Why Recilic isn't a stranger network
The most reasonable question a technical person asks about Recilic: "You've built end-to-end encryption, erasure coding, and proof-of-possession audits. That's most of a decentralized storage network. Why not add an incentive layer and let strangers store each other's shards — why not be Filecoin for backups?"
Short answer: because friendship is a better protocol than tokens. Longer answer below.
The graveyard
"Turn the world's idle disks into a storage marketplace" is an idea so good it gets seriously attempted every five years. MojoNation (2000). Wuala's storage trading (2008). Symform (2009). Space Monkey (2012). Storj's early consumer nodes. Sia. The consumer versions of all of these are dead or pivoted. The survivors — Filecoin, today's Storj — serve web3 projects and S3-compatible enterprise storage, not people backing up family photos.
The failures were structural, not accidental:
Stranger churn is poison. A stranger's home machine leaves the network permanently, without notice, and owes you nothing. The only defense is aggressive over-replication plus constant repair traffic — which consumes the very bandwidth participants were promised they'd barely notice. More repair → a worse deal → more people leave → more repair. A death spiral with excellent documentation.
The incentive layer eats the product. For strangers to store your data reliably, they must be paid; payment needs pricing; pricing needs a marketplace; a marketplace needs reputation, penalties, and dispute resolution. Filecoin built an entire blockchain to answer this — proof-of-replication, proof-of-spacetime, slashing, a token economy — and it is still not a thing a person with a Mac and a photo library can simply use. Every engineering hour spent keeping the market honest is an hour not spent making backup better.
Two-sided markets have brutal cold starts. No supply without demand, no demand without supply. A friend mesh has no such problem: you and your two Macs are already a minimum viable network.
And the platform reality: token economics are, practically speaking, incompatible with the Mac App Store — which is exactly where a consumer backup app needs to live.
Friendship is an older protocol
The entire incentive layer exists to answer one question: why would this node keep holding my data?
Between friends, the question answers itself. Reciprocity is the payment: I hold your shards, you hold mine. No pricing, no reputation system, no slashing — because if your friend's Mac goes offline, you can text them. Churn correlates with real relationships: a node leaving your mesh usually means something you already knew. "Moving house. Replacing the Mac mini."
And the trust being asked for is smaller than it looks. Everything is encrypted before it leaves your machine; no node — nor any coalition below the recovery threshold — can reconstruct anything. What you trust friends with is availability, not confidentiality. Confidentiality is guaranteed by math. Availability is checked by the audits: nodes must periodically prove, byte-for-byte, that they still hold your shards.
The honest limits
The friend model has real constraints. You need machines — yours, or friends who'll trade space with you. If you own exactly one Mac and know no one with spare disk, Recilic today has less to offer you than a cloud backup subscription does — our own comparison page says so.
For people who want an always-on node without asking a friend, there's a middle path that doesn't require a stranger market: an official, company-operated storage node you could subscribe to as one member of your mesh. It would hold ciphertext fragments like any other node, answer the same possession audits, and reconstruct nothing below the threshold. Even we couldn't read what we'd store. That's a service relationship, not a marketplace — no tokens required.
The door isn't locked
Nothing in the architecture forbids stranger nodes. Shards are ciphertext; audits don't care whom they challenge; placement doesn't care who owns the failure domain. If a genuinely sustainable open-network model emerges someday, the primitives are ready.
But the roadmap is unambiguous: make backup between people who already trust each other so good it feels inevitable. The mesh you can build this weekend — your Mac mini, the office Studio, a friend across town — is the product. Everything else is a maybe.