OpenAI should build Slack

Original Article ↗ Hacker News Discussion ↗

Slack vs. Teams vs. Other Chat Apps

  • Many see Teams as a “solid success” only in adoption, driven by bundling with Microsoft 365 rather than product quality.
  • Experiences with Teams are sharply split:
    • Some say it’s “fine” and meets basic enterprise needs (chat, video, calendar integration, recordings, transcription, SSO).
    • Others describe it as slow, buggy, unreliable in messaging, search, notifications, multi-org use, and audio/video handling.
  • Google Chat is viewed as barebones: acceptable for basic chat, worse than Slack on features, but more reliable than Teams by some accounts.
  • Alternatives mentioned: Discord (good product but gamer-branded, not compliance-focused), Mattermost/mostlymatter, Rocket.Chat, Zulip, IRC, Signal (if it had better APIs), and self-hosting.

Slack’s Strengths, Weaknesses, and Network Effects

  • Slack is generally preferred over Teams/Discord for day-to-day work: simple, good UX, “just enough features.”
  • Pain points: heavy Electron footprint, slowness, no code syntax highlighting, unreliable/brittle integrations, perceived quality decline recently.
  • Slack Connect and broad external adoption are seen as its main moat; switching would hurt unless partners move too.
  • Some dislike real-time chat entirely (information overload, “electric shoulder tapping”) and prefer better email tools.

AI, OpenAI, and a “Slack Killer”

  • Some argue chat is a commodity; AI-native workflows (email, scheduling, deployments, approvals, code changes, alerts) inside a chat UI could be a real differentiator.
  • Others push back: Slack’s value is human async communication, and more AI “features” would be distracting or “slop.”
  • Skepticism that OpenAI should dilute focus further (already doing search, images, video, agents, app store).
  • Doubts that OpenAI can build something reliable given current LLM limitations and its own “vibe-coded” tooling.

Trust, Privacy, and Federation

  • Strong concern about handing all internal communications to OpenAI; comparisons to giving all email to an ad company.
  • Questioning whether OpenAI would be any more benevolent than Salesforce; risk of data mining emphasized.
  • Desire for federated or open solutions exists, but commenters note federation conflicts with most corporate incentives and has historically failed at scale.