Anurag Arjun@anuragarjun · 2026/08/18赛道影响RT @AvailProject: Nexus Rising is now live 🕊️
And we're here to give apps a faster, more scalable way to reach users and liquidity across…
AI：偏主题投资表达，释放对基础设施与新叙事的偏多信号。
Anurag Arjun@anuragarjun · 2026/08/05观点输出RT @ShieldTXxyz: A ShieldTX trade uses a fresh Hyperliquid wallet.
Your entry, size, and PnL are still public state. What is shielded is t…
AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。
Anurag Arjun@anuragarjun · 2026/07/29赛道影响fair. if your requirement is "the venue is private and the operator learns nothing," then HL today doesn’t meet that, and ShieldTX doesn’t fully solve that yet either.
we started off with a fully private system but that ran into the liquidity bootstrapping issue and so we have come up with a system that works for already liquid venues.
the wedge is narrower: for people already trading on HL, public wallet attribution leaks a lot. ShieldTX targets that first - known wallet -> funding path -> position wallet.
so I’d frame it as: not operator-blind privacy yet, not claiming that. but also not nothing for the HL users getting copied/liquidation-hunted off their known wallets today.
different threat model. operator-blind execution is the harder next step.
AI：偏主题投资表达，释放对基础设施与新叙事的偏多信号。
Anurag Arjun@anuragarjun · 2026/07/29赛道影响fair criticism. current ShieldTX is not a trustless dark pool / fully operator-blind system.
we don’t and won’t trade against users - no prop desk, no internalizing user flow, no using intent data to take the other side.
but tbh that’s also the real constraint with Hyperliquid today: if an account trades there, its size/liq/fills are public.
so the practical privacy layer is attribution privacy - breaking the link between known wallet -> funding path -> HL position wallet.
hiding the position itself needs HL-native privacy or off-HL execution. ShieldTX starts with what works on HL today: shielded funding + isolated position wallets, then we reduce operator visibility over time.
on the web2 analogy - this is not web2 - the non-web2 part is custody/settlement correctness: the Arbitrum vault accepts state roots only with valid proofs, so infra can’t just invent balances or settle invalid spends. but execution privacy from the operator is a separate problem and not what we’re claiming yet.
AI：偏主题投资表达，释放对基础设施与新叙事的偏多信号。
Anurag Arjun@anuragarjun · 2026/07/29赛道影响yes - current version is not an operator-blind protocol. the sequencer infra has privileged visibility because it has to process notes, nullifiers, intents, and position workflows.
the privacy claim today is against public observers: wallet graph, copy-traders, liq hunters, etc. not "ShieldTX infra learns nothing."
the zk part is for settlement correctness: the Arbitrum Vault only accepts valid root updates, so the sequencer can’t settle invalid spends/double-spends/value breaks. but zk is not currently being used to make the sequencer blind to all user activity. but its on the roadmap, yes.
AI：偏主题投资表达，释放对基础设施与新叙事的偏多信号。
Anurag Arjun@anuragarjun · 2026/07/29赛道影响so if someone is tracking your main wallet, they don’t see "this wallet opened X size with Y liq." they see a deposit into ShieldTX, and a separate position wallet trades on HL.
if they somehow identify the position wallet, the position is public. the goal is breaking that link at scale.
AI：偏主题投资表达，释放对基础设施与新叙事的偏多信号。
Anurag Arjun@anuragarjun · 2026/07/29观点输出yeah - liq hunting is one of the big reasons this exists.
nuance: Hyperliquid position data is still public for the wallet that is trading. ShieldTX doesn’t make HL itself private.
what we hide is the attribution path: your known/source wallet -> isolated position wallet -> size/liq/PnL/order history.
AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。
Anurag Arjun@anuragarjun · 2026/07/29观点输出in the current design, the user authorizes spends by signing the input note commitment. the Nightshade sequencer proposes the transition: marks the nullifier and creates the new output/change notes.
the SP1 circuit re-checks it: valid sigs, nullifier not spent, value conserved, new commitments correct. the Nightshade Vault contract on Arbitrum only accepts the new roots if the proof verifies.
current privacy target is public attribution: source wallet -> position wallet -> trade graph.
AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。
Anurag Arjun@anuragarjun · 2026/07/29观点输出@TheDhawalS for now, no - we have some heuristics but it kind of relies on a 30s time threshold of the user taking a trade after the original trade - open to ideas though
AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。
Anurag Arjun@anuragarjun · 2026/07/29观点输出@TheDhawalS yeah while doing analytics work for https://t.co/t1nFp5aB5D, we realized that there is massive industrialization of copy trading, especially on HL
AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。
评论
暂无评论