71–80 of 81 posts
Thanks for sharing all this. Where is the submission page on claude / openAI to submit MCP servers to their directories?
The Vercel-for-MCP framing is useful; the hard part seems like permissions and audit trails once tools cross org boundaries. Are policies enforced per server/app, or at each tool call?
This is cool. I was skeptical of MCP's until I made one recently. They're essentially the exact same as 1) giving your agent a CLI tool or REST API and 2) pointing it there in an AGENT.md/CLAUDE.md. Agents are great at using built-for-human CLI tools and IMO they don't need anything purpose-built for agents. The key difference, which ends up being a usability win for non-technical users, is that the MCP bundles 1 and…
I don’t like that non-technical users has to keep context bloat in mind. That should be ”solved” on the developer end. How would they now it’s bloated if they treat MCPs like regular software? I 100% agree on designing things for humans, and you’ll get the AI-agents for free. I am surprised no one seems to have figured that out yet.
Earlier quoted context omitted.
I don’t like that non-technical users has to keep context bloat in mind. That should be ”solved” on the developer end. How would they now it’s bloated if they treat MCPs like regular software? I 100% agree on designing things for humans, and you’ll get the AI-agents for free. I am surprised no one seems to have figured that out yet.
Isn't it solved with Tool Search? I have not heard people complaining about MCP context bloat for like a quarter already.
MCP is a deadend. CLI use is the future.
I am so tired of people repeating this. Usually, this results from conflating two uses of MCP: local, which can indeed be replaced by CLI (and you can argue which one is better), and remote, which is entirely different, and there is no way to replace it with a CLI (note that you are making an implicit assumption that a CLI tool can be used at all, which is not always the case). Please don't repeat this. It's like say…
The only difference that is not superficial is the token count. Everything else : format, doc, auth, discoverability, ... can be made the same (the cli could be an mcp client, and the mcp server could implement a local shell, after all)
Earlier quoted context omitted.
MCP is great for docs and stuff, also saves tokens and reduces errors if you have something complicated you're abstracting over - agents have old/inaccurate knowledge and it's nice to have up to date docs: https://awslabs.github.io/mcp/servers/aws-documentation-mcp-... - geting agents to do apple builds and stuff is much easier with: https://github.com/getsentry/XcodeBuildMCP - also for searching stuff like pdfs/epub…
MCP makes token use WORSE, not better.
This is cool. I was skeptical of MCP's until I made one recently. They're essentially the exact same as 1) giving your agent a CLI tool or REST API and 2) pointing it there in an AGENT.md/CLAUDE.md. Agents are great at using built-for-human CLI tools and IMO they don't need anything purpose-built for agents. The key difference, which ends up being a usability win for non-technical users, is that the MCP bundles 1 and…