ＬＵＮＡ

On a public chain, a conditional instruction is visible before it executes.
A stop-loss, a rebalancing trigger, a bid ceiling: each must be published before it can fire. That transparency makes chains verifiable, and lets anyone position ahead of the condition.
@Nillion's Covenants take a different approach. An instruction is encrypted, tied to a condition, and posted on Ethereum. Observers can see that a Covenant exists but not what it says. When a live data feed meets the condition, nodes publish their shares; once enough are available, the payload is reconstructed and verified on Ethereum before any action runs.
Privacy here is tied to a condition: the instruction opens when it is met, and can expire unrevealed if it never is. That differs from Nillion's earlier blind computing, which keeps data private while it is processed.
Dusk, Nillion's first roadmap phase (not the Dusk Network project), went live on Ethereum mainnet in early October. The initial design is limited: price-based conditions on a select list of assets, windows up to 30 days, and no cancelling after posting. Multi-chain support (Darkfall) is planned, not live. The $NIL ERC-20 migration was a separate, earlier event.
Open questions: the system relies on the node set and data feeds working correctly, and fully private conditions are a later phase (2027). An Oct 24 $NIL unlock, roughly 16–19M tokens depending on the tracker, adds supply and could add selling pressure; it doesn't determine price.
Which onchain conditions should stay sealed, and which should stay public?

#nil $NIL @Nillion #NIL

If an AI agent can find a useful paid API on its own, why does a human still have to pull out a card every time?

That's the gap Kite is working on. Its Agent Passport gives an agent an identity and a funded wallet. The part I find most relevant is the limits. You approve a session up front with a passkey and set the budget, time window and which merchants it covers. The agent can only spend inside those lines, and each payment is designed to link back from session to agent to you, with a receipt anchored on-chain.

That pairs with x402, which puts the old HTTP 402 "Payment Required" code to work. A server asks for payment, the agent pays, and access is granted. Picture an agent comparing data APIs, picking one under your cap, and settling the bill while you do something else.

It's still early. Kite says Passport and its mainnet went live on April 30, but its own docs still list parts of the payments guide as in progress, and not every service supports x402. Whether this becomes common depends on how many services adopt it and how well the spending limits hold up in practice.

The question it raises is a good one: how do you let software spend money without handing it a blank check?
#KiteBBW $KITE

GO...ND claim...🌟

🚀 $BTC Bullish Recovery

Trade Setup (Long)

Entry Zone: 81,200 – 81,900
Take Profit Targets:
TP1: 82,500
TP2: 83,500
TP3: 85,000

Stop Loss: 80,500

Price recovering with residual buying pressure around the short-term averages. Continuation higher is likely if it holds above 81,200.

Disclaimer: This is not financial advice. Cryptocurrency trading involves high risk of loss. Always do your own research and trade responsibly at your own risk.

Buy and Trade $BTC
🚀