Episode Summary
Executive Summary: The episode centers on Glue, a new enterprise chat platform from David Sacks and Evan Owen that rethinks workplace communication around threads instead of channels, while tightly integrating AI as a contextual “teammate.” The discussion covers product design, Slack migration, AI-powered expertise discovery, pricing, onboarding, privacy, and why offices still matter for building strong teams.
Main Topics: Glue’s thread-first communication model (Priority: 5/5): Glue replaces channel-centric chat with threads as the primary unit of work, aiming to reduce channel fatigue and irrelevant noise while allowing threads to be shared across groups and people. AI as a contextual teammate in enterprise chat (Priority: 5/5): The founders argue that AI should live inside the workplace chat app so it can use conversation history, app integrations, and org context to answer questions, summarize work, and eventually chime in proactively. Slack migration and backwards compatibility (Priority: 4/5): Glue emphasizes low-friction adoption by importing Slack history and selected channels, so teams can migrate without losing organizational memory or starting from scratch. Onboarding and product-market fit through white-glove setup (Priority: 4/5): Rather than letting users self-serve blindly, Glue uses guided onboarding to ensure teams connect coworkers and data sources, making the product immediately useful and lowering churn. Pricing, AI usage, and enterprise economics (Priority: 4/5): The team discusses a $7 per-user monthly plan, included AI usage, optional API-key bring-your-own-model support, and how lower costs than Slack GPT-style enterprise pricing could be a competitive edge. Privacy, model choice, and not training on customer data (Priority: 4/5): Glue says it does not train foundation models on customer data and instead routes queries to models like OpenAI and Anthropic, with automatic model selection based on task and context. Remote work vs. office-based culture (Priority: 3/5): Sacks and Owen argue that offices remain important for mentorship, serendipity, apprenticeship, and management visibility, even if remote work can be a benefit for some senior hires.
Key Arguments: Enterprise chat should unify human conversation, AI assistance, and app context in one place rather than splitting work across separate tools. Thread-based collaboration reduces channel fatigue because people can participate in only the conversations relevant to them instead of joining entire channels. AI becomes far more valuable when it has access to current chat history, company context, and connected tools like Salesforce, Google Drive, GitHub, and Notion. The strongest enterprise AI use cases are managerial and operational: asking what people are working on, identifying subject-matter experts, and surfacing risks or priorities. A product like Glue needs enough initial setup—coworkers, groups, and integrations—to be useful, which justifies guided onboarding instead of pure self-serve. Model quality is improving quickly, so Glue should remain model-agnostic and allow routing across providers rather than betting on one foundation model. Customer data should not be used to train Glue’s models; trust and privacy are core to enterprise adoption. Distributed work can be practical, but fully atomized remote organizations are harder to manage; offices improve speed, apprenticing, and communication. Backward compatibility matters: importing Slack history and memberships reduces switching costs and preserves institutional memory.
Data Points: Waitlist size: 5,700 - Number of people on Glue’s waitlist at the time of the interview. Price per user/month: $7 - Glue’s listed per-seat pricing after the trial period. Free trial length: 3 months initially; 1 month long-term - New signups get an extended launch trial, later standardizing to a shorter trial. Team size: 7 full-time employees - The team size mentioned for Glue’s R&D/build team at launch stage. Model usage included: Some Glue AI usage included with every seat - The product includes baseline AI access in the subscription before usage limits apply. Slack history import: Public chat history and selected channels - Glue can import data from Slack exports to preserve prior context and memberships. Years since Yammer launch/sale: Launched in 2008; sold in 2012 - Sacks references his earlier company to explain why he returned to this space. Twist appearance number: Twist number 5 - The hosts note Sacks appeared on the show very early in its history.
Pivotal Quotes: "I think the AI becomes powerful enough eventually that you don't even have to ask it a question. It'll just jump in and provide help as needed, like a virtual teammate would." — David Sacks: Describing the future of AI inside enterprise chat as proactive and contextual. "We think of it as the hub between your humans, your apps, and your AI." — David Sacks: Explaining the meaning of the name Glue and the platform’s core positioning. "We've basically jailbreaked the threads from the channels." — Evan Owen: Summarizing Glue’s main product idea of making threads, not channels, the primary collaboration unit.
Implications: Glue reflects a broader shift toward contextual enterprise AI embedded in workflow tools. If successful, it could pressure Slack-like products to rethink structure, onboarding, and AI integration while reinforcing the value of office-based teamwork and organizational context.
About This Week in Startups
Jason Calacanis covers startups, tech, markets, media, and all the hottest topics in business and technology. He also interviews the world’s greatest founders, operators, investors, and innovators.