The Twenty Minute VCCitrini Research Breakdown: Agents, "Ghost GDP", Consumer Spend | Figma Earnings Beat
At a glance
WHAT IT’S REALLY ABOUT
AI agents reshape software economics, markets panic, winners fight back hard
- Anthropic’s security product announcement triggered a sharp cybersecurity stock selloff, which the panel argues was partly an overreaction to “already-possible” capabilities but rational given how perfection-priced multiples amplify small increases in perceived risk.
- A central debate emerges on whether AI agents will “maim” SaaS incumbents by siphoning incremental value (turning apps into commoditized databases) versus the counterview that incumbents and AI-native vertical startups will mediate foundation models into enterprise workflows.
- The group critiques the Citrini Research “2028 crisis” framing by separating micro adoption realities from macro conclusions, while still engaging the “Ghost GDP” fear that productivity gains accrue to fewer people and could soften consumer spending in the short term.
- OpenAI’s leaked spending and revenue projections are framed as ambitious but belief-dependent, with concern about massive forward spending on future products, alongside recognition that frontier AI firms are historically unprecedented in growth.
- Figma’s strong quarter is used as a case study of an incumbent “fighting back” with AI integration, while also highlighting the difficulty of underwriting durability when foundation models may replicate core design capabilities.
IDEAS WORTH REMEMBERING
5 ideasMarket selloffs can be “old news” repriced, not new capability.
The panel notes that Claude-powered security auditing and even penetration testing already existed in workflows like Replit/Vercel; the shock came from a high-salience launch that forced investors to re-underwrite competitive risk.
When a stock is priced for perfection, small narrative shifts cause large drawdowns.
CrowdStrike is cited as an example where strong growth and margins still sit on lofty multiples, so any incremental chance of disruption—however remote—compresses valuation quickly.
Agent disruption may be less about replacement and more about value leakage.
Even if incumbents survive (e.g., DocuSign, DoorDash), the agentic layer could capture the incremental value and slow growth enough to “maim” outcomes for shareholders.
Public incumbents are behind on agents because deployment is operationally hard.
Jason argues agents remain effectively custom: they require onboarding, data cleansing, tuning, and scarce forward-deployed talent, which large horizontals serving many verticals struggle to execute at scale.
A realistic AI thesis needs micro adoption judgment before macro panic.
Rory rejects “civilization ends” framing and pushes a two-step evaluation: first assess whether specific industries will actually be displaced quickly; only then reason about GDP, labor, and consumption effects.
WORDS WORTH SAVING
5 quotesWhen you are priced for perfection, anything less than perfection will be a kick in the nuts.
— Rory O’Driscoll
Every... The further you go, the more folks are at risk of being maimed by AI. Just maimed.
— Jason Lemkin
Ghost GDP is that this productivity is not going to human workers that then spend it.
— Jason Lemkin
Across the scope of history, productivity is fricking awesome. It's the only thing that's made us rich.
— Rory O’Driscoll
Almost all the B2B software we use today is terrible now. It's terrible.
— Jason Lemkin
High quality AI-generated summary created from speaker-labeled transcript.
