Can Jack Dorsey Reinvent the Internet by Making Twitter More Like Email?

The case for a technical free speech fix


On Wednesday morning, Twitter co-founder and CEO Jack Dorsey surprised many people by tweeting out an ambitious plan to hire an independent team, called "Bluesky," to explore re-architecting Twitter as an open, standards-based "protocol" rather than its current state as a "platform." As part of that tweetstorm, Dorsey cited a paper I wrote earlier this year, published by the Knight First Amendment Institute at Columbia University, entitled "Protocols, Not Platforms: A Technological Approach to Free Speech."

Assuming you don't feel like reading that 7,000-word long academic piece on what this all could mean, I will make the simplified case here for why this could be a very big deal, while also noting that there's an even greater probability that it will mean very little.

Let's start with a simple analogy. Compare email, an open standardized set of protocols including SMTP and IMAP, with Facebook Messenger, a proprietary, centralized messaging tool created, owned, and controlled entirely by a single company: Facebook.

With email, anyone could (with a bit of technical knowhow) set up their own email server. But since few people want to do that, there are plenty of places to get your own email address and client (the interface through which you receive and send emails). Your internet provider often will give you an account, or you get a Gmail account from Google like basically everyone has. Here's the interesting bit, though: even though Google has a huge market share of the email business with Gmail, the fact that it's based on open email protocols means that you don't have to rely on Google for anything. You can bring your own email address into Gmail if you want. Or you can use Gmail using a third party client, like Mozilla's Thunderbird.

And, perhaps even more important, it doesn't matter which combination of these things you use because you can still communicate with anyone using any other email system. Don't like what Google is doing with email or worried that the company might spy on you? No problem, export your emails and contacts and go use something else. Nothing breaks. You don't lose access to anyone else. Indeed, Google actually has tremendous market-based incentives to "not be evil" in this scenario, since it's so easy for you to go elsewhere.

Compare that to Facebook Messenger. If you want to communicate with someone on Facebook Messenger you need a Facebook account. And you need to use Facebook's app. And you can only communicate with other people on Facebook using Facebook Messenger. If you don't trust or like Facebook Messenger, you can certainly move to a different messaging app—but you lose your history, you lose your contacts and you can only communicate with others who use your new choice of apps. And, worse, if Facebook decides it doesn't want you on Facebook any more, you're entirely out of luck. Facebook is, literally, the monopoly provider of Facebook Messenger.

The email/protocol example is the way much of the internet used to be in the early days. The Facebook Messenger example is what much of the internet became during the web 2.0 timeframe. Old-school open protocols were seen as less user-friendly, and less sustainable without a big company backing them.

What Dorsey is proposing, however, is to take Twitter—a proprietary, closed system—and see if it's possible to move it to the historical, more open protocols of the early internet. This would mean giving up centralized control, pushing more power and control out to the end users, and creating a more competitive market for a better version of Twitter. 

While it has resulted in plenty of eye-rolling, Dorsey's nod to the possibility of cryptocurrency/blockchains is quite interesting here as well: "Blockchain points to a series of decentralized solutions for open and durable hosting, governance, and even monetization." As he notes at the end, this offers up a potential business model that could keep a protocol sustainable (unlike in the past) without the need to resort to sucking up all your data and targeting ads.

It has the potential to, at the very least, shift the discussion on three of the biggest complaints concerning the big internet companies today: competition, privacy, and content moderation. By default, it enables more competition. Depending on how it's structured, you can reduce the privacy questions by (1) reducing the need for data since you might not need targeted ads to support it; (2) handing the data control back to the ends or to more trusted third party data stores; and (3) incentivizing better behavior because of the ease of switching. And, it allows for better content moderation options by allowing there to be both competition at the moderation/filter level, but also in allowing end users to opt into their own level of moderation comfort, rather than leaving it all up to a single monolithic entity.

If this works, it would completely upend much of how the internet is looked at today, potentially limiting the internet's biggest annoyances, while retaining its best features: the ability to freely connect and communicate with people around the globe. But that's a very big "if" at the beginning of this paragraph. There is no history of taking a proprietary, closed platform and turning it into a broad, open protocol. However, Twitter is well-positioned to make it happen—perhaps better than almost anyone else. Other efforts to build a new social media protocol from scratch (and there are many) suffer from a failure to attract a large enough audience. Twitter brings scale.

There are many reasons why this could fail. But if it does succeed—and I believe it could—it would represent a big shift in how an internet service at scale can operate, and it would change much of the discussion we've been having over the past few years concerning the position and power of big internet sites.

NEXT: California Freelancers Suffer From Totally Predictable 'Unintended Consequences' of Gig Worker Protection Bill

Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Report abuses.

  1. If you read the twitter thread where he talked about it, there were a ton of people talking about ActivityPub. This is because someone has already invented the technology that Jack wants to make and standardized it with the Web’s standards bodies, and there’s a number of platforms out there already using it. The exceptions are Facebook, Twitter, and a handful of other monolith holdouts are against implementing the standard. I think Jack sees the train coming down the track of enforcing this sort of interoperability, and instead of an open standard he wants to try and force everyone onto a standard that Twitter controls.

    1. instead of an open standard he wants to try and force everyone onto a standard that Twitter controls


      The only way this has an upside for Twitter is if its licensed and part of that license will be a provision that server operators must conform to Twitter’s censorship rules if they want to have inter-connectivity with other servers.

    2. Yep. The rise of all those interoperable platforms (as opposed to Twitter’s so called ‘platform’ that actually is not) also risks making it obvious that Twitter is not always and everywhere an interactive computer service, but is often an information content provider and therefore not deserving of any section 230 liability protections when they act as such.

    3. This is exactly what I came here to say. All of this technology exists now, but no one wants to use it… mainly because no established company can monetize (or has figured out how) a federated internet.

      I agree that Jack sees the train coming and he’s trying to position himself as possibly the sole license holder for his new proprietary “open” technology.

      1. It’s also going to kill it. If one were to remove the system behind likes and retweets the platform would crumble. Since both of these things generate tons of cash twitter picks and promotes what will make them the most. This is what has fueled interactions on all of these platforms. The reason why these technologies took off was because of the content delivered and the manner it was delivered got people hooked.

        Dorsey knows that regardless of who sets up an open Twitter his is going to win. It’s already too established of a brand and will have more money behind it than any other. The other thing to take note of is that Dorsey knows times are changing. We see others like Youtube attempting to do the same. If you have a decentralized system who faces liability?

        Be very cautious of decentralized internet because if it involves you transmitting the data on it then you aren’t protected. If you have to transmit in order to watch then you’re responsible for transmitting whatever it is you watched. You’re no longer just watching and reading in that instance and any methods you use to cover your tracks could be construed as you knowingly participating in wrongdoing. It’s all a legal grey area and almost none of this has been tested. If you’re a test case then expect to go broke.

  2. Who but dipshit media people actually use Twatter?

    1. It’s the MySpace of the future.

    2. It’s primary usefulness is that it gives us a window to see what these scumbags really think about America and normal Americans.

      Like for example, when Welchie Boy goes on there and openly daydreams about massacring everyone he disagrees with.

  3. still never tweeted. or been on fb

    1. And I can assure you that you are a better man for that.

      1. i would use twitter to compliment people “i liked you in that thing i saw you in” but i hear that’s not what it’s for

  4. I’ve been describing this sort of idea since 2008, in the context of arguing why Facebook’s sustainable competitive advantage may not be indefinitely sustainable. An open system or protocol could allow users to interact and access the same content with the platform or app of their choice. You could even have something that imports all of your content from Facebook. Like exporting your iTunes library to some other player, back when streaming hadn’t taken over yet.

    1. We’ve had such systems and protocols since the 1980s. The latest iteration is ActivityPub and the Fediverse.

      They are not catching on yet because of Section 230 and a history of internet neutrality, which allowed Twitter, YouTube, and Facebook to monopolize markets.

      1. They’re not catching on because they’re boring and a replacement for something that already exists. Millions are spent discovering and implementing ways to make the big tech platforms interesting and addicting. The majority of those that use social media aren’t having censorship hinder their experience. Those that are then flock to the alternatives and the alternatives look like a dumpster filled with diapers from the elderly that’s on fire. As long as they’re there then the majority will steer clear.
        Youtube monopolized the market by going into the red. You have to lose money in order to compete. If you look at those that went to Bitchute you’ll see a small fraction of the numbers they have or would get on youtube. An alternative only works if something dies first and people may not switch, but rather move on to something entirely different like going outside.

        1. An alternative only works if something dies first and people may not switch,

          I agree. And hopefully killing Section 230 and Internet neutrality will cause Twitter, Facebook, and YouTube to die. It’s long overdue.

  5. From a user perspective, I can see this as great.

    From a buisness perspective… the fuck? Killing the goose that lays a golden egg? Is he high?

  6. I must be missing something here. Email as protocol works fine because everything is private and targeted. Everything I send has specific destinations. Everything I receive came from specific sources who specifically targeted me.

    Twitter and Facebook (the main Facebook, NOT Facebook Messenger) are public. If I post something on either of them, everybody can see it (ignoring the privacy controls for the moment). This works because the data is not “sent” to anybody but Facebook and Twitter. Their client apps only display what people have selected, but when Joe Q Blow reads my post or tweet, he is seeing what he retrieved from Facebook or Twitter, not what was sent to him.

    Email works not just because the protocols are open and public, but because the messages are transmitted once between specific endpoints. I could simulate Facebook and Twitter by having an IMAP service with no credentials so anybody could read it, but Facebook and Twitter could never simulate email because they never send those messages anywhere; they remain on their central servers. This is similar to individualism simulating socialism with contracts, but socialism not even tolerating individualism, let alone simulate it.

    I don’t believe Twitter or Facebook could convert to protocol-based services. All we’d end up with is isolated Balkanized mini-Facebooks and mini-Twitters.

    I must be missing something.

    1. I don’t think you are because that is exactly what I thought. The whole point of Twitter and Facebook is for things to be public and for you to interact with a web of contacts. I don’t see how that could ever operate like email.

      1. I imagine it is going to be like private messaging. You can send emails to multiple people. Picture downloading some kind of software that runs off of files on your computer and everything is peer to peer. It’s not necessary and everyone is going to tire of it in a month or so. People aren’t on social media because their friends are interesting. They can text for that. They’re on it because the platform shows them things they like based on what they have clicked before and other data they have mined. Some may find it useful, but there are already other methods of doing this and since it will completely lack the narcissism factor that social media provides most will find it boring.

    2. “All we’d end up with is isolated Balkanized mini-Facebooks and mini-Twitters.”

      If by Balkanized, you actually mean end user controlled then you are correct.

      You could easily configure a platform such that the sender decides who can view the message (e.g. specified individuals, groups, or ‘anyone.’) And other users could choose what messages they wish to view – individuals, groups (public or private) or ‘anyone.’

      Ideally, in such a system users would also be able to select and utilize their own end encryption, thus limiting the government (or anyone else’s) ability to access back doors.

      You see this as a problem?

      1. I don’t think it is a problem. I just don’t think it fills the needs people use social media to fulfill.

        1. I could see how it could exclude people who are too stupid or too lazy to do any sort of basic account configuration, but that’s the sort of detail that the winning platform(s) would do best.

          Think of it like a giant set of personals pages, but one where you can sort and select the posts that you view.

          This could end up supplanting all sorts of email and facebook groups for clubs, groups, and other social organizations. Same goes for more dedicated sites like pinterest. Again – done right – this sort of service could easily allow you to do much of what pinterest does, but with much more user definition and user control (assuming it was willing to handle the data/bandwidth.)

      2. Like John says, that isn’t Facebook or Twitter any more.

        I don’t see any way to simulate Facebook or Twitter with open protocols.

        Facebook or Twitter could publish their interface specs and let anyone connect, with proper credentials. But that would be merely changing the client apps. It would have nothing to do with multiple Facebook and Twitter servers, which is what email does.

        1. “that isn’t Facebook or Twitter any more.”

          To the extent that Facebook and Twitter are actually curating and controlling the content you view, yes I’d have to agree with that statement.

    3. Something along these lines (assuming its not a Twitter licensed protocol)

      A user joins a server instance running this protocol and to all servers running this protocol, that user is known as KeyboardActivist@VirtueSignal

      Our brave soy-boy can go out and do battle with imaginary nazis on any server instance that allows VirtueSignal instance users to post. Our couch commando can also follow users on other instances or be followed by users on other instances.

      The owner of the VirtueSignal instance is probably the sort that see’s nazis everywhere and so this instance will certainly have an extensive list of instances that are not permitted to post on VirtueSignal.

      The VirtualSignal instance likewise being host to a bunch of rage flinging morons is probably also gonna be blocked by a lot of instances, and so our tide-pod muncher won’t be allowed to post there.

      An intrepid user, one who doesn’t want to be subject to blanket bans simply because the instance they signed up on is disfavored could also just as easily host their own server instance consisting of just a single user. Then they could post and follow users throughout the network. That user could then still have access to a plethora of other instances, even if they get banned from a handful.

      1. Blue check brigade hardest hit.

        How can you have a Vision of the Anointed if everyone can do the anointing???

      2. That is a great explanation. Thanks. And it makes sense. But Twitter is one of the worst companies on earth. I find it hard to believe they would do that.

    4. There’s ways around it.

      Think of how Tor and things like that work, where there is no real “central repository”, but different people hosting different things, validity verified by hashes and such. There is no central control (and thus no central ability to say “you can’t post”), but you can still “see” everything.

      Unless you’re using a client that restricts what you can see, that is. Your client might add filters/censoring, but other clients will just let everything through.

      Another alternative is to have an expectation of servers that host, but the servers then talk to each other, but that when anyone makes a querry of any server, it then digs up things from every server. So basically what Twitter is already doing, but instead of it owning the servers, it just defines the protocols for how servers (whoever they’re owned by) talk to each other.

      And probably other ways that I can’t think of.

      It’s a fun problem, but not an insurmountable problem.

      1. The big trick is how do brands, or those with some sort of notoriety prove “I am me” without a central authority.

        In that same line of thinking, Brands and well known people aren’t going to use a protocol that doesn’t afford them the means of preventing others from impersonating them.

        I suppose the answer would be servers using the protocol that specialize in confirming the identity of users. That would likely be something Twitter would continue to do. If you saw MegaBrandX@Twitter you’d know its the legit account, and MegaBrandX@SomewhereElse is an impostor.

        So Twitter goes on to specialize in verified users where brands live and send them sweet sweet ad money.
        Non-verified users that stay on twitter have to follow the SJW CoC. Or find a federated server and follow their favorite verified users on twitter that way. AI will be used to make a ‘Twitter safe space’ blocking problematic tweets sent by federated users to verified users.

        1. I mean, Facebook started off with “confirming the identity of users” back when it required a .edu e-mail address, but that kind of dropped on the wayside as it got bigger. Twitter confirms some user identities, but most users aren’t verified.

          So yeah, some server might try to establish itself as the “verifier”, but it might also be that folks that are worried set-up their own server. You know it’s the the real POTUS, ’cause it’s POTUS@gov or some-such.

          Alternatively (and this is probably a bad option) governments could really get involved and tie your universal “social media” account to a national ID or some-such. Kill anonymity dead.

          1. Twitter certainly doesn’t want to verify everybody, or even most users.
            What I think Twitter really wants is to have the money making blue checkmarks be as popular and profitable as ever, without Twitter also having to expend resources moderating or mediating pitched battles between ten million different 20-follower twitter accounts.

            Since anyone can set up a server and call it what they want, I suspect most brands would prefer Twitter verify that they are the real whoever just to save themselves the trouble of proving that a server called ‘RealMegaBrandX’ is legit. If anything they would set up such a server, lock it down and redirect to Twitter.

            1. I think Twitter wants to not be “public” and I think they want you to host the files and not them. Twitter is available around the world so if one country has a rule Twitter has to follow or they could get blocked. (Russia fined them 46 bucks once and failure to comply could lead to blocking.) Then the UK makes some law regarding nonconsensual adult images and then Twitter has to follow it. Then the US could amend CDA 230 and now Twitter is liable for this and that and they can’t properly moderate so instead of just throwing it away they try this method out. Then you’re liable because you’ve just taken on what Twitter didn’t want to bother with.

        2. The same way corporations verify who they are on Websites and Email. Through the existing DNS system on the internet and certificates. It wouldn’t be hard to extend that schema to Twitter accounts.

    5. We’ve had such systems for decades and they work fine: USENET and Fediverse are common examples.

      Subscriptions and distributions are federated and aggregate based on user demand and site policy.

  7. “There is no history of taking a proprietary, closed platform and turning it into a broad, open protocol” seems to me to be directly contradicted by “The email/protocol example is the way much of the internet used to be in the early days”.

    The idea may well fail but it won’t be due to lack of precedents.

    1. Internet protocols were open and distributed from the start.

      Companies like Twitter and Facebook basically replicated these open systems with proprietary ones and monopolized their markets, in large part because net neutrality and immunity from law suits created the conditions that allowed that to happen.

  8. I might have to check out this enticing new offer from Twitter! It sounds almost as good as an enticing new offer from a clown driving an unmarked white van with a “free candy” sign spray-painted on the side! I’d be a fool not to!

    1. You’d think free candy from a van would be of low-quality, but that clown has some good shit. Whole candy bars. Not those mini ones.

  9. You can bring your own email address into Gmail if you want.

    Um…no. Unless I’m missing something, that’s a pretty bad error on Masnick’s part.

    1. Yes you can: Gmail will host your domain for you, and they will also let you use their front end for an existing IMAP server.

  10. The question is, how does Twitter Inc make money off of an ‘open protocol’ version of itself?

    Right now, all ‘Tweeting’ goes through Twitter’s servers & they can data-mine that to sell ads. I don’t know if they are profitable, but they at least have an argument that they can be at some point.

    SMTP, HTTP, IMAP and other major ‘open’ internet protocols were all developed by research institutions, or standardized by industry… Which lead to their use by multiple competing companies producing both clients and servers as part of larger commercial products….

    What’s in it for Twitter if they become a protocol? They won’t be able to data-mine effectively (Because they won’t actually own the data, the way they do when people log into to use their proprietary service)…. So the ad-based ‘Google Style’ model is out…

    That leaves software licensing – but who would buy a ‘Twitter Server’ the way companies buy Microsoft Exchange Server for email? And how would twitter survive the eventual open-source reverse-engineering of their new ‘protocol’

    The whole ‘block-chain’ idea has further problems – how do you ‘reward’ people for participating – would anyone ‘mine’ tweets to keep the network working? That idea really only works for the application it is used for now: speculative investment (Dutch tulip style) and as a platform for cybercrime/money-laundering….

    1. “The question is, how does Twitter Inc make money off of an ‘open protocol’ version of itself?”

      If you have to download a program the program could mine crypto while you use it. If that’s the case they won’t have to worry about mining data for ads as well as fearing advertisers leaving the platform.

  11. If you expect anything remotely freedom oriented from Dorsey, you’re a fool.

  12. “There is no history of taking a proprietary, closed platform and turning it into a broad, open protocol.”

    It really depends on how you define “open” but Microsoft ActiveSync Protocol was an exclusive Proprietary protocol used exclusively by Microsoft Email servers to Microsoft Email clients on Windows Mobile phones. It was a direct competitor to Blackberry and GOOD technologies mail syncing, and was seen as a more comprehensive solution than IMAP.

    When the iPhone first launched in 2007, it only could do IMAP and was at a serious competitive disadvantage when it came to corporate email syncing. By the second Gen Apple had licensed ActiveSync from Microsoft.

    Nowadays if you have a smartphone syncing email from a server (Android, GMail, Apple, Lotus, corporate, whatever) there is a very strong likelihood you are using that ActiveSync protocol. It no longer is licensed in the same way. Microsoft makes you pay royalties for use of the IP so it isn’t open source. But the code is widely available and can be implemented by anyone and that is why there are hundreds of applications in app stores that do ActiveSync messaging. And hundreds of thousands of servers that support it.

    In the case of twitter, making it an open protocol is the easy part. It is determining who pays for the storage.

Please to post comments

Comments are closed.