Intentional News

Intentional Social Media Design and Architecture

6 min read

Intentional Social Media Design and Architecture

Origin: the owner’s seed, 28 Aug 2026, in the wake of the Meta $17–18bn teen-safety settlement (260826 Meta 18bn Settlement Teen Harm) and its Democracy Now! follow-up (Amba Kak, 27 Aug). The owner: “The envisioning folder needs something that goes to intentional social media design and architecture. My RS vault has something already around walled garden approaches. Society cannot trust, at this stage, private companies to do the right thing with the interests of our children. Despite this win. We can try like a multi-prong approach — meaning considering national networks be established using open source, like Truth Social did with Mastodon. Can you seed the vision.” Seeded 2026-08-28.

The thesis (the owner’s seed + Alf’s read)

The Meta settlement is real progress, but it is not the answer. It is a fine and a design patch imposed from outside on a company whose business model is the problem — and the same model (surveillance advertising, engagement-at-any-cost) will keep producing the same harms wherever it runs. The deeper move is to stop asking the incumbents to be better and instead architect different social media:

  1. Walled gardens for children — safe, age-appropriate, moderated digital spaces (the RS-vault’s “social media ban” strand, below), so the default for kids is a healthy space rather than an addictive one they must be rescued from.
  2. A multi-prong approach, not one silver bullet — the owner’s framing: regulation (rules of the road, as Amba Kak argues) plus education plus building the alternative.
  3. Public-interest networks built on open source — the owner’s specific seed: “national networks be established using open source, like Truth Social did with Mastodon.” The precedent matters: a private venture (Truth Social) forked the open-source Mastodon/ActivityPub codebase to stand up a whole network. If that is technically trivial for a company, it is available to states, public bodies and cooperatives too — a national/public-interest network is a buildable thing, not a fantasy.

The through-line: we cannot trust private companies to do the right thing with our children’s interests — so build spaces whose architecture is not their profit motive.

The walled-garden strand (from the RS vault)

RS-vault note: “social media ban” (100 - myjourney/114 on tech today/social media ban.md). The owner’s earlier thinking there:

  • A total ban (Australia’s under-16 approach) is regressive — it creates a “cliff edge” at 16, drives kids to unregulated underground networks, and treats symptoms not causes.
  • Better: wall-garden social networks — age-verified, moderated, developmentally appropriate spaces with no algorithmic feeds for under-16s, no targeted ads, built-in well-being tools (screen-time limits, mindfulness), and educational integration.
  • Options explored there: tiered walled gardens (12–14 → 14–16 → 16+), “social media passports” / graduated access, digital-detox zones and hybrid physical-digital spaces, algorithmic-accountability reform, digital literacy as a core subject.

That note is the child-safety half of the vision. This note adds the ownership and architecture half.

The open-source / public-network strand (the owner’s seed)

  • The Truth Social precedent (verified): Truth Social was built on Mastodon’s open-source codebase (AGPL-licensed). Mastodon’s own blog confirmed Trump’s platform “found using Mastodon code” in Oct 2021; it initially violated the AGPL licence by withholding source before later publishing it. The point the owner draws: standing up a full social network from open source is a solved technical problem — the barrier is never code, it’s governance, funding and adoption.
  • National / public-interest networks: if open-source forks are easy, a state, public broadcaster, cooperative or civic body can run a network whose incentives are not advertising. This is the “public square as public utility” idea — the social web as infrastructure, like the post office or the BBC, rather than a private enclosure.
  • The Fediverse as the substrate: Mastodon, PeerTube, Pixelfed, Lemmy run on the open ActivityPub protocol (W3C standard since 2023); BlueSky uses the AT protocol; BridgyFed bridges them. Decentralised = no single corporate owner, users can move between instances, no single point of failure or control. (RS-vault: “Fediverse”, 100 - myjourney/114 on tech today/Fediverse.md.)
  • Interoperability as a policy lever: a national network built on open protocols can federate with others, so “public” and “commercial” networks can interoperate on the user’s terms — the user owns the graph, not the platform.

Why this matters

  • Directly answers “what next?” after the Meta win: the settlement changed one company’s defaults; this is about changing the architecture so no company is in a position to choose badly.
  • Connects to the vault’s AI-governance thread: Amba Kak’s “rules of the road” (surveillance-advertising business model as root cause) and the Humanist AI Manifesto’s “common good / democratic governance / protection from manipulation” principles — a public-interest network is the concrete institutional form of those principles.
  • The owner’s globalist/cooperation lens: national networks that federate (rather than wall off) fit the vault’s preference for shared, interoperable, commons-based solutions over nation-state-first enclosures.

Honest limits (Alf)

  • “Public-interest” does not guarantee “good.” A national network run by a state can be a surveillance/enforcement tool; a walled garden run by the wrong hands is a gated prison. Governance and independence are everything, and they are the hard part — open source solves the code, not the custody.
  • Adoption is the bottleneck. Open networks exist (Mastodon etc.) but are tiny relative to the incumbents; a network is only valuable if people are on it. This is why the owner’s multi-prong framing matters — regulation constrains the default while the alternative builds.
  • Truth Social is a cautionary tale too: it used open source to build a hyper-partisan, toxic space. Open architecture is neutral; the social design (moderation, community norms, funding) decides the outcome.
  • The settlement showed defaults matter more than rules — any envisioning must specify defaults, not just permissions.

Open questions

  • Which walled-garden designs exist in practice (school networks, “BeReal for Schools”, EU Digital Passport analogues) and what’s the evidence on outcomes?
  • What would a UK/EU “public square as public utility” network actually look like — who runs it, how is it funded, how is it kept independent?
  • Does the Truth Social/Mastodon fork route work for a public operator under AGPL (source-release obligations are a feature for transparency)?
  • How do children’s walled gardens graduate into the open fediverse at 16 without a cliff edge (link to the RS “social media passports” idea)?

Graph View

Backlinks