I've been thinking for a long time about atmospheric publishing, from our work on to the ecosystem emerging around , and how we can beat Substack at their own game.
Last month I did a talk at on Atmospheric Stacks — a new approach to building cohesive full-stack social experiences collectively in the Atmosphere.
This post is an adaptation/expansion, still a little messy, but you can watch the talk or skim the slides below for the TL;DR!
Also below: a couple addenda, still WIP, not included in the talk…
recording of my talk from 7.7.26
slides from my talk!
Intro
This talk is about how the Atmosphere can compete with successful incumbent social apps — not by recreating them app for app, but by, as an ecosystem, decomposing them and replacing them, collectively!
Replacing closed, proprietary stacks with open stacks.
I want to use the example of Leaflet, Standard Site, and the emergent longform social publishing ecosystem as an ongoing case study, and how we can continue to advance our shared mission to supplant Substack.
The problems we need to solve — together
Successful social apps typically combine:
a bunch of features that work well together
business model / incentive model that gets people coming back
good product + good incentives = sticky network dynamics
The big issue: we don't just need feature parity. We need to solve the distribution problem, the social ecosystem problem, the "creators need viable businesses" problem!
We also need to face reality: incumbents have huge advantages. Resources, bigger teams, more $, and most critically, entrenched network effects.
So, it's an incredibly tall order for small emergent builders to replace a big successful app-slash-social-network like Substack — let alone a mega-platform like YouTube or Instagram — alone.
But…what if we do it together???
This talk is about what that means, realistically as well as aspirationally. What can we do today, where can we go, and how can we collaborate?
Atmospheric stacks: what's different?
In the Atmosphere, composability is a superpower.
(Just one that must be wielded carefully, so as to not…y'know, end up creating experiences that suck.)
Yes, this means more complexity and coordination costs. But the alternative is no coordination / interop at all. So, we figure it out!
The upside is collectively creating something more flexible, more powerful, more alive, and more resilient.
Something better for users and better for (non-monopolist) builders too!
old stacks: skyscrapers — all the parts in one place, big but kinda boring
new stacks: rainforests — many layers, diverse ecosystem, symbiotic, vibrant
New, Atmospheric stacks might look more like…
several popular tools with an overlapping set of core functionality
surrounded by a rich ecosystem of other tools and services with more specialized functions, experimental interfaces, and more
ways for people to put these things together into cohesive experiences
Standard Site and an emergent stack for social publishing
As it happens we've got a great case study of an emergent Atmospheric stack for longform social publishing!
I'm going to talk about Leaflet and Standard Site and the wider atmospheric publishing ecosystem, vs. Substack, because it's a good example and I know it firsthand, and the "stack" thing in the name, but later I'll point at how a similar approach might work in other areas.
There's already a pretty complex ecosystem taking shape here, with lots of dynamics and potential for interop and other exciting things.
In some places we see coopetition, like those of us making Standard Site publishing tools. In other areas it's pure win-win symbiotic dynamics, like publishing + reading + social discovery tools that can interoperate. And lots of untapped potential yet to explore, like ways we can merge e.g. publishing and video tools into powerful experiences.
So, we're building new stacks here! Collectively assembled stacks! This one is for social publishing — not just web publishing, like WordPress, but publishing with a social network, which today is mostly: Substack.
What's Substack's…stack?
So, what makes Substack unique here? What's Substack's stack?
It's not one thing, but many! And although there are a ton of other publishing tools out there, this stack is a pretty unique set of things:
great design
blog (web) + newsletter (email)
easy to customize
easy to monetize
free to publish
social features that are nice to use
social network, actual follow graph, newsletters + short posts
strong cultural footprint, lots of use across many verticals
strong brand, people are familiar with the model
all in one platform
long tail of advanced features (podcasts, livestreaming, guest posts, multi-author publications, etc.)
Sure, lots of competitors do, say, half of these things quite well! But right now there are basically none that match the full stack.
If this is too many things for any one of us to go head to head on all axes by ourselves…how can we tackle it collectively?
Oh yeah and it's a whole damn social network
First, just to hammer on one important point a bit more…
Key thing with Substack, and similar for other big social apps: it's not just a way to publish stuff, it's a successful social network!
Same for YouTube; you post there not just for the free hosting (though that's nice too) but for the distribution.
Substack's actually somewhere in the middle; most successful writers on the platform are coming in with their own audiences, and it's of course in Substack's interest to overstate the extra reach they get people…but it's not nothing, and exaggerated is very different from fake.
So, the stack we're taking on is a bundle of [publishing app] AND [social app], which, hey, is something atproto can be great for!
But the replacement, of course, looks rather different. For one, the openness of the ecosystem means you structurally can't lock up all the data and keep the network to yourself. Instead…
So, what's the new Atmospheric publishing stack?
How do we replace Substack? With a network of our own!
The replacement for Substack is not Bluesky, or Leaflet, or any single app…it's the Atmosphere!
The new counter-stack is an ecosystem!
It's apps like Leaflet and pckt and Offprint…
It's the wider Standard.site community…
It's the entire Atmosphere…
It's other channels like email and incumbent social networks
It's the whole internet, if we can bring them onboard!
We can think of this as a many-layered circle of readers, writers, publishing tools, discovery surfaces, communities, utility / glue apps, and more:
At the center of the circle, we need to get creators active and give them powerful tools…
At the edges of the circle, we draw people in with cool new ways to find interesting things and read and engage with communities…
Activity ripples inward, activity ripples outward; we weave connections between layers, like some mycelial network…
Okay, let's get more specific!
How does my experience change?
What's different in a user's experience using the Atmospheric stack for longform publishing, instead of Substack?
In this world…
It's possible for your experience to live within one app, but it doesn't have to
People move between different apps, or use many, for different aspects of what they need
Not everyone needs everything — there's more room for custom, personalized experiences, for both readers and writers
There's a mix of common layers many people use, and more unique ones that cater to long tail niche needs — all in the same network
Let's contrast some flows in Substack vs. the new Atmospheric stack:
Finding things to read
Substack: social media or within the Substack app
Atmospheric stack: social media, blogs, other pubs, custom feeds, reader tools x social graph, cross-app recs…
Following / subscribing
Substack: follow or subscribe in Substack
Atmospheric stack: subscribe with atproto; email is secondary…and more like RSS, read and follow your subs in any app you like
Participating in convo
Substack: comments on posts; 'notes'
Atmospheric stack: comments section, Bluesky, Blacksky, future private communities, other social apps…
Starting a publication from scratch
Substack: you gotta do it on Substack
Atmospheric stack: on Leaflet, Offprint, pckt…or your own custom site!
Building a custom API integration
Substack: lol you can't
Atmospheric stack: sure do whatever you like!
Sketching the new stack: parts of the stack we need to replace
Let's look again at the parts of Substack's stack — and lay out some specific things we might try with Leaflet and the Standard Site ecosystem:
blog + email newsletter:
many alternatives, including atmospheric ones like Leaflet and Offprint — one of the easy parts
flexible and customizable:
also many alternatives; hard to do well but still not the hard part
Leaflet: themes, custom domains, lots of block types etc.
clear way to make money:
some alternatives, like Ghost and Beehiiv; harder to do well; Atmospheric alternatives are still WIP…but soon!
can't overstate the importance of "HELP PEOPLE MAKE MONEY"
interesting space here; Substack is actually pretty limited in this respect…much we can try besides the obvious paid subs model
monetization + social network = lots of cool potential
nice social features:
varying degrees…Leaflet has comments, many apps have easy social sharing and recommendations (Standard Site)
lots to play with here, but STILL not the hard part! because that's more about…
strong network effects:
critical for community, discovery, keeping people in the ecosystem — the Atmosphere has potential, but too Bluesky-heavy right now, far from realized…
lots we can do with e.g. cross-publication recommendations, social graph aware discovery, and communities
free to publish:
basically no one else does this
so…can we reduce the importance of free email? as a longer term goal, we can aim to mitigate that moat
strong cultural footprint:
the Atmosphere has a fairly diverse user base, but not enough; many areas where we need more big accounts or 'anchor tenants'
need more verticals + more recruitment efforts…and of course this goes hand in hand with stronger incentive models!
long tail of advanced features:
things like podcasting, livestreaming, guest posts and multi-author publications, and beyond: less core features for everyone, but many things important for various creator niches and larger orgs
great space for atmospheric opportunity; lots of cool integrations and interop if we team up across apps / lexicons
strong brand / familiarity:
the Atmosphere's in its infancy; some people know about Bluesky but basically no one knows about the wider ecosystem
people do understand blogging, and the basic paid sub model, not only from Substack but Patreon, the New York Times, and so on, so we have some simple conceptual anchor points
all in one easy to use platform:
this flipside of making interop easy: great UX is still hard
we have a ways to go to make Atmospheric stacks function legibly to end users as unified cohesive experiences, and not just kludgy assemblages (do NOT mention self-hosting or cross-protocol bridging when pitching people on a solution here lol)
Lots of cool stuff already!
As I said, it's still early and we have a ways to go here. The good news is there are actually a lot of cool pieces coming together already!
Some things are pretty serious and promising already; others are nascent or prototype phase but point at big potential later. We have…
search tools like pub search
bluesky's initial standard.site link card integration
Wordpress has v1 of a plugin for atmospheric publishing
Sequoia for publising a static site
reader apps like Standard Reader and Skyreader
And of course the big three atmospheric publishing apps: Leaflet, pckt, and Offprint. All of us are doing some interesting things!
Recent integration with pckt + Marque, the atproto domain registrar, is a cool example. We'd like to make this happen in Leaflet too. Cool experiments that work can become patterns.
Standard Reader is a great example on the reader side, launched pretty recently and has shipped a bunch of new things quickly, including some more experimental stuff like a way to publish 'collections' with a more magazine-like reader experience, still compatible with Standard Site…
And lots more to explore…
There's lots to explore at all the layers of the stack, and I'm especially excited for more weird cool atmospheric stuff, things only really possible in this kind of open interoperable ecosystem.
We're at kind of the first stage of just getting close to having the basics in place for solid publishing tools / platform…
Pretty soon I think we can enter the second phase where we start to get a bit weird and see what unique things we can do together!
Weird cool valuable publishing things:
embeds with any app, universal backlinks, etc. etc.
Weird cool valuable social things:
deeper Bluesky integration, mentioning anything, more ways to comment and engage, finding things via friends / social graph, etc.
Weird cool valuable monetization things:
one-off purchases and other $ models
…and things that combine all the above!
like gating certain content w/ social incentives e.g. free or discount if you share, if > N ppl sub, etc.
A huge thing on the way is permissioned data and groups / communities, lots we'll be able to do there! Like we could make it so paid members of a Leaflet publication get access to private communities, and those communities can themselves exist across many places, not locked into a single app.
We can make this work not only with Bluesky, though that'll be one good entry point. We can do e.g. with communities on Blacksky or Acorn as well. Big publishers like The Verge can do it themselves, too!
To succeed here we've got our work cut out. We've gotta follow the money, build for communities, solve discovery…
We've gotta grow Standard Site a thousand-fold, let's say, as a start.
We've gotta try a whole lot of things, together. Let's get to it!
Addendum: things we've gotta do!
We've gotta follow the money
I think a really big missing piece here is the incentives at play, both social and financial, and ways we can help people make money.
It's not about the tech but the business model — successful social networks, from Substack to YouTube to Instagram, give creators a way to build businesses.
We need to go deeper here and make this a priority!
Bluesky right now is the big node in the network and a good place to bootstrap your network, but the value in owning direct relationships with your audience goes beyond that, and we have a lot of levers to pull.
Instagram and X and Substack and other closed platforms can't compete with the diversity of experience possible here, and they basically all trend worse over time.
If we can hit parity with the user experience and financial incentives, winning is inevitable. It's just thermodynamics!
We've gotta solve discovery and reach
Giving people not only better tools but ways to reach people is a big challenge, and there are no quick and easy answers here. The surface area is vast, and I think we need lots of people experimenting.
We can't rely on Bluesky: they're a fairly big network but we need to find places to reach audiences that go well beyond — email lists, other social channels, word-of-mouth, direct connections to IRL communities.
It probably includes being smart with advertising, figuring out mutually beneficial partnerships, creating good feedback loops that reinforce audience engagement, and more!
We've gotta build for communities
We have to help specific people and communities to be successful and sustainable, and they must stand independent of Bluesky.
Leaflet aims to help creators and communities reach each other and make their own businesses and activities sustainable. First we have to create value; only then can we can capture some small portion of that value and grow the ecosystem.
We need growth, which means we need to reach new audiences, which means we need to make things valuable for them specifically.
has a great post, Building for Organizers, that everyone should read:
Post not found.
We've gotta grow Standard Site a thousand-fold
Standard Site lexicons are currently the top non-Bluesky lexicons in use. But the absolute numbers are still tiny, single digit thousands of monthly active users. What does it look like when millions of people are reading and writing on Standard Site?
All websites should be more social and more interconnected if they want to be. We can give them the tools to do so.
Coopetition makes sense in this environment — it's actually mutually beneficial for us to work together by making things that are larger than ourselves right now. It's also an ambitious vision, and this kind of collaboration will help signal the power of all this, and the possibilities ahead, which is a positive sign for growth.
It's partly a storytelling problem, communicating why and how this is the future of the social internet, and demonstrating the diversity and the interoperability here — and how we can do this by not only making great new platforms, but building connections and onramps for an enormous slice of the existing internet. That can go a long way to making this value clear and be a selling point for helping to onboard the next thousand and then the next million creators and writers and publishers and communities.
We've gotta try a ton of things, together
Collaborative discovery and audience reach is a major existential question; how can all of us smaller apps and the whole atmosphere ecosystem help each other get bigger together?
We can delegate some of the more open questions around reader experience and discovery to the community, and as part of this we should publish some kind of collective call to action both for existing atmosphere folks and wider communities e.g. journalism, local, etc.
We should work on tools for creators to make money, and we should enable interesting valuable things they can only do in atmospheric contexts.
This means thinking collaboratively, fluidly, gathering feedback on the actual needs folks have and identifying areas we could offer something unique and useful.
We need to reach people beyond Bluesky, and help them convince their own audiences to use our tools and migrate from other platforms. This means not only demonstrating real unique value, but making it easy to use, minimizing friction, and actually proving that creators can get good results — that they can reach people and make money!
Addendum: Laws of Atmospheric Stacks [WIP]
The Atmospheric stack is only as strong as its weakest parts
For publishing: we need great writing tools AND great conversation tools AND great discovery tools AND great integrations.
The Atmospheric stack is amplified by openness
The parts of the stack being open and to varying degrees interchangeable make the whole thing more powerful.
The Atmospheric stack mixes layers common and unique
One way to win is better custom experiences while still benefiting from shared network dynamics; some layers serve as common infrastructure while others are context-specific, more niche or bespoke.
The Atmospheric stack is fractal
The above applies to atproto as a whole, and to social publishing, and to more specific things like particular community niches: at each level, the network can share the core / meta layers and differentiate on the more granular details and user experiences.
The Atmospheric stack expands the surface area of solutions
It's many builders working together…more perspectives, possible approaches, subsets of the problem tackled, overlapping efforts. Wrangling these emergent dynamics can be challenging, but powerful.
The Atmospheric stack is more user-first
Empowering because it's so flexible, customizable, adaptable…the fatter the stack the more you can really dial in your own experience!
The Atmospheric stack is more antifragile
Permeable: things can enter and expand and evolve the stack.
Resilient: gaps are filled, pain points are patched over, affordances are collectively improved, any layer can be replaced…
The Atmospheric stack is cooperative and cooepetitive
We can all find a niche, share foundations, carve out space and grow the entire ecosystem together; coopetition at different layers helps evolve new equilibria.
Addendum: other sorts of stacks
I didn't have time in the talk, but I think the "atmospheric stack" framing could be a useful lens for pointing at how a similar approach might apply to all sorts of other apps and stacks!
As our friend Victoria once put it, telling people to host their newsletter on Ghost instead of Substack is a bit like telling people to host their tweets on S3, or put their videos on Vimeo instead of YouTube. It's a replacement for a narrow slice of the stack, but by no means a solution for the whole thing.
With atproto we can build the network effects collectively while still having diverse tools that specialize in doing particular things well. "RSS plus social network" for example is a hugely powerful thing to build on, but we've got a ton of work to do with all the UX and interop and creator adoption story etc as mentioned above.
Other potential examples:
YouTube / Twitch --> Streamplace, Spark, etc. (other video apps?)
bookmarking? no big incumbent really, maybe Pinterest?
geo / location stuff? Beacon Bits etc
GitHub --> Tangled, npmx, other social coding tools
Twitter --> Bluesky, Blacksky, future community apps etc
Reddit --> TBD!
Goodreads --> Bookhive & various others!
LinkedIn --> Sifa + more TBD
Of course very interested to here how others are thinking about these dynamics, from coalitions of related products and lexicons and services, to symbiotic interop across apps, to missing layers like shared payments infra and beyond!