July 31, 2026
AI office drama enters the chat
qm
This new work AI wants your whole team — but the comments are already side-eyeing it
TLDR: QM is a new startup-focused workplace AI that works in Slack and on the web, aiming to help teams share one smart assistant safely. Commenters were less impressed by the pitch than by the irony: an AI project asking for human-only contributions sparked jokes, skepticism, and confused side-eyes.
QM is pitching itself as the office AI for startups: one helper that can live in Slack, show up on the web, remember what different people and group chats need, and even handle background tasks like checking projects or sorting messages. In plain English, it wants to be less like a solo chatbot and more like a shared digital coworker that can help a whole company without everyone stepping on each other’s toes. It also promises companies can run it in their own setup and choose different AI engines instead of being locked into one provider.
But the real action was in the comments, where readers instantly zoomed past the product pitch and into maximum irony mode. The biggest eyebrow-raiser? QM’s contribution policy, which asks for human-written proposals instead of AI-generated code. That set off a wave of "wait, an AI project that doesn’t want AI-written contributions?" reactions. One commenter basically said the whole thing felt like the industry disappearing into its own weird AI spiral, while another laughed at the idea of an AI-heavy project asking contributors not to let AI "artificially expand" their ideas.
Then came the classic internet pile-on: some users questioned whether QM is really a serious cloud tool or just another thing that feels like it runs on "one Mac/virtual machine" with fancy branding. Others name-dropped rival agent tools and asked whether those are simply better. And in the middle of all that, one beautifully blunt comment cut through the noise with: "What is yc software?" Honestly, that may have been the most relatable reaction of all.
Key Points
- •QM is described as a multiplayer agent harness for startups that supports both individual workspaces and shared collaboration in Slack channels, group messages, and projects.
- •Each person and room has its own scoped memory, files, keychain view, permissions, crons, web apps, and durable sandbox.
- •The platform is designed to be vendor-flexible, with the same core able to run on Pi, OpenCode, Codex, and Claude Code.
- •Example use cases include cross-source search, internal app building, inbox triage, repository work such as tests and PRs, and project updates in shared channels.
- •Its architecture uses a headless core with Postgres persistence, isolated per-scope sandboxes, and a stack that includes TypeScript, Node, Fastify, Bolt, Vite, and Lit, with security modes of Strict, Auto, and Dangerous.