Episode 823 | Hot Take Tuesday: Is A.I. Killing B2B SaaS?, ChatGPT Ads, OpenClaw

Summary of Episode 823 | Hot Take Tuesday: Is A.I. Killing B2B SaaS?, ChatGPT Ads, OpenClaw

by Rob Walling

41m•March 10, 2026
Listen to Episode

Overview of Episode 823 | Hot Take Tuesday: Is A.I. Killing B2B SaaS?, ChatGPT Ads, OpenClaw

Host Rob Walling is joined by Tracy Osborne (Head of Product, TinySeed & MicroConf) and Anar Volset (co‑founder, TinySeed; founder/managing partner, Discretion Capital) for a rapid-fire Hot Take Tuesday. The group covers whether AI is killing B2B SaaS, recent moves toward ads in ChatGPT/consumer LLMs, the rise of agent apps like OpenClaw, practical experiences with current LLMs (ChatGPT, Claude, Claude Code), and a quick M&A note about stock vs. asset sales for SaaS exits.

Key topics discussed

  • Is AI “killing” B2B SaaS?

    • Debate whether AI will replace subscription SaaS or simply rebrand it (agents, embedded AI, etc.).
    • Market reaction: public SaaS stocks under pressure; some view declines as a buying opportunity.
    • Consensus: SaaS as a category isn’t going away—software will persist—but product go‑to‑market and defensibility will change.

  • LLMs and perceived model progress

    • Panelists’ hands‑on experiences with ChatGPT, Claude, and Claude Code.
    • Observations: incremental improvements rather than sudden leaps; harnesses (tools around models) are often bigger UX wins than raw model updates.
    • Practical failures: ChatGPT sometimes “spirals” on specific tasks (Zapier examples); Claude/Claude Code preferred by panelists for certain workflows and customization.

  • ChatGPT / consumer LLM ads and monetization

    • Anthropic ran a provocative Super Bowl ad; Sam Altman publicly criticized the ad’s portrayal of ChatGPT.
    • Discussion on inevitability of ad monetization for consumer AI (when scaling to billions of non‑paying users).
    • Opportunity for startups: early ad inventory in LLM interfaces could become a low‑cost, high‑ROI channel—like early Google/Facebook paid channels.

  • Agent apps / OpenClaw (formerly Claw Bot)

    • OpenClaw is viewed as an exciting, practical example of a local agent that ties into a user’s files, email, and calendar.
    • Biggest concerns: security, privacy, and reliability of agent decisions (examples of Superhuman auto‑managing email and missing items).
    • Apple as a likely future interface/guardian (secure, device‑centric data nexus) for everyday agent experiences.

  • M&A practical note (from Anar)

    • Most transactions in the $2M–$20M ARR range are buyer‑opaque (often PE‑owned).
    • Rule of thumb: once you get into the ~$1M+ ARR / several million exit price range, stock deals become more common; by ~$5M+ stock purchases dominate.
    • Asset buys can yield more cash at closing but are simpler and often preferred by buyers for tax reasons.

Main takeaways

  • “AI is killing SaaS” is an overblown headline. AI is a powerful new software capability that will reshape products and competition, but it’s still software—incumbents that integrate AI well can survive and thrive.
  • The immediate risk is disruption to incumbents who fail to adapt quickly; the bigger structural change is easier creation of highly targeted apps (the “vibe‑code” niche apps), so incumbents must strengthen defensibility (data, integrations, network effects, deeper UX).
  • Model improvements feel incremental for many practical workflows; the larger wins today come from orchestration, tooling, and “harnesses” (e.g., agent frameworks, code‑aware assistants).
  • Ads for consumer LLMs are likely inevitable; founders should watch for early advertising/placement opportunities in those new channels.
  • Agent/desktop AI (OpenClaw style) is promising but raises legitimate privacy/security concerns—many users will wait for a trustworthy, integrated solution (Apple positioning highlighted).
  • For founders considering exits: know how stock vs. asset sale structures affect taxes (QSBS), legal cost, and buyer preferences—use resources like Discretion Capital’s guide.

Notable quotes / concise insights

  • Tracy: “The landscaper down the street is always going to need some sort of software to help them run their business.” (B2B demand persists.)
  • Anar: The only way SaaS dies is if “we basically invented god”—i.e., superintelligent AI that makes all software moot.
  • Rob: “It’s just software by a different name.” (Agents/AI will be another form of software monetization, not a fundamentally different business.)
  • Practical UX point: harnesses around models (like Claude Code) are often more impactful for users than marginal model version updates.

Action items / recommendations for founders

  • Don’t panic — evaluate how AI can augment your product, not immediately replace it.
  • Make your product defensible: focus on integrations, data scaffolding, enterprise workflows, and UX that’s hard to replicate by quick “vibe code” clones.
  • Experiment with new ad/placement opportunities in LLM interfaces early if they become available—early inventory can be inexpensive and high impact.
  • Test multiple LLMs (ChatGPT, Claude, Claude Code) for different tasks; consider Claude for deeper content/personality alignment and Claude Code for codebase‑aware workflows.
  • Watch agent apps (like OpenClaw) for new UX patterns; think about privacy and trust frameworks you’ll need to support if you grant deeper access to user data.
  • If planning an exit, read the Discretion Capital guide and get early legal/tax advice on QSBS, and the tradeoffs between asset vs. stock sales.

Resources mentioned (links to check in show notes)

  • Discretion Capital M&A guide: discretioncapital.com/guide
  • Mercury banking (episode sponsor): mercury.com
  • Designly prototyping sprint (discount for listeners): designly.co/for‑the‑rest‑of‑us (promo mentioned)
  • Articles / threads referenced in episode:
    • “AI is killing B2B SaaS” (headline article discussed)
    • Articles about OpenClaw / agent apps (e.g., “OpenClaw changing my life”, “OpenClaw: what Apple intelligence should have been”) — see the episode show notes for direct links
  • Startups for the Rest of Us email list — signup at startupsfortherestofus.com for two exclusive episodes and episode recaps

Hosts / guests:

  • Rob Walling — host
  • Tracy Osborne — Head of Product, TinySeed & MicroConf
  • Anar Volset — co‑founder TinySeed; founder/managing partner, Discretion Capital

This summary captures the episode’s main debate points, practical experiences with current AI tools, and tactical takeaways for founders. Check the episode show notes for direct links to the articles and guides discussed.