Intermediate · Knows crypto, not Kaspa

The one-sentence difference

Bitcoin selects a chain and orphans the rest. Kaspa represents concurrent proof-of-work in a DAG and orders it with GHOSTDAG. The double-spend check still happens after ordering. A DAG is not a free pass to spend twice.

If you already think in UTXOs, fees, and confirmation policy, you can skip Door 1. Stay here until you can place Kaspa without borrowing Ethereum’s words.

Against the rest of the map

Bitcoin. Same PoW cash lineage, UTXO ledger, no premine. Kaspa’s block interval is much shorter. That does not copy Bitcoin’s settlement culture. Recipients still pick a policy. Node bandwidth and validation cost rise with the extra blocks. If a Bitcoiner asks “is it final?”, the honest answer is: inclusion, acceptance, and the recipient’s wait are three different things. Same as Bitcoin, faster blocks.

Ethereum. Ethereum moved to proof of stake and an account virtual machine. Kaspa stayed with miners and UTXOs. Toccata added spending rules (covenants), covenant IDs, a ZK precompile, and sequencing lanes. That is not the EVM. You do not deploy a Solidity app onto Kaspa L1 today and get Ethereum’s tooling. “Smart contracts on Kaspa” is a sentence that needs a label: covenants are live protocol; application compilers are not all stable.

High-throughput PoS and L2s. Fast inclusion is often a sequencer or a validator set. Kaspa’s speed is many proof-of-work blocks. Honest work that arrives late is kept, not discarded. 100 BPS is a research target, not a spec. Crescendo’s 10 BPS is live. If you came from Solana, do not map TPS slogans onto block rate. Nodes still receive, validate, and store every block.

UTXO, not an account VM

Coins live as unspent outputs. A payment consumes outputs and creates new ones. Change is an output back to you. A covenant is a spending rule attached to an output: wait, limit, name the recipient, require a proof. That is closer to Bitcoin Script than to an EVM contract with storage.

Shared execution among many users (a lending pool, an AMM with one global book) is a different problem. Covenants do not by themselves give you that. vProgs are the research name for based programs with validity proofs. They are not a product testnet you can treat as live.

Live versus later

Use these labels. Mix them and you are pitching.

• Live. GHOSTDAG. 10 BPS (Crescendo). Toccata protocol rules on mainnet. rusty-kaspa v2.0.1 is the maintenance tag checked here.
• Release candidate. SilverScript v1-rc1. Experimental. Prefer Testnet-10 until a tagged v1.
• Prototype. Argent. Examples exist. The README says it is not release-ready.
• Research. vProgs. Repository existence is not production.
• Proposed. DAGKnight. The rusty dagknight branch is not merged. No mainnet activation is documented in the checked KIP or node releases.
• Local model. Dated 8 September 2026. Inherited budget, atomic offer completion, competing execution, a correction bounty, and a permission arena. You can break them. They do not sign or submit.
• Testnet lab. Existing covenant primitives on Testnet-10. This site’s Town is that lab. One browser holds every key. That is not three independent people.
• Withdrawn. Parker’s hosted V5 and V6, 8 September 2026. Town at /covenants stays. MIX does not ship those versions.

A merged Active KIP is law. A tweet, a Discord rumor, and a forum thread are not. Toccata is consensus. silverc is a compiler. People still mix those two in chat.

What speed does not buy

It does not remove miner concentration. It does not make every wallet support covenants. It does not make a repository into a product. It does not make a price chart into evidence. The skeptical list is the same as for any PoW coin, plus the extra node cost of a fast DAG.

Door 3 is the dated machine: KIPs, builder network, node. If you only wanted the map, open status and the playground next.