Before open web standards became widely adopted, sharing information across systems often meant navigating formats, protocols, and permissions. Then one open standard changed everything. Any browser. Any server. Any country. Suddenly the web was interoperable by default. Now think about where AI agents are today. They can read, plan, and act. But without a shared protocol, agent-to-service connections can turn into custom negotiations. Researchers and standards bodies have warned that without open interfaces for AI systems, the future of the web could be dominated by a few proprietary platforms acting as chokepoints between users and content. That warning is not abstract. It is the same structural risk the early web faced before open standards won. So far, we established that the developer's role is shifting upward. Agents absorb execution. Architects set boundaries. Now the question is bigger. What does it mean for the open web itself when agents can plan, sequence actions, and call tools autonomously? AI agents differ from traditional software in a fundamental way. They combine language understanding, planning, and tool use to pursue a goal. They do not execute a fixed script. That distinction matters enormously for what comes next. The Model Context Protocol is an emerging standard for connecting AI agents to external tools, data sources, and services in a structured, interoperable way. Think of it as analogous to how HTTP standardized communication between clients and servers. The key idea is interoperability by design. The success of the web's open standards model suggests that open, widely adopted protocols for AI agents could promote competition and innovation by allowing many independent services to interoperate. Tim Berners-Lee's semantic web concept anticipated software agents that could understand and act on structured information. Modern AI agents connected via protocols like MCP can be seen as a practical step toward realizing parts of that vision. Suppose you want to understand what making the open web programmable actually looks like in practice. WordPress is the clearest example. Its open-source architecture allows third-party plugins and APIs to be used by AI agents to read and modify content in controlled ways. That is not a small surface. WordPress powers a substantial share of the web's content management ecosystem. Research on tool-using language models confirms that giving models access to external tools, such as databases or APIs, significantly improves accuracy and reliability compared with models generating text from internal parameters alone. Respira is exactly that external tool layer, built specifically for WordPress. Early web automation used crawlers and software agents to navigate content. Modern AI agents connected through Respira are a more capable continuation of that lineage. Now, all of this needs guardrails. Research on autonomous agents has found that unconstrained optimization for open-ended goals can lead to unintended behaviors. That reinforces the need for monitoring and human-in-the-loop oversight for agents acting on live systems. Modern web APIs use OAuth 2.0 and similar frameworks to grant third-party applications limited access. Analogous permission and consent mechanisms are expected to be necessary for AI agents on the open web. The European Union's AI Act explicitly addresses transparency, accountability, and risk management for AI systems. [emphasis] These frameworks will shape how agentic services operate. Carefully scoped permissions and audit logs are not just good engineering. They are the architecture that regulatory reality is converging toward. The takeaway, Mihai, is precise. The open web was built on a bet that shared protocols beat proprietary lock-in. That bet paid off. The agentic web is making the same bet again. Proposals for AI-native web protocols include ideas where web resources expose machine-readable affordances, not just data but also allowed actions, so agents can safely understand what operations are supported. Open-source and open-weight models provide an alternative to proprietary AI services, enabling self-hosted agents that interact with web content without sending all data to centralized providers. Respira sits at that intersection right now. It turns WordPress, which already powers a substantial share of the web's content management ecosystem, into a programmable environment for AI agents. That means the open web does not have to wait for a distant future. For you, Mihai, it starts with the sites you already manage.