Lily雪莉呀

How can you know the joy of fish if you are not a fish?

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

There’s something I’ve been wrestling with: I wouldn’t think twice about giving an AI Agent tens or even hundreds of dollars to check market prices, buy data, or call models. But hand over control of my entire wallet? No way. Just because AI can get things done doesn’t mean it’s always right. If it makes a bad call or invokes something by mistake, who’s responsible once the money is gone?

With that question in mind, I looked into Kite. What appealed to me most was its Agent Passport design. It’s not just about giving an agent a wallet; it ties identity, permissions, and payment limits together. You can set per-transaction caps, an overall budget, a validity period, and which assets are allowed in advance. The Agent can only operate within those limits. For example, you could give it permission to spend just $100 on data services, with no access to your other funds. Put simply, it lets AI work inside a cage without letting it get out.

Now for the fundamentals: the mainnet launched in April 2026, with Chain ID 2366, EVM compatibility, and stablecoin settlement support. The testnet has recorded more than 1.9 billion Agent interactions, with a peak of 30 million calls in a single day. Those numbers look impressive, but they’re from the testing phase. We’ll have to wait and see what real paid usage looks like on mainnet.

As for the token, the total supply is capped at 10 billion: 48% for the ecosystem and community, 20% for modules, 20% for the team and early contributors, and 12% for investors. Value accrues through fees on AI service transactions, with some revenue converted into KITE to support ecosystem development. Whether this mechanism can sustain the token ultimately comes down to the business: are enough users and enterprises willing to keep paying for data, models, and services?

So I’ll be keeping a close eye on three things: mainnet transaction growth, actual settlement volume, and the protocol’s real revenue. Whether all three grow together matters more than simply looking at the number of addresses.

The AI Agent economy has significant potential, but a promising sector doesn’t mean $KITE is guaranteed to have value. Even the best payment infrastructure can’t create demand on its own. Kite is using x402 to bring HTTP 402 back to life, combined with state channels for high-frequency micropayments, which really does lay the groundwork for machine-to-machine transactions. Now we’ll see how well it delivers. @KITE AI 中文 #KiteBBW $KITE

1

#nil $NIL @Nillion
The more I look into Nillion’s Encrypted Markets lately, the more I think there’s something to it—especially the Covenants part.

In the past, placing stop-losses or limit orders on-chain was basically like putting your plans on display. Once people saw what price you were waiting for, MEV and snipers would be right on your tail. Covenants takes a straightforward approach: encrypt your trading intent and seal it on-chain. Until the conditions are met, neither nodes nor the project team can read it. When the price or time condition is met, the nodes work together to unlock and execute it. It relies on threshold cryptography: the secret is split among a group of nodes and only reconstructed after enough of them confirm. No keeper, no single trusted party.

This is already beyond the white paper stage. Dusk has launched on Ethereum mainnet—that’s the first step. It’s not just private transfers; it brings conditional execution directly to L1. Sealed orders, conditional payments, and even Agent strategies can run on it. Nodes need to stake at least 70,000 NIL, with rewards coming from 0.5% annual inflation plus trigger fees. Developers can integrate it using the TypeScript SDK, while identity-based encryption handles the underlying cryptography so they don’t have to.

What I’m most optimistic about is its positioning as a privacy overlay that requires no migration. It stays on Ethereum, so composability remains intact; it simply encrypts the trading intent itself. Before, you either hid in the mempool or moved to an entirely different chain. Covenants takes the middle route.

Over the next seven eras, Darkfall and Skyglow are set to expand to multiple EVM chains and even enable observation on chain A with settlement on chain B. Nightside and Neon, meanwhile, will bring custom conditional logic and a dedicated contract language. All of this is still in the works, so it’s too early to count it as a win.

When people think of privacy in crypto, many think only of mixers. Nillion is working on privacy at the strategy layer. Keeping instructions private doesn’t mean every trigger condition is hidden, and it certainly won’t make money for you. But at least the plans you stayed up all night researching don’t have to become someone else’s free homework. The rules can be publicly verified; there’s no need to show your hand to the whole table ahead of time.