Web3 gal

One Thing AI Should Never Take Away From Us: Control. I've been thinking about how much we're starting to trust AI agents with. They can already handle tasks, interact with services and make payments on behalf of their users. That's exciting, but I still have one concern: what happens when an agent needs to make a decision that should involve its owner? This is one reason I find Concordium's upcoming Human Consent Request interesting. The idea is simple but important. An agent or merchant can request human approval when consent is needed, with that approval anchored on-chain so it can be independently verified. I like this because giving AI more responsibilities shouldn't mean giving up human control. There should still be a clear point where the person behind the agent gets to say yes or no. As agentic commerce grows, I think features like this will become increasingly important. @Concordium is exploring a side of AI adoption that deserves more attention. I'm looking forward to seeing how this develops.

AI Agents Don't Just Need Wallets. They Need Permission. Imagine giving an AI agent access to your money. It can book services, purchase data, pay other agents and complete tasks while you're away. Sounds useful, right? But would you give it unlimited authority? That's a question the agentic economy cannot ignore. I believe the next major step for AI agents is not simply making them capable of more actions. It's making sure they can prove who authorised them and what they're allowed to do. This is one reason I'm paying attention to @Concordium. Its infrastructure brings together verified identity, privacy-preserving proofs and tools for controlled settlement. Think about the possibilities: 💧 An agent backed by a verified human or business. 💧 Credentials that can be checked without exposing private information. 💧 Payments through x402. 💧 Sponsored Transactions so agents don't need to hold $CCD to cover network fees. 💧 Protocol-Level Locks designed to hold funds until agreed conditions are met. These are important building blocks for an economy where software doesn't just assist people but actually acts on their behalf. The interesting part is that Concordium's verification can extend to agents on other networks too. Developers don't have to abandon the ecosystems they're already building on. For me, this is a much more meaningful direction for blockchain and AI than simply adding another chatbot to Web3. The future belongs to agents that can do useful work ~ and infrastructure that makes their actions accountable. That's a narrative worth watching. 🚀

$CCD is having a strong day, up 23.20% in the last 24 hours, with trading volume reportedly increasing by more than 300%. 👀 But what interests me more is the progress happening beyond the price chart. As AI agents become capable of handling tasks, interacting with services and making payments, one important challenge remains: how do we know who is behind an agent, and how can its actions be held accountable? This is the problem @Concordium is working to address. The Agent Registry provides a foundation for registering and verifying AI agents, while the Verified by Concordium Badge connects an agent to a verified human or business without publicly exposing personal information. The MCP Server takes this further by making verification signals readable by machines, so AI agents and connected applications can interact with trust information programmatically. Then there is x402, an HTTP-native payment protocol that enables services to request payments as part of digital interactions. With x402 payments live and Concordium involved in the x402 Foundation, identity and payments are becoming part of a much bigger conversation about the agentic economy. Add Concordium's participation in the Mastercard Crypto Partner Program and the developments discussed during Town Hall #6, and there is a broader story worth paying attention to. What makes this interesting is how these pieces connect: Agent Registry helps establish who an agent is. MCP makes its verification signals understandable to machines. x402 provides a way for agents to pay for services. AI agents will need more than intelligence to operate meaningfully in the real world. They will need identity, accountability and the ability to transact. That is the infrastructure Concordium is working towards. A green candle can attract attention, but building solutions for an emerging economy is the bigger story.

What If You Could Own an NFT Without Revealing Everything Inside It? NFTs have always been associated with transparency. You mint a collectible, its artwork and traits become visible, and people immediately start comparing rarity and value. But what if digital ownership didn't require revealing every detail? That is the idea @FungoLabs is exploring on Ethereum through Fully Homomorphic Encryption (FHE), powered by @zama. Its Genesis collection introduces a different approach: NFTs that arrive sealed, with two separate layers created and encrypted by the smart contract. The first is The Look, which contains the artwork. Owners can privately open it whenever they choose, publish it for everyone to see, or keep it sealed. The second is The Essence, containing the NFT's class, element and four stats. These attributes are designed to remain encrypted permanently. There is no reveal function for them. Even more interesting, these attributes are generated through Zama-powered FHE randomness rather than being manually selected by the team. The goal is to make the process fair without exposing the hidden information. Now, privacy isn't limited to the NFT itself. Fungo is also applying this thinking to its Genesis auction. Instead of a traditional auction where everyone can monitor competing bids, participants submit encrypted bids in cZAMA. Their bid prices and quantities remain hidden while the auction is open. Let's make that easier to understand. Imagine there are only 3 NFTs available, but 4 collectors want one: 💧 A bids 3,000 cZAMA 💧 B bids 1,500 cZAMA 💧 C bids 1,000 cZAMA 💧 D bids 500 cZAMA The three highest bidders win. But here is where the auction gets interesting: they all pay the same final price, which is 1,000 cZAMA in this example. (A) bid 3,000 but pays only 1,000, receiving 2,000 back. (B) bid 1,500 and receives 500 back. (C) pays exactly 1,000, while (D) loses the allocation and receives the full 500 cZAMA refund. In other words, bidding higher doesn't automatically mean paying more. Your bid represents the maximum amount you're willing to pay, while the clearing price determines what successful bidders actually pay. These are illustrative figures, not actual auction prices. Bids are binding, and the encrypted bidding fee is non-refundable, so participants should understand the rules before submitting. What makes this particularly interesting is how both ideas connect. Fungo is exploring private NFT attributes and confidential auctions, two situations where revealing information too early can influence decisions and outcomes. And the ambition goes beyond its 3,232 Genesis NFTs. Its Launchpad is designed to help other creators build encrypted collections, customize reveal rules, run sealed auctions and develop games around hidden NFT attributes. For me, this is a more interesting direction for NFTs than simply creating another collection with different artwork. It raises a bigger possibility: what if the next generation of digital collectibles gives owners control not only over what they own, but also over what others are allowed to know about it? That's the experiment @FungoLabs is bringing to the NFT space.