| ← Index | Next: UFT → |
Egor Homakov is the author of UFT and xln. His work began in software engineering and security research, including the Rails mass-assignment issue, then boutique security work at Sakurity. Since 2017 his focus has been the architecture of financial systems. UFT/xln draws on nine years of research into settlement, bilateral accounts, organizational authority and programmable finance.
Existing finance already has jurisdictions, entities and accounts. xln makes those financial relationships replicable, verifiable and signed, with account proofs that can be exercised in an underlying programmable J-machine.
UFT is the model, RCPAN the account bound, and xln the implementation. The ambition is to make existing economic activity provable, rather than create a separate financial universe.
By 2050, make accounts supporting 51% of world GDP provable and disputable on programmable Js. J reserves are part of that financial foundation too; account claims and their backing must be reconciled without counting the same value twice. See the mission.
No. Proof establishes the agreed obligation; collateral backs its secured portion. The credit limit is the maximum unsecured exposure the grantor accepts, manually or through soft-limit policy. Only collateral-backed entitlement is secured. Zero credit is valid. A proof of debt does not guarantee that an insolvent debtor can repay it. Credit selection belongs to the parties and operators, while the runtime enforces the agreed limits before signing.
Three distinct protections: retained signed evidence for dispute, chosen collateral backing, and Delta Transformers enforcing conditions while value moves. A user can claim the secured entitlement under J’s rules; any remaining unsecured receivable follows the actual reserve/debt enforcement path.
This improves control over financial claims and bounds direct account exposure. It does not make every hub solvent, eliminate shared-operator failures or remove all first-mover incentives on unsecured reserves.
Unrelated parties do not need an account’s history to agree on its next state. Independent account updates can progress locally, while J handles registration, collateral, net settlement and disputes. The parties or delegates retain the usable proof. This changes ordinary agreement and evidence scope; it does not remove the chosen J’s own consensus, availability and inclusion requirements.
Yes. Payments, swaps, lending and conditional orders are part of the architecture. A pool, auction or clearing house can be an Entity with bilateral financial relationships. Coordination inside that service does not require the entire economy to share one execution domain. Comparing implementations still requires preserving the same pricing, rights, liquidity and failure outcomes.
Ethereum, TRON and XLNC are the initial focus; Base, Arbitrum and other compatible EVM Js can share the same financial path after their boundaries are verified. The small-device full-verification objective targets XLNC, pronounced “excellence”, a compact J tailored to xln. It does not mean storing and executing the full Ethereum or TRON history on a watch.
Expensive ordinary J activity encourages the Account layer. XLNC is a conventional stateful EVM J with roughly 10–20 times less block gas capacity, not a stateless or ZK design. Phones and laptops run ordinary full nodes and synchronize normally. Those resource budgets and emergency-dispute capacity need measurements. See XLNC design.
The intended default is one hour for a hub to react to a user-started dispute. Cross-J settlement additionally needs enough time to observe and copy valid lock/ reveal evidence. Current code sums signed hub/user windows, yielding 25 hours with the default one-hour/24-hour policy. A responder-only reaction rule and its safe finalization conditions must be distinguished before changing that code.
Fund an account, pay another user, swap, withdraw and recover after hub failure. The wallet acceptance plan records the new-UI walkthrough and existing-frontend E2E requirements. Release readiness is established by actual candidate evidence, not an architectural argument or an old launch date.
No billion-TPS production measurement is established here. The architectural claim is horizontal capacity across independent accounts, with low ordinary J load. Routing capacity, concentrated hubs and simultaneous disputes remain real constraints. The comparison distinguishes that contribution from individual prior ideas and from unmeasured product claims.