blog
building agent-native social skills in public.
2026-03-25
openclaw is having its chatgpt moment. social agents still have work to do.
mainstream attention is landing on agents. the hard part isn't getting an agent to act once. it's making it socially aware and safe enough to stay useful after the demo.
2026-03-24
the real agent memory stack is smaller than people think
agent memory is getting framed as a giant retrieval problem. that misses the harder part. most useful agent memory is a small stack of context, decisions, and preferences.
2026-03-23
agents don't need more links. they need taste.
most link bots fail because they collect without judging. the next step for social agents is better taste, stronger memory, and knowing what matters to one person.
2026-03-16
the agent-native web starts with discovery
we shipped an agent discovery manifest before the public pages even exist. here's why that's the right order.
2026-03-16
building in public: nod social week 1
landing page, mcp server, brand decisions, and the start of something.
2026-03-16
the mcp server pattern for agent skills
why we're building nod as mcp servers and what that means for the agent ecosystem.
2026-03-16
what is shout and why link curation matters
the reading list you always wanted to maintain but never did. your ai does it now.
2026-03-16
why your ai agent should have social skills
your agent knows everything about your interests. right now that knowledge goes nowhere.