A structured atlas of how decentralized systems became real.
Not a feed. Not a hype reel. A structured atlas of how blockchain, cryptography, and decentralized systems evolved — and why each turn happened. Timeline. Technology layers. Protocol comparison. Zero-knowledge. Ecosystem graph. A deep dive on Psy Protocol. The future trajectories.
2008 → present → next.
Filter by era. Each event records the technical innovation, the cause, and the downstream impact. Hover the spine for context; click an event to expand.
Four layers, four trajectories.
Consensus, smart contracts, scaling, and privacy. Each row is a layer. Each cell is a generation, with what it solved and what it left unsolved.
Sybil resistance via cost; longest-chain rule resolves forks probabilistically.
Energy cost; mining centralization; probabilistic finality.
Bitcoin
Energy cost cut ~99%; near-instant economic finality with slashing.
Validator centralization risk; complex re-staking surfaces.
Ethereum (post-Merge), Cosmos, Aptos
Higher throughput by parallelizing block proposal; sub-second finality.
More complex incentive analysis; harder to reason about MEV.
Aptos, Sui (Narwhal/Bullshark)
Lets new chains rent existing validator security instead of bootstrapping their own.
Correlated slashing risk; cross-AVS contagion.
EigenLayer, Symbiotic, Babylon
Conditional spending: multisig, time locks, hashlocks.
Not Turing-complete; limited stateful logic.
P2SH, Lightning HTLCs
Turing-complete VM with global state and deterministic execution. Solidity becomes the lingua franca.
Reentrancy and integer-overflow bug classes; gas-pricing complexity.
Ethereum, BSC, Polygon, every L2
Resource-typed asset model: 'coins are objects', not balances. Reentrancy structurally impossible.
Smaller ecosystem; tooling lags Solidity.
Aptos, Sui
Programmable accounts: gas in any token, social recovery, session keys, batched UX.
Bundler/paymaster centralization; specification complexity.
Safe, Argent, Coinbase Smart Wallet
Off-chain mutual updates; on-chain only for open/close. Near-instant payments.
Liquidity routing; channel watchtower assumptions.
Lightning Network
Off-chain execution; L1 fraud-proof window enforces correctness.
7-day withdrawal delay; assumes one honest watcher.
Optimism, Arbitrum, Base
Off-chain execution; L1 verifies a succinct proof. Withdrawals as fast as proof.
Prover hardware cost; proving system complexity.
zkSync Era, Scroll, Linea, Polygon zkEVM
Cheap short-lived L1 data lane optimized for rollups.
Long-term DA still depends on external committees or DA layers.
Ethereum + EigenDA / Celestia / Avail
Hide sender, recipient, and amount via zk-SNARK proofs over a Merkle tree of commitments.
Trusted setup; shielded usage rate has historically been low.
Zcash
Break direct on-chain linkage between sender and recipient.
Anonymity set is small; metadata leaks; legal status fraught.
Wasabi, Tornado Cash (sanctioned 2022)
Each receipt lands at a one-time address derivable only by the recipient.
Sender side still observable; full privacy needs additional ZK layers.
EIP-5564, Umbra
Privacy + cross-chain transfer with no honest-majority assumption: every state transition is proven.
Compliance UX; proof costs at peak load; integration surface.
Psy Protocol, Aztec, Iron Fish
Same four axes. Different bets.
Every L1 / L2 / privacy chain is a position on the same four axes — decentralization, scalability, security, privacy. The radar shows the position; the table shows the price.
| Protocol | Type | Consensus | Effective TPS | Finality | Decentralization | Scalability | Security | Privacy |
|---|---|---|---|---|---|---|---|---|
Bitcoin | L1 | Nakamoto PoW | 7 | ~60 min (6 conf.) | 95 | 25 | 95 | 25 |
Psy Protocol | Privacy L1 | ZK-verified bridge & privacy layer; chain-agnostic | 1,000 | Proof-bound (seconds–minutes per chain) | 88 | 78 | 92 | 95 |
Ethereum | L1 | PoS (Casper FFG + LMD-GHOST) | 18 | ~13 min (epoch boundary) | 85 | 40 | 88 | 30 |
Zcash | Privacy L1 | Equihash PoW | 26 | ~25 min | 80 | 25 | 85 | 92 |
Cosmos Hub | L1 | Tendermint BFT | 1,500 | ~6s | 75 | 70 | 70 | 30 |
Arbitrum | L2 | Optimistic rollup over Ethereum | 250 | ~7 days (challenge window) | 70 | 80 | 82 | 30 |
Scroll | L2 | ZK rollup over Ethereum | 150 | ~30 min (proof cadence) | 70 | 75 | 86 | 30 |
OP Stack chains | L2 | Optimistic rollup; superchain | 200 | ~7 days | 65 | 78 | 80 | 30 |
zkSync Era | L2 | ZK rollup over Ethereum | 200 | ~24h (proof cadence) | 65 | 78 | 86 | 35 |
Starknet | L2 | ZK rollup; Cairo VM | 300 | ~30 min | 65 | 82 | 85 | 35 |
Aztec | Privacy L1 | Privacy ZK rollup over Ethereum | 50 | ~30 min | 60 | 60 | 86 | 90 |
Solana | L1 | PoH + PoS (Tower BFT) | 4,500 | ~12.8s | 55 | 88 | 70 | 25 |
Compute proven. Inputs hidden.
Zero-knowledge proofs let one party prove a statement is true without revealing why. The spec is small. The applications are not. We list the schemes, the systems built on them, and what trust assumptions each retires.
such that a public statement S(x, w) holds. I will not show you w. I will give you a small proof π that anyone can verify in constant-ish time.
I learn that the statement is true and I learn nothing else about w — that is what 'zero-knowledge' literally means.
Smallest proofs (~200B) and fastest verification (constant). Pre-quantum security; requires a circuit-specific trusted setup ceremony.
Trusted setup per circuit; circuit changes mean a new ceremony.
Zcash Sapling, original Tornado Cash, many fixed-circuit verifiers.
Honest verifier; broadcast-time witness availability.
Universal SNARK with a single setup that supports any circuit up to a fixed size. Custom gates, lookups, and recursive composition.
Universal trusted setup; larger proof and verification cost than Groth16.
zkSync Era, Aztec, halo2-based provers.
Per-circuit ceremonies.
Hash-based, transparent (no trusted setup), post-quantum. Larger proofs than SNARKs but provers are faster on big circuits.
Larger proofs and higher verification cost on-chain; fewer verifiers in production EVM than SNARK families.
Starknet, Polygon Miden, RISC Zero.
Trusted setup; quantum-vulnerable assumptions.
PLONKish proofs that can verify other proofs of themselves — recursion without a separate verifier circuit.
Engineering complexity; lookup-table sizing for verifier circuits.
Zcash Orchard, Scroll, Pluto.
Single-shot proofs as the only deployment model.
Cryptographic primitive that proves the signer is one of a fixed set without revealing which. Predates ZK in privacy use.
Anonymity set bounded by ring size; no general computation.
Monero, MimbleWimble (related).
Sender identifiability for the specific ring.
A bridge with no committee.
Psy Protocol is a ZK-verified bridging and privacy layer. Every state transition — deposit, transfer, withdraw — is enforced by recursive zero-knowledge proofs that any chain's verifier contract can check. There is no multisig. There is no validator committee. The bridge is the proof.
Architecture overview
User wallet
Holds keys, generates ZK proofs of ownership and inclusion locally before broadcasting.
Psy node
Maintains the global Merkle state of notes; serves inclusion paths; processes blocks of shielded transactions.
Recursive prover
Aggregates per-transaction proofs into a single recursive proof per block. Bridges its statement across chains.
Relayer
Submits proofs and metadata to each chain's verifier contract. Stateless: cannot censor a user who runs their own.
On-chain verifier contract
On every supported chain. Verifies the recursive proof against a hard-coded verification key. Releases or accepts assets accordingly.
Indexer
Reads chain events and rebuilds public-state views. Anyone can run one. The protocol does not depend on it for safety.
ZK-based verification
Each shielded transaction generates a client-side proof: 'I own a note in the tree; I am not double-spending; value is conserved; output notes are correctly re-randomized.' Per-transaction proofs are aggregated into a single recursive proof per block. That recursive proof is then verified by the 'verifier contract' on each supported chain. One proof. Uniform spec. Many chains.
- · state-rootbefore
- · state-rootafter
- · list of nullifiers (spent notes)
- · list of new commitments (new notes)
- · public payouts (withdrawals only)
Cross-chain design
Every supported chain holds a verifier contract (Solidity / Move / Cairo / Bitcoin Script-extension, depending on the chain) with a hard-coded verification key. Psy assumes no 'message bus' or 'signing committee'; the same recursive proof shape is verified on each chain separately. State travels across chains carried by proofs, not by signatures.
The bridge is not infrastructure. The bridge is the proof.
Deposit & withdrawal flows
User sends asset to the verifier contract on chain X. The contract emits an event with a commitment hash, not the user's address linkage.
User locally generates a note (commitment, nullifier seed, value, asset, owner pubkey). Note is added to the Psy Merkle tree by the Psy node when the deposit event is finalized.
No client proof needed for deposit — the chain itself attests to the deposited value. Inclusion in the Psy tree is checked at next-step proof time.
Versus traditional bridges
| Axis | Traditional bridge | Psy |
|---|---|---|
| Trust model | Honest majority of multisig signers or validator committee. | Cryptographic verification. No honest-majority assumption. |
| Failure mode | Key compromise → unbounded loss (≥ $2.5B lost 2022–2024). | Bug in verifier circuit → bounded by code review and formal verification. |
| Censorship | Operator can block transfers per address. | Anyone can act as relayer; user can self-relay. No address-list policy on the protocol. |
| Privacy | Public source and destination on both chains. | Source and destination unlinked; amount hidden in private mode. |
| Cross-chain composability | Custom message-passing per pair of chains. | One verification key per chain, one recursive proof shape across all chains. |
| Latency | Bound by signer round-trips and committee finality. | Bound by proof generation + verifier finality on the destination chain. |
| Operational surface | Hot wallets, key ceremonies, KYT pipelines per integration. | Public verifier contract per chain. No keys to rotate. |
What it solves
From 2022–2024, more than $2.5B in user funds were lost to bridge exploits. Almost all of these were committee or multisig compromises — not protocol bugs. Psy makes the failure mode 'a bug in a public, audited verifier circuit', not 'a key in a private process is stolen'.
Why it is different
- 01
Psy treats the bridge as an artifact of cryptography, not of governance. Removing the committee removes the largest historical attack surface in the field — bridges are responsible for the majority of crypto losses since 2022, and almost all of those compromises trace back to multisig or signer key theft.
- 02
Recursion is load-bearing: per-transaction proofs aggregate into per-block proofs, which aggregate further across chains. Without recursion the verification cost would not be feasible on EVM chains.
- 03
Privacy is built in, not bolted on. The same circuits that prove correctness already hide the sender, recipient, and amount on the Psy side; selective disclosure is added by viewing-key, not removed by audit.
- 04
Position in the evolution: Psy is what 'a bridge' looks like once cryptography catches up to the field's biggest unsolved governance problem. It belongs alongside zkEVMs and shielded payments as the third leg of the ZK-everywhere thesis.
Protocols, builders, applications.
Click a node to anchor it; the graph highlights its dependencies and dependents. The graph is built from public-record relationships, not market lists.
ZK bridge & privacy layer.
Three trajectories. Each with a load-bearing assumption.
Scaling. Privacy adoption. AI × chain convergence. We sketch each as a band of possibility — what the optimistic, baseline, and pessimistic curves look like, and which assumption flips one into another.
Scaling — effective L2 TPS
Proof-generation cost continues to fall ≥2× per year and EIP-4844 follow-ons (full Danksharding) ship before 2028.
If completion of Danksharding slips past 2028 OR proof-cost curves stall, baseline collapses to pessimistic.
Privacy adoption — % of L2 value in shielded form
Compliance UX matures (selective disclosure + traveling viewing keys) and at least one major institution pilots a shielded settlement leg.
A high-profile sanctions enforcement action against a major shielded protocol pulls baseline toward pessimistic.
AI × chain — % of on-chain transactions initiated by autonomous agents
Agent-grade signing UX (passkey-bound session keys + per-action policy proofs) ships at scale; agent reputation systems exist.
A coordinated agent-led economic incident triggers blanket throttling at the wallet/RPC layer.
Blockchain is what cryptography looks like once it has consequences.
History first, hype never.
We organize the field by the technical decisions that produced it: a satoshi paper that solved double-spend, a stateful machine that turned currency into computation, a rollup that turned computation into compressed proofs. Every section is built on what came before.
Architectures over coins.
We compare protocols on the same four axes — decentralization, scalability, security, privacy — and show what each design buys and what it pays. Tickers come and go. Architectures persist.
The Psy section is the load-bearing one.
Psy Protocol is included not as advertising but as the present-day end of the cryptographic line that started in 2008: ZK-verified state, no multisig, no committee, cross-chain by proof. We show why it is technically distinct, and where it sits in the evolution.
Future curves, not promises.
The Future section sketches scaling, privacy, and AI×chain trajectories as bands of possibility, not point estimates. The atlas tells you where the field is, where it could go, and what would have to be true for it to get there.