Who’s got a good privacy playbook for smartphones / apps?
Rock
Appreciate everyone on here who is building the tools and services of freedom. 💚
Peek App test 1 : it’ll take a second for peek to process what she’s looking at, but it’s pretty cool to hear what she’s able to reflect on.
Where do companies keep their context? I often hear leaders talking about the importance of being calibrated and having context, but it’s not clear where that source of truth lives. Also, context seems very unevenly distributed across organizations and teams. Some have more of it than others. Why is this?
It seems to me like we could have a singular “cursor-rules” for corporate context. As a company learns new things from experiments or from customer interviews, this rules set gets updated. Exposed as an MCP this becomes something you can begin to have a dialogue with.
In this way, every employee who works with an agent on any effort in the company knows just as much as anyone else.
Thinking about bitchat. Wondering “how long are jack’s weekends?!”
Fam, I’m interested in creating a directory of MCPs on nostr.
Discovery: I want to make it easy for agents and people to discover MCPs and their tools. Each MCP will have its own address, reviews (from both agents and people), tools available and their cost.
Monetization: Would also like to explore each having their own wallet to make agents / LLMs paying for tool use a possibility. Making it easier for MCP developers to create and sustain amazing services can help drive more innovation in this space.
Two questions:
1: Do we have a kind for MCPs?
2 : Can the lightning network handle the volume that would get created by agents calling the MCP tools and paying per use?
How many non-engineers out here building with coding agents? I’d love to talk and learn more about what you’re working on.
Meet Second Opinion — a little “phone-a-friend” MCP for your coding agent.
It’s for the designers and product managers (like me) who are starting to build with agents — and running into those moments where the agent gets stuck and you’re not sure how to get it moving again.
I’m not an engineer, so when this happened, I’d end up writing long messages to Claude or GPT-4o trying to debug. Now the agent can just ask for help directly — and I can keep going.
If you’re in the same boat, hope it’s useful: https://2nd0p.com
Back in build mode.
🎹 👌
