After mulling over my bullet points, it occurred to me that the network problems I was dealing with—background cloud sync, editing across multiple devices, real-time collaboration, offline support, and reconciliation of distant or conflicting revisions—were all pointing to the same question: was it possible to design a system where any two revisions of the same document could be merged deterministically and sensibly without requiring user intervention?

It’s what happened after sync that was troubling. On encountering a merge conflict, you’d be thrown into a busy conversation between the network, model, persistence, and UI layers just to get back into a consistent state. The data couldn’t be left alone to live its peaceful, functional life: every concurrent edit immediately became a cross-architectural matter.

I kept several questions in mind while doing my analysis. Could a given technique be generalized to arbitrary and novel data types? Did the technique pass the PhD Test? And was it possible to use the technique in an architecture with smart clients and dumb servers?

Concurrent edits are sibling branches. Subtrees are runs of characters. By the nature of reverse timestamp+UUID sort, sibling subtrees are sorted in the order of their head operations.

This is the underlying premise of the Causal Tree. In contrast to all the other CRDTs I’d been looking into, the design presented in Victor Grishchenko’s brilliant paper was simultaneously clean, performant, and consequential. Instead of dense layers of theory and labyrinthine data structures, everything was centered around the idea of atomic, immutable, metadata-tagged, and causally-linked operations, stored in low-level data structures and directly usable as the data they represented.

I’m going to be calling this new breed of CRDTs operational replicated data types—partly to avoid confusion with the exiting term “operation-based CRDTs” (or CmRDTs), and partly because “replicated data type” (RDT) seems to be gaining popularity over “CRDT” and the term can be expanded to “ORDT” without impinging on any existing terminology.

Much like Causal Trees, ORDTs are assembled out of atomic, immutable, uniquely-identified and timestamped “operations” which are arranged in a basic container structure. (For clarity, I’m going to be referring to this container as the structured log of the ORDT.) Each operation represents an atomic change to the data while simultaneously functioning as the unit of data resultant from that action. This crucial event–data duality means that an ORDT can be understood as either a conventional data structure in which each unit of data has been augmented with event metadata; or alternatively, as an event log of atomic actions ordered to resemble its output data structure for ease of execution

To implement a custom data type as a CT, you first have to “atomize” it, or decompose it into a set of basic operations, then figure out how to link those operations such that a mostly linear traversal of the CT will produce your output data. (In other words, make the structure analogous to a one- or two-pass parsable format.)

OT and CRDT papers often cite 50ms as the threshold at which people start to notice latency in their text editors. Therefore, any code we might want to run on a CT—including merge, initialization, and serialization/deserialization—has to fall within this range. Except for trivial cases, this precludes O(n2) or slower complexity: a 10,000 word article at 0.01ms per character would take 7 hours to process! The essential CT functions have to be O(nlogn) at the very worst.

Of course, CRDTs aren’t without their difficulties. For instance, a CRDT-based document will always be “live”, even when offline. If a user inadvertently revises the same CRDT-based document on two offline devices, they won’t see the familiar pick-a-revision dialog on reconnection: both documents will happily merge and retain any duplicate changes. (With ORDTs, this can be fixed after the fact by filtering changes by device, but the user will still have to learn to treat their documents with a bit more caution.) In fully decentralized contexts, malicious users will have a lot of power to irrevocably screw up the data without any possibility of a rollback, and encryption schemes, permission models, and custom protocols may have to be deployed to guard against this. In terms of performance and storage, CRDTs contain a lot of metadata and require smart and performant peers, whereas centralized architectures are inherently more resource-efficient and only demand the bare minimum of their clients. You’d be hard-pressed to use CRDTs in data-heavy scenarios such as screen sharing or video editing. You also won’t necessarily be able to layer them on top of existing infrastructure without significant refactoring.

Perhaps a CRDT-based text editor will never quite be as fast or as bandwidth-efficient as Google Docs, for such is the power of centralization. But in exchange for a totally decentralized computing future? A world full of devices that control their own data and freely collaborate with one another? Data-centric code that’s entirely free from network concerns? I’d say: it’s surely worth a shot!

With the comprehensive application of Artificial Intelligence into the creation and post production of images, it seems questionable if the resulting visualisations can still be considered ‘photographs’ in a classical sense – drawing with light. Automation has been part of the popular strain of photography since its inception, but even the amateurs with only basic knowledge of the craft could understand themselves as author of their images. We state a legitimation crisis for the current usage of the term. This paper is an invitation to consider Synthography as a term for a new genre for image production based on AI, observing the current occurrence and implementation in consumer cameras and post-production.

In fact, more people are technologists than ever before, insofar as a “technologist” can be defined as someone inventing, implementing or repurposing technology. In particular, the personal computer has allowed anyone to live in the unbounded wilderness of the internet as they please. Anyone can build highly specific corners of cyberspace and quickly invent digital tools, changing their own and others’ technological realities. “Technologist” is a common identity that many different people occupy, and anyone can occupy. Yet the public perceptions of a “technologist” still constitute a very narrow image.

A technologist makes reason out of the messiness of the world, leverages their understanding to envision a different reality, and builds a pathway to make their vision happen. All three of these endeavors—to try to understand the world, to imagine something different, and to build something that fulfills that vision—are deeply human.

Humans are continually distilling and organizing reality into representations and models—to varying degrees of accuracy and implicitness—that we can understand and navigate. Our intelligence involves making models of all aspects of our realities: models of the climate, models of each other’s minds, models of fluid dynamics.

We are an unprecedentedly self-augmenting species, with a fundamental drive to organize, imagine, construct and exercise our will in the world. And we can measure our technological success on the basis of how much they increase our humanity. What we need is a vision for that humanity, and to enact this vision. What do we, humans, want to become?

As a general public, we can collectively hold technologists to a higher ethical standard, as their work has important human consequences for us all. We must begin to think of them as doing deeply human work, intervening in our present realities and forging our futures. Choosing how best to model the world, impressing their will on it, and us. We must insist that they understand their role as augmenting and modifying humanity, and are responsible for the implications. Collective societal expectations are powerful; if we don’t, they won’t.

Nevertheless, the interface pushes back with its prescribed methodologies, workflows, and limitations. Interface and artist are an antagonistic pair. Perhaps the best description of the polemic between the two is one of productive cannibalism. Just as the interface evolves under the pressure of innovation to accommodate new pragmatic uses, the artists’ will continue to deconstruct and push its aesthetic and behavioral properties to their limits.

Mainframes were followed by PCs, and then the web, and then smartphones. Each of these new models started out looking limited and insignificant, but each of them unlocked a new market that was so much bigger that it pulled in all of the investment, innovation and company creation and so grew to overtake the old one. Meanwhile, the old models didn’t go away, and neither, mostly, did the companies that had been created by them. Mainframes are still a big business and so is IBM; PCs are still a big business and so is Microsoft. But they don’t set the agenda anymore - no-one is afraid of them.

We’ve spent the last few decades getting to the point that we can now give everyone on earth a cheap, reliable, easy-to-use pocket computer with access to a global information network. But so far, though over 4bn people have one of these things, we’ve only just scratched the surface of what we can do with them.

There’s an old saying that the first fifty years of the car industry were about creating car companies and working out what cars should look like, and the second fifty years were about what happened once everyone had a car - they were about McDonalds and Walmart, suburbs and the remaking of the world around the car, for good and of course bad. The innovation in cars became everything around the car. One could suggest the same today about smartphones - now the innovation comes from everything else that happens around them.

Although Amazon has clocked staggering growth, it generates meager profits, choosing to price below-cost and expand widely instead. Through this strategy, the company has positioned itself at the center of e-commerce and now serves as essential infrastructure for a host of other businesses that depend upon it. Elements of the firm’s structure and conduct pose anticompetitive concerns—yet it has escaped antitrust scrutiny.

This Note argues that the current framework in antitrust—specifically its pegging competition to “consumer welfare,” defined as short-term price effects—is unequipped to capture the architecture of market power in the modern economy. We cannot cognize the potential harms to competition posed by Amazon’s dominance if we measure competition primarily through price and output. Specifically, current doctrine underappreciates the risk of predatory pricing and how integration across distinct business lines may prove anticompetitive.

These concerns are heightened in the context of online platforms for two reasons. First, the economics of platform markets create incentives for a company to pursue growth over profits, a strategy that investors have rewarded. Under these conditions, predatory pricing becomes highly rational—even as existing doctrine treats it as irrational and therefore implausible.

Second, because online platforms serve as critical intermediaries, integrating across business lines positions these platforms to control the essential infrastructure on which their rivals depend. This dual role also enables a platform to exploit information collected on companies using its services to undermine them as competitors.

there are dozens of companies that remix some combination of lists, tables, charts, tasks, notes, light-weight databases, forms, and some kind of collaboration, chat or information-sharing. All of these things are unbundling and rebundling spreadsheets, email and file shares.

LinkedIn tried to take the flat, dumb address book and turn it into both structured flow and a network of sorts. But by doing that for everyone, it has the same problem as a spreadsheet, file share or email - it’s a flat, lowest-common-denominator canvas that doesn’t capture the flows that many particular professions or tasks need.

There’s clearly a point in the life of any company where you should move from the list you made in a spreadsheet to the richer tools you can make in coolproductivityapp.io. But when that tool is managing a thousand people, you might want to move it into a dedicated service. After all, even Craigslist started as an actual email list and ended up moving to a database. But then, at a certain point, if that task is specific to your company and central to what you do, you might well end up unbundling Salesforce or SAP or whatever that vertical is and go back to the beginning.

every application category is getting rebuilt as a SaaS web application, allowing continuous development, deployment, version tracking and collaboration. As Frame.io (video!) and OnShape (3D CAD!) show, there’s almost no native PC application that can’t be rebuilt on the web. In parallel, everything now has to be native to collaboration, and so the model of a binary file saved to a file share will generally go away over time

an entire generation now grew up after the web, and grew up with smartphones, and assumes without question that every part of their life can be done with a smartphone. In 1999 hiring ‘roughnecks’ in a mobile app would have sounded absurd - now it sounds absurd if you’re not. And that means that a lot of tasks will get shifted into software that were never really in software at all before.

The internet is a firehose. I don’t, myself, have 351 thousand unread emails, but when anyone can publish and connecting and sharing is free and frictionless, then there is always far more than we can possibly read. So how do we engage with that?

So your feed becomes a sample - an informed guess of the posts you might like most. This has always been a paradox of Facebook product - half the engineers work on adding stuff to your feed and the other half on taking stuff out. Snap proposed a different model - that if everything disappears after 24 hours then there’s less pressure to be great but also less pressure to read everything. You can let go. Tiktok takes this a step further - the feed is infinite, and there’s no pressure to get to the end, but also no signal to stop swiping. You replace pressure with addiction.

Another approach is to try to move the messages. Slack took emails from robots (support tickets, Salesforce updates) and moved them into channels, but now you have 50 channels full of unread messages instead of one inbox full of unread messages.

Screenshots are the PDFs of the smartphone. You pull something into physical space, sever all its links and metadata, and own it yourself.

Email newsletters look a little like this as well. I think a big part of the reason that people seem readier to pay for a blog post by email than a blog post on a web page is that somehow an email feels like a tangible, almost physical object - it might be part of that vast compost heap of unread emails, but at least it’s something that you have, and can come back to. This is also part of the resurgence of vinyl, and even audio cassettes.

The film-camera industry peaked at 80bn consumer photos a year, but today that number is well into the trillions, as I wrote here. That’s probably why people keep making camera apps with built-in constraints, but it also prompts a comparison with this summer’s NFT frenzy. Can digital objects have value, and can a signature add scarcity to a JPEG - can it make it individual?

there are now close to 5bn people with a smartphone, and all of us are online and saying and doing things, and you will never be able to read everything ever again. There’s an old line that Erasmus, in the 15th century, was the last person to have read everything - every book that there was - which might not have been literally possible but which was at least conceivable. Yahoo tried to read everything too - it tried to build a manually curated index of the entire internet that reached 3.2m sites before the absurdity of the project became overwhelming. This was Borges’s 1:1 scale map made real. So, we keep building tools, but also we let go. That’s part of the progression - Arts and Crafts was a reaction against what became the machine age, but Bauhaus and Futurism embraced it. If the ‘metaverse’ means anything, it reflects that we have all grown up with this now, and we’re looking at ways to absorb it, internalise it and reflect it in our lives and in popular culture - to take ownership of it. When software eats the world, it’s not software anymore.

Generalists can pursue broader themes, questions, and lenses which, across their interests give them a deep perspective from breadth.For example, a specialist is someone who is obsessed with chess and spends their waking hours practicing, playing, and studying.A generalist is someone who is obsessed with the idea of game-play, and has researched and gone deep on sports, childhood psychology, board games, and philosophy.

Embracing being a coordinate on the map for a point in time is about allowing yourself to be seen as something specific. Generalists can feel trapped by that but the truth is being specific, and being on the map for others is a way of being in service. If you never pin yourself down (just for a time) you miss the benefits of being connected or in service.

This is all build-up to my proposal for what Musk — or any other bidder for Twitter, for that matter — ought to do with a newly private Twitter. First, Twitter’s current fully integrated model is a financial failure. Second, Twitter’s social graph is extremely valuable. Third, Twitter’s cultural impact is very large, and very controversial. Given this, Musk (who I will use as a stand-in for any future CEO of Twitter) should start by splitting Twitter into two companies. One company would be the core Twitter service, including the social graph. The other company would be all of the Twitter apps and the advertising business.

TwitterServiceCo would open up its API to any other company that might be interested in building their own client experience; each company would: Pay for the right to get access to the Twitter service and social graph. Monetize in whatever way they see fit (i.e. they could pursue a subscription model). Implement their own moderation policy. This last point would cut a whole host of Gordian Knots:

A truly open TwitterServiceCo has the potential to be a new protocol for the Internet — the notifications and identity protocol; unlike every other protocol, though, this one would be owned by a private company. That would be insanely valuable, but it is a value that will never be realized as long as Twitter is a public company led by a weak CEO and ineffective board driving an integrated business predicated on a business model that doesn’t work. Twitter’s Reluctance