How does having a main agent on one computer running other agents on other computers compare to this: https://www.youtube.com/live/h19gDMGi7Xs?si=15AFSdsaigHU-Itt

Has anyone been digging into the OpenClaw announcement from Build 2026? (Properly kicked off June 2 with the Windows Developer Blog post; main keynote coverage hit June 6.) Quick summary for anyone who missed it — Microsoft is turning Windows 11 into a native host for AI agents. Four-pillar stack: - OpenClaw containers — hardware-isolated agent compartments using VBS + Hyper-V, vTPM-backed enclaves, JSON manifest-driven security policy - MXC (Microsoft Execution Containers) — cryptographically signed, immutable agent images on a "just-enough OS" minimal kernel - Scout — real-time agent telemetry via ETW, sub-second anomaly detection, pushable through Group Policy / Intune - Project Solara — multi-agent orchestration via publish/subscribe message bus, declarative YAML capability definitions OpenClaw on Windows ships as a Feature Experience Pack before end of July, public preview in 24H2, GA early 2027. The article explicitly names Claude Code and GitHub Copilot CLI as supported. Hermes isn't on the named list — which is the bit I'm trying to figure out. Three things I'd love community input on: 1. Has anyone tested Hermes inside an OpenClaw container in Insider builds? OpenClaw started life as an open-source Linux spec, so technically it should host any Python-based agent runtime — but I haven't seen anyone confirm Hermes specifically. 2. Project Solara vs Hermes Kanban — overlapping or complementary? Solara is doing multi-agent orchestration via pub/sub. Hermes Kanban (v0.13+) is also multi-agent orchestration but via a durable task board with heartbeats and zombie detection. Curious whether Hermes-on-Windows would run inside Solara, replace Solara for our use cases, or just coexist. Feels like a real architectural question worth pressure-testing. 3. Nick — separate question for you: Orgo soon to drop Windows VMs in beta. Will the Windows Orgo VMs get OpenClaw / MXC support once the Windows feature lands GA? The unlock I'm imagining is cloud-hosted, OS-level hardware-isolated agent containers running Hermes — but that's three layers stacking which may or may not actually compose cleanly.

Two weeks heads-down on Sonigo's first agentic production deployment — Hermes on Orgo, running end-to-end for one client. Drafts replies across the channels they use, queues for human approval via Telegram, fires only inside safe time windows. Three production lessons that don't seem to be in the YouTube tutorials: 1. Approval-token expiry has to match the use case. 5 minutes works for live ops where you approve and it fires immediately. For batches queued to send in a later window, you need 8–12 hours or the tokens die before they fire and you've burned all the prep work. One env var per tool, configurable per risk profile. 2. Pre-send thread-freshness check is non-obvious but essential. Before any outbound send, the agent has to re-check the live thread on the destination platform — if a human's already replied (because they saw the notification first), refuse the queued send. Without this, double-messages destroy reply rates and signal automation harder than any timing pattern ever will. 3. Curated tool surface > raw permissive surface. Hiding the escape hatches (cronjob / file / terminal / computer_use / delegation) and replacing them with curated wrappers that bake approval gates into the tool layer itself is the difference between "the agent could go off-script" and "the agent literally cannot do X without a human in the loop". Took a real audit pass to find every gap. Worth it. The real question for the room: Sonigo's positioning agentic workflows as the core offering. The next operational unknown for me is multi-client scale — at what point did you graduate from "I personally approve every outbound action across every client" to "I trust the agent to fire X without my touch"? What was the trigger? A specific number of clients? Number of successful sends across a workflow without incident? A specific architectural shift like introducing a reviewer agent? Or did you build it differently from day 1? Trying to design the operating cadence right before it's painful to change later.

Hey everyone, I've been heads down with Orgo and working with real businesses to automate their workflows with computer-use agents, but I want to get way more active in here with content that actually helps you. I've got a bunch of ideas for what to create next and I want to hear from you on what would be most valuable. Drop a comment with your top 1-2 picks (or suggest something I'm missing): Build tutorials & walkthroughs - Step-by-step: building a CUA from scratch (beginner-friendly) - Swarm agents — how to have a main agent spawn and coordinate sub-agents on their own computers - Using the Orgo API vs the playground — when to use what Real business use cases (with demos) - Insurance form automation — filling carrier portals with no APIs - E-commerce creative pipeline — scraping competitor ads, generating variations, testing on Meta - Email-triggered agents — auto-responding to customer inquiries with real data lookups - Data entry for legacy software that has no API ("the API gap problem") Business & monetization - How to sell CUA builds to businesses (pricing, pitching, delivering) - "Boring businesses" you can fully automate with agents - How to land clients Templates & starter kits - Downloadable agent templates you can open in Cursor and start tweaking - Prompt engineering & templates for computer-use (what works, what doesn't) Live sessions - Live build sessions where I build an agent from scratch in real-time - AMAs — ask me anything about Orgo, CUA, building an AI business What's most useful to you right now? And if there's something specific you're trying to build or a problem you're stuck on, drop it below!