A new research implementation called Kurrent entered the R&D chat with an Eltoo-inspired latest-state channel design for Kaspa. The core idea is payment-channel-style state that mostly changes off-chain: participants keep agreeing on newer channel versions, and if settlement or dispute reaches the chain, the latest valid version should beat an older stale one. Eltoo is the research model behind that "latest state wins" approach; Kurrent borrows the principle, not Bitcoin's exact transaction mechanism. (Kurrent README) Arthur said the local-devnet gate currently passes on real Kaspa devnet, the Toccata covenant VM, and real Lightning Network regtest, with replayable evidence from ./scripts/check.sh. (source)
The Kaspa angle is that a latest-state channel needs the base layer to enforce a narrow rule: preserve the channel identity, advance the state number correctly, reject stale or non-continuing spends, and allow only the surviving state to settle. Kurrent is testing whether Toccata-era covenants and transaction introspection can express that rule in a Kaspa-shaped UTXO system. (Kurrent research brief)
The important part is the evidence shape. The project includes a live state-channel flow with a real funding transaction, a real S1 to S2 state update, a real settlement transaction, and a stale spend rejected by a node with the actual signature-script error. (source) It also includes a live factory example with funding and materialisation transactions, plus a live atomic Lightning Network to Kaspa swap in both directions with LND logs and the observed preimage. (source)
Kurrent also has a typed Rust protocol model using secp256k1 and sha2 to re-validate the on-chain artefacts, state headers, and settlement templates. (source) That makes this more than a sketch: reviewers can replay the raw transactions, witnesses, txids, preimages, and model checks rather than only reading a proposal. (source)
Arthur followed with a research note explaining the design frame: the latest-state idea, the Kaspa and Toccata covenant angle, and the open requirements around ordering, response windows, fees, monitoring, and factory materialisation. (source) He was careful not to oversell it, saying it is "not a production claim so far" and should be read as a mechanism note to make the prototype easier to review. (source)
All sources link to public messages in the Kaspa Core R&D (public) Telegram channel.