, for humans, business entities, and properties during a due diligence process. In order to get the longitudinal picture of an asset or human — and to make sure they aren’t encumbered by legal or financial liabilities — one must visit a dazzling array of online and offline resources some synchronous, some asynchronous. Even then, one missed thread can make a difference. With a uniform view of the truth, the friction goes away. Blockchain isn’t the entirety of the system, and I don’t know enough to declare any solution a winner, but on the surface it makes sense.Set that aside. Many times it is possible to address new challenges with the same toolsets, and they work just fine. Until they don’t. For example, point to point solutions work great until the attributes of a single SKU sit in several systems/databases, and you need to not only update them based on events in yet another database — but to also expose those changes via an API that gives you a slot in a supply chain that you would otherwise not have. At that point, using old tools begets the layering of spaghetti that turns into data cement. That way lies madness.In terms of the blockchain, the idea that certain data interaction types (probably not even the right terminology) should have their own trains intuitively makes sense. And, given the amount of money thrown at it (remember, a lot of the early solar companies went bust, but it’s flourishing now on the wings of persistent IP), I think blockchain, too, will continue to emerge, survive and thrive. As a natural extension of server virtualization, it seems to couple nicely with the shattering of large monolithic systems into loosely coupled collections of orchestrated interactions.

      2. Michael Elling

        All networks tend toward centralization. The key is to nurture the edge. This starts with the core recognizing their value is mostly derived from the edge; and not vice verse. Equilibrism is the answer: http://bit.ly/2iLAHlG

  7. Ajith

    Interesting post, thank you for sharing. Two quick thoughts:1. Decentralisation may not be the best option for matters of life and death, when we need a centralised player to be liable for any losses — Decentralised uber may be a terrible idea. By the time the network penalises a bad actor, irreversible damages could occur.2. CD appears to be conflating decentralisation with ‘cryptonetworks’. Cryptonetworks with blockchain architecture similar to bitcoin and ethereum are unsuited for processes that aren’t zero-sum and those that don’t need an immutable ledger (which is different from an immutable database of records).I love bitcoin and ethereum and am an avid participant of this great socioeconomic crytpo experiment but honestly don’t know what to make of all the blockchain hype!

    1. William Mougayar

      Then you’re touching on the topic of governance, which Chris mentions, although it deserves a deep dive itself.

      1. Ajith

        Hmm yeah, sorry, my example may have over simplified the point i was trying to make. Essentially, even in a decentralised network, central players become necessary due to liability/safety issues.

        1. cavepainting

          Nothing will ever be 100% decentralized. But the current model where a few hold the data and fates of the billions is an unsustainable one. Directionally, value needs to reverse its course from flowing into the center to flowing to the edges. The rate at which happens remains to be seen.

  8. kidmercury

    I don’t get the part where he says every participant in a crypto network has the same incentive of seeing the currency appreciate. Outside of Bitcoin, can someone provide an example of that being the case?

    1. jason wright

      what’s the definition of “participant” in a crypto network? the ‘how many’ is a whole other head scratcher.

  9. jason wright

    “Decentralization is a commonly misunderstood concept. For example, it is sometimes said that the reason cryptonetwork advocates favor decentralization is to resist government censorship, or because of libertarian political views. These are not the main reasons decentralization is important.”Misunderstood from the perspective and priorities of a venture capitalist perhaps, but not necessarily so from other perspectives. Resisting government censorship may be a very germane and important reason to advocate decentralisation. A broader perspective may be a healthier perspective. Capital is just the juice.

  10. MildredMarianne

    Decentralization has the potential to lead to the creation of New Inclusive Economic and Political Space that are aligned with the ideals of Martin Luther King and Gandhi.Decentralized technology must have an ethical framework of policy and governance that is aligned with a commitment to a world where justice, peace, and love replace, hatred, exploitation, and violence.I will only use (and build) technology that is aligned with this pledge https://douglass.io/bcpledghttps://uploads.disquscdn.c

  11. Rob Underwood

    Strongly recommend this week’s Sam Harris podcast with Niall Ferguson as a good complement to Chris’s post — https://samharris.org/podca

    1. sigmaalgebra

      I took your advice and tried the Sam Harris thing. Gee, no thanks.(1) Their reasoning is far too vague. Especially the guest kept going back to history for all his answers. So, right, if the only tool one has is a hammer, then they see every problem as a nail.(2) They both got all wound up about how Russia affected the election but gave no details about any significant effects. Surel “liberals” on the two coasts may have paid lots of attention to Facebook political ads (I go to Facebook occasionally but don’t remember ever seeing a political ad) and are concerned about ads placed by some Russians (from some recent news, apparently without much money to spend), but in the “flyover” states the Trump voters (A) no doubt paid much less attention to Facebook ads, (B) had some rock solid pocketbook reasons to vote for Trump, especially on immigration, our trade deficits, NAFTA, and other trade deals, (C) getting US plants and businesses going again, (D) getting rid of the “mandate” in ObamaCare, (E) tax reductions, (F) stopping the inheritance tax (which has devastated many US farming families), etc.(3) The guest has had lots of fears about Trump for two days a week and Harris seems to think Trump is a disaster, but neither of them gave any actual, specific things wrong about Trump. Sure, Trump has enemies: There are some powerful people who really liked Obama and the Clintons and are totally pissed off at Trump, especially over immigration and new trade deals. Well, these Trump enemies have for nearly two years now thrown against Trump all the dirt they could find, but so far the worst they have found was something about two scoops of ice cream. Since the Trump enemies are no doubt throwing the worst dirt they can find and are throwing nothing at all significant, the enemies are in effect giving Trump a really clean bill of health.One big point in the podcast was that more communications doesn’t always lead to agreement but leads to divisions, cliques, tribes, etc. Well, the communications DO cause more good information being available. But, still, objectives differ, and getting consensus and power is still highly competitive. So, the divisions they are seeing are basically just competition.But the best information on-line — e.g., Wikipedia, what can be found via Google, reports from Department of Labor, stock market data, economic data, on Twitter with what Trump actually, really, in fact, word for word actually SAID — has improved the information quality enormously. In fact the good information is so good that the Trump enemies are pushed to pump out just totally made up nonsense, “fake news”, propaganda as in the Goebbels “If you tell a lie often enough, then people will believe it. Even you will come to believe it.” (IIRC).The two guys in the podcast, for all their veneer, trappings, and belief in rationality are being just emotional and irrational about Trump — their fears are no more rationally founded than some giant monster hiding in the dust under the bed.Guys, if you have some solid evidence where Trump did something significantly wrong, then I’m all ears — cough it up or just shut up.

      1. Rob Underwood

        I disagree. I think it’s a well done episode, especially the early part about network models and their historical context. That they use history as a primary guide I don’t mind.Also, they both make the explicit point that the Russia impact on the election may be getting overblown and the interviewee specifically says that Trump would have won without Russia intervention. Don’t know why you state that “They both got all wound up about how Russia affected the election” when they both say that’s likely getting too much attention.

        1. sigmaalgebra

          > That they use history as a primary guide I don’t mind.I’m not thrilled by the quality of information from historians. E.g., they don’t do much with causes. Their methodology is poor. For arguments about today and the near future, I’d rather have the data and rational points from today and not the first printing press.E.g., on the lazy S curves today, here at AVC, I’ve posted several times a lazy S curve that one day for the FedEx BoD got our two BoD representatives from General Dynamics to stay and that, thus, saved the company. And there is some history of success with lazy S curves, but I never pursued or argued for a lazy S curve based on history and, instead, always argued based on a fairly narrow business context and there word of mouth advertising. From that, I wrote out a differential equation. I got the lazy S curve only from and after solving the differential equation. That there is some historical support, okay, but I argued based on word of mouth advertising and a differential equation, not history.> Don’t know why you state thatBecause it seemed that way in the early parts. IIRC, they believed that without the Russian ads on Facebook, Trump would have lost.But I confess: IRCC the podcast was 2+ hours long, and I confess I listened to only maybe the first 45 minutes of it!!!Maybe I get three demerits and have to write assembly code for an 8086 that prints out 1000 times “I will listen to all of a podcast before commenting on it”.

    2. Michael Elling

      Ok, pure centralization is bad. And pure distribution (decentralization) has not been proven to work. So what’s the answer????

      1. Rob Underwood

        Federated models?

        1. Michael Elling

          Yes. But a) when have those worked in humanity’s winner takes all history, and b) assuming you buy into federation, what are the settlement (incentives and disincentives) mechanisms between the federated groups. Just look at state vs federal rights issues over the past 240 years in our country; or what’s going on in the EU presently.The internet is completely imbalanced, with risk being mostly placed on the receiver. It’s not a true “inter-network of networks” without a terminating settlement that forces the sender to put skin in the game. None of the decentralization crowd are talking about marginal difference between actors with-in a network or between networks. Those differences are only aligned with economic incentives/disincentives.To accord $1 trn or $100bn, let alone $1bn of value to the crypto sector that you can’t even call half-baked from a technology perspective, let alone pointing to a proxy business model that worked in the past is absolutely ridiculous.Thanks for pointing out the podcast. I had jotted down Square and Tower a few weeks ago, but don’t remember where I’d stumbled on the reference. I like Ferguson’s approach to the analysis of history and constantly looking for counterfactuals. History must be understood from the moment before it occurred; not the moment after.

  12. sigmaalgebra

    I read the whole thing. So, he discusses “decentralization” of hardware (e.g., the Internet), software, and organizations then somehow comes to believe that the blockchain ideas will become really important.I believe that rationally he is standing on rotten lumber of a pier and seeing things in the fog.His cases don’t really work: E.g., is AWS (Amazon Web Services), a “cloud”, centralized or decentralized? Well, Amazon may by now have some of the server farms in different US states or on different continents and in that sense have decentralized hardware. But Amazon is run by Bezos so is centralized, a hierarchy. But the AWS cloud has lots of businesses with their software and data at AWS, and these businesses and their software and data are very decentralized.If we want some decentralized computing for block chains, then maybe put them all at AWS where Bezos is over them all?That is, his dichotomy of de/centralized is highly fuzzy and not very clear.The essay ends with a claim, mostly just by analogy, that the block chain ideas will catch on, grow exponentially, because they are in some respects decentralized (of course the open source software is centralized on Github or some such), all without saying just what value the block chain ideas will bring to just what pressing problems.The claim that usually people didn’t see what problems the Internet solved is not true: (1) The advantages of email were seen right away. (2) Lots of people in business worked with huge catalogs, really bulky, nearly always out of date, a pain to use, very expensive to produce. Putting those catalogs on-line as Web sites with HTML files was obviously a good deal.People flocked to PCs for one really important reason — “document processing” or in simpler terms getting rid of the typewriters. That was NOT subtle or obscure.Generally, if there is an application that needs to be distributed in some way, then we can do that now, with or without a block chain.

  13. cavepainting

    At the beginning of the S-curve, the third party wants to do a good job for its primary stakeholder, the user. There is no one else who is more important. It will do everything possible to win the users’ approval and trust.But as the third party acquires more users and eyeballs, its business model gets dependent on advertisers and it starts taking users for granted. Its primary stakeholders are now the real customers (advertisers) and the stockholders.If the customer is always the user (and a paying one at that), and there is significant competition, there is unlikely to be extraction. The third party simply cannot afford to lose a customer.So, the root cause of the evils of centralized internet behemoths are two fold: a) the advertising business model that marginalizes the user, and b) network effects consolidating traffic and hence ad spend into Google and FB.So, the question is: Are we really solving these with decentralized apps? How?It is almost ridiculous to see so many blockchain companies managing advertising data on-chain. We are kind of missing the whole point of this again.

  14. jason wright

    what is this “…“no platforming” of users” Chris refers to in his post?

    1. PhilipSugar

      I couldn’t help but see the term Nylon. Maybe it means something different, but it doesn’t show the end. Living in Delaware: https://www.delawareonline….

      1. Michael Elling

        We are brainwashed into focusing on success, not failure. As a former analyst I called the 80s and 90s the bull-sheet market. It was easy with a spreadsheet for those Finance BS and MBAs to develop models with constant growth over 5 or 10 years. Never once was there a “down” year.

        1. PhilipSugar

          Truer words never have been said.

    2. sigmaalgebra

      Cute. There are a lot of lazy S curves in business growth.Often here I justified one case via word of mouth advertising via the differential equationy'(t) = k y(t) ( b – y(t) )where t is time, y(t) revenue at time t, b the upper limit on revenue (least upper bound, asymptotic limit) from the size of the market, and k a constant. So the equation is from assuming that the rate of growth in revenue, y'(t), is proportional both to the current revenue y(t) from current customers talking about the company and target revenue (b – y(t)) from customer hearing the talking.And, there y(t) is a lazy S curve, apparently a specialization of the logistic curve.But that is a narrow derivation of a lazy S curve. In practice the lazy S curves appear to be more common than cases where the word of mouth advertising is justified as the cause.E.g., you mentioned nylon. It may be that Intel and Microsoft are also examples. So, Intel went from individual transistors to integrated circuits to the 4040 or some such as the basis for a calculator to the 8088 for the IBM PC, to several generations of more and more powerful microprocessors, to 64 bit addressing, to processors for high end server farms on the Internet. And, Microsoft’s software went through corresponding steps.So, is there some math for describing this larger behavior?

  15. Gayatri Sarkar

    I think this is one of the most powerful entrepreneurship story I have ever heard. True decentralization lies in empowering others.https://www.youtube.com/wat

  16. Matt A. Myers

    We don’t have a well-functioning and cooperative internet because of governance; remember how governance is the final layer of competition?If people believe that decentralization is the only part to the solution to these woes, then they’re simply wrong.Decentralization is part of the solution and necessary as a backup access point to data (information, communication, etc; every home perhaps should be part of a mesh network – this would prevent companies like Comcast from behaving how they do, along with organizations on other layers), however we shouldn’t by default be in fear of centralized systems because of what ifs.If the population is 1) aware they need quick access to decentralized systems for safety against (perhaps this should be added to the Constitutions of countries), and 2) their government is supportive of this, then the competitive nature of our world and the internet will eventually weed out the bad governors and good will prevail.Platforms and services, if they want to make agreements to cross-link/integrate with each other, how do they decide what their value is and what reciprocal value will partners they’d make need to return? They’ll surely both want it to be equal (or Boards of Directors likely don’t want to be on a seemingly lesser end of a deal), even though it might not be equal – which then should lead us to lessons in childhood of sharing with others, playing nicely. Otherwise you get what we have today, people who build and govern these online Kingdoms, while not working together to better the Kingdom as a whole – increasing efficiencies, productivity, etc.Above along with a belief in abundance, where anything you put your mind to is possible – that unfortunately most of us are not or supported to believe or reach to succeed in this way by our parents or community, we need to believe that once we have enough for ourselves that we can share everything else. And once we have enough for our immediate country and population, we should take pride in and gratefully find how we can support others to reach this level of understanding/consciousness — and away from fear, war, violence.Else we fight others to secure “enough” resources. I believe however we’re smarter than that, and I feel we all intuitively see the abundance (which makes the suffering we see and experience even more difficult to swallow and most often leaving people feeling disconnected and alone. This sense of abundance just out of grasp is why we must be tricked and manipulated (subtly programmed or controlled) into believing there’s this relentless evil and that violence against others is okay; imprinting at birth into the context and environment your born into and other.This also leads into the idea of abundance vs. scarcity – scarcity being fear based.The only way scarcity can exist, especially in today’s time, is because of bad governance. We have the technology and resources to feed everyone in the world, and in a very quick timeframe, if we could be governed to work together – and not worry/fear; we’d of course have to do everything sustainably which then adds some cost and time, however with abundance that’s fine – there’s infinitely more resources to use, something Elon’s helping us access exponentially faster with his SpaceX efforts.Once we fix governance of these major platforms (whether the platform of digital technology service or that of a country or global organizations), whether by means of creating competitors or working together to improve relationships between these important landmark services/platforms, then everything will run and more efficiently and people will be safer, and live more productive and happier lives full of joy.@jameshrh:disqus and @sigmaalgebra:disqus also made good supportive points here.Namaste.