We've integrated $CVT as a payment option on Switchboard. @LaunchCovenant is building the governance layer for AI agents. We're building the exchange those agents route inference through. Now their token works directly on our platform - deposit $CVT, we do the rest. The agent economy works better when the pieces connect. This is one of them.

We've integrated $CVT as a payment option on Switchboard. @LaunchCovenant is building the governance layer for AI agents. We're building the exchange those agents route inference through. Now their token works directly on our platform - deposit $CVT, we do the rest. The agent economy works better when the pieces connect. This is one of them.

Talk to Me Tuesday · Alex Salsali (Covenant) — AI Agent Governance Live | Crypto Bros nitter.net/i/broadcasts/1rGmqpRRp…

UPDATED guest card — Talk to Me TUESDAY 🔥 Alex Salsali @AlexJFRSalsali (Covenant) live with us — AI agent governance. Come watch → @AlexJFRSalsali @LaunchCovenant @innit2winniTllc @XRPGoldGod NFA

The Covenant Framework organizes agents into domains that mature through four stages: Forming, Storming, Norming, Performing. We visualized this using L-system fractals grown by branching grammars of increasing complexity The same way a snowflake crystallizes from a simple seed into an intricate structure.

I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.

$CVT Covenant Foundation: Building the Governance Layer for the Agentic Future The next phase of AI isn't simply about making models smarter. It's about giving AI the ability to act. Agents will increasingly write software, operate businesses, interact with financial systems, communicate with other agents, retain long-term memory, and execute tasks with decreasing levels of human supervision. That creates an enormous problem that doesn't get nearly as much attention as AI capability itself: Who governs the agents? 🤔 This is where @LaunchCovenant Foundation becomes interesting. $CVT is approaching AI from the opposite direction of most projects. Instead of asking "How do we make an agent capable of doing more?" they're asking: "How do we make increasingly autonomous agents operate within persistent, enforceable rules?" Their framework introduces a constitution, specialized agents, enforcement hooks, persistent memory and multiple layers of governance—including hard rules, heuristics, examples and behavioral defaults. In other words, they're attempting to build structure around autonomous intelligence. And that may become increasingly important. Today's AI assistant usually exists inside a session. You ask something. It responds. The interaction ends. Tomorrow's agents won't necessarily work that way. They could operate continuously, maintain memory, interact with software and other agents, spend money, make decisions and execute complex workflows over long periods of time. At that point, prompting alone isn't governance. You need permissions. You need memory. You need identity. You need auditability. You need rules that persist even when the human isn't standing over the agent. And ultimately, you need infrastructure capable of answering one fundamental question: What is this AI allowed to do? That's the problem Covenant's developers appear to be thinking about remarkably early. The project describes governance across multiple depths rather than relying on a single system prompt. It also emphasizes persistent memory—allowing the system's understanding of intent to survive individual sessions and evolve over time. That distinction matters. The more autonomous AI becomes, the less useful a brilliant agent is if nobody can reliably constrain, audit or understand its behavior. The developers behind Covenant appear to recognize something that could become obvious several years from now: The AI revolution won't only need intelligence infrastructure. It will need governance infrastructure. The internet needed identity, permissions, security protocols and operating systems as it matured. Cloud computing needed IAM. Financial systems needed authorization and audit trails. Autonomous AI may eventually require its own equivalent. Whether Covenant ultimately becomes that infrastructure remains to be seen. This is still an early project, and execution, adoption and developer traction will determine whether the thesis becomes reality. But the underlying problem they're attacking is difficult to dismiss. Everyone is racing to give AI more power. Covenant is asking how that power should be controlled. If autonomous agents become as important as many people expect, that could turn out to be a very important question to have started answering early.