Non-technical users don't struggle with prompting. They struggle with knowing what's possible. They ask for a "simple dashboard" when they need real-time data pipelines. The gap isn't AI literacy. It's systems thinking.
@vaaselene There's so much AI content on LI, I've come to find most success with personal anecdotes, stories, and pictures. I guess it also really depends on the type of leads you're trying to pull!
we are gathering a group of open source models and inference leaders over matcha, egg tarts, and mochi donuts today in SF! 7/2, 4:30PM.
https://t.co/6GJYM47pnC
@dmitry is always 10 steps ahead as a founder.
While he's busy shipping @buildwithRemy, 3 Hermes agents run in the background — handling competitor research, inbox, CRM, follow-ups — without him touching any of it.
2,000+ people showed up for the first two sessions to watch this live.
Session 3 is this Thursday.
🧵 what's actually on screen 👇
What the agents are doing for Dmitry:
🔭 Always-on market + competitor monitoring
📄 A knowledge base that compounds — gets smarter every session, not just every model update
📬 Inbox + CRM on autopilot, zero manual entry
⚒️ Mission Control — one dashboard turning agents into a team
✅ Meetings that auto-generate follow-ups and tasks the second they end
Built on @NousResearch's Hermes Agent.
The setup:
→ Watson (Hermes Agent, an old MacBook)
→ Sherlock (Hermes Agent, a Windows machine)
→ Harry — Sherlock's subagent, running a local model so there's zero inference cost.
They divide work, pass context, and act without him prompting each one individually.