Nostr isn’t just a Twitter clone. Block turned it into its own AI communication layer. Ten new NIPs just shipped. But here’s what nobody’s telling you: when AI agents coordinate on a protocol where you can’t buy attention, the next six months will decide everything.

The Protocol Just Got a Massive Upgrade

On July 5, the Swift Nostr library dropped version 0.6.0—a protocol and tooling release so substantial it’s worth paying close attention to. Ten new NIPs landed simultaneously. That’s not incremental improvement; that’s a leap forward.

We’re talking about NIP-46 (Nostr Connect) for remote signing, NIP-13 for proof-of-work, NIP-23 for long-form content, NIP-49 for private key encryption, and NIP-50 for search capabilities, among others. The team also fixed a critical NIP-44 encryption bug that had previously made payloads non-interoperable across clients—payloads from earlier releases won’t decrypt, but they never interoperated anyway.

This matters because interoperability has been Nostr’s Achilles’ heel. Different clients speaking different dialects of the same protocol created friction. That friction is now significantly reduced.

The Marmot Spec Is Now Official

The Nostr Compass #30 newsletter, published July 8, confirms that the Marmot spec has been marked adopted. MDK cut v0.9.0 through v0.9.3 over just three days—July 6 through July 8—with encrypted group avatars, external signer support, and MarmotKit iOS and Android bindings.

That’s the kind of velocity you see when a project reaches escape velocity. Three version bumps in as many days tells you the development cycle is accelerating, not slowing down.

Mostro also shipped Transport v2 on NIP-44 direct messages with anti-spam gates, and Bitchat 1.6.0 added NIP-13 proof-of-work. The ecosystem isn’t just growing; it’s maturing.

Block’s Quiet Bet on Nostr

Here’s where it gets interesting. Jack Dorsey’s Block has been developing Nostr-compatible products. But the company made two announcements quietly over Easter weekend that deserve far more attention than they received.

First, Sprout—a Nostr relay built specifically for AI and humans to coordinate. It’s essentially a Nostr-native communication layer with AI agents built in, enterprise authentication, and a native MCP server.

Second, Mesh LLM—which lets people pool spare GPU capacity across a mesh network, coordinated over Nostr, so smaller machines can run larger AI models together.

Think about what this means. Block is making Nostr the communication layer for agentic AI. And here’s the kicker: users of these tools won’t even know they’re on Nostr. That’s the point. The protocol becomes invisible infrastructure.

The Attention Economy Can’t Be Gamed Here

One of the most compelling aspects of Nostr that’s finally getting attention is its resistance to traditional attention hacking. As Heather Larson and Derek Ross discussed on the Soapbox Sessions podcast, you cannot buy attention at the protocol level on Nostr.

Your demographic data isn’t available to advertisers. If you want someone to read your message, you pay them directly. Heather has run geo-zap campaigns through Club Orange for as little as five dollars to reach hundreds of people at specific locations. Derek ran one for Nostr Valley targeting people within 100 miles for roughly 20 dollars.

Compare that to the traditional model where marketing agencies spin up fake accounts to simulate organic trends. As Eliza McLamb exposed, agencies are running narrative campaigns for record labels, creating fake fans and simulated organic trends. That’s payola for 2026.

On Nostr, that game doesn’t work. The economics are fundamentally different.

What’s Next?

The W3C Nostr Community Group published did:nostr v0.1.0 on July 2. The ContextVM SDK for JavaScript/TypeScript was released, bridging Nostr and the Model Context Protocol to enable decentralized access to computational services. OpenSats continues to fund projects through The Nostr Fund, with over $32 million allocated to open-source projects across 40+ countries.

The bear case? Fragmentation could still kill momentum. Different implementations, varying client support, and the inherent complexity of decentralized systems remain real challenges.

The bull case? Nostr is becoming the default communication layer for things far beyond social media. AI coordination, GPU pooling, decentralized identity—these aren’t speculative use cases. They’re shipping right now.

Summary

Nostr isn’t just surviving; it’s evolving faster than most people realize. The protocol just received its most significant upgrade yet with ten new NIPs, the Marmot spec is officially adopted, Block is building AI infrastructure on top of it, and the attention economy on Nostr operates on fundamentally different economics than traditional platforms. The next six to nine months will be critical as AI capabilities comparable to what we’re seeing in closed models reach open source. Nostr’s positioning as a decentralized coordination layer for both humans and agents couldn’t be more timely. Whether it becomes the default communication infrastructure of the next internet or remains a niche protocol depends on one thing: whether developers build on it faster than the alternatives. Right now, the momentum is undeniable.