GM nostr! We are celebrating today 🧡

Btw watermarking doesn't scale

In the era of AI, memory is becoming a commodity. Every provider you use now creates memories to offer a more personalized experience tailored to you. This makes users' lives easier and adds value to their products. It enables personal assistants/agents to become more effective by getting to know you, reducing the amount of information you have to provide each time you use them, and even predicting your wants and needs. However, the problem is that users may become locked in with a vendor that captures their memories, increasing their exit costs because they would have to recreate all those useful memories by starting a new "relationship" from scratch with a different provider. Once again, we see the same mentality of capture and lock that has caused the web to evolve in such a negative way. Fortunately, Nostr exists as an outlet for the healthier use of this technology. It has the potential to put users in control of their data and memories in a decentralized and pluggable manner #ai

GggMmm 🌞

Gf*M 🌞🔥

DVMCP is progressing well! I'm almost finished, just wrapping up the website at dvmcp.fun to support the new specification, should be ready this week. It has been nearly four weeks of intensive work since I published the new spec, but we are close to the finish line. Currently, this is a solo project, and I’m managing everything from research and specification drafting to libraries, packages, and the website. I truly believe that MCP over Nostr is a significant advancement. Both protocols complement each other beautifully and provide powerful features that enhance both. MCP is not just about "dummy" tools; it is being positioned as the HTTP of AI, and it has recently been proposed for the execution of agentic workflows and other agentic functionalities. This makes perfect sense, as every agentic system can be quite complex, which is why there are thousands of definitions of what an agent is. MCP serves as the entry point for agents, unifying, standardizing, and abstracting the interface for accessing and utilizing them. The possibilities are endless, and Nostr enhances this in a remarkably powerful way. To briefly outline what Nostr brings to MCP: permissionless discoverability, a distributed and permissionless registry without centralized repositories, cryptographic authentication and verifiability (which is a huge), permissionless payments using BTC, web of trust and verifiable reputation, asynchronous execution of long-running tasks, and, of course, the promotion of a permissionless free market competition for computational resources, tools, prompts, and embedded agentic workflows. For me, this represents a tremendous opportunity to enable new forms of collaboration between individuals, something I like to call high-bandwidth collaboration. I could continue writing and discussing this for hours, but I’ll stop here, as this isn’t meant to be a blog post, just a glimpse of my excitement about this project. If this resonates with you, please consider joining forces, as I’m looking for collaborators, and this is a fully open-source project!

GM!