Bing dodges $100B bullet & IVP's Tom Loverro on the looming startup collapse | E1678

Episode Summary

Executive Summary: The episode examines the legal sale of sensitive mental-health data and the risks of confidently inaccurate AI tools, then features IVP’s Tom Loverro on a coming startup funding crunch. He urges founders to preserve runway, prioritize sustainable economics, communicate with investors, and choose engaged board members—while noting downturns can create opportunities for focused companies.

Main Topics: Mental-health data privacy: A Washington Post report described 11 companies willing to sell bundles of sensitive health information. Because HIPAA generally covers healthcare providers and certain entities—not many apps and data brokers—this sharing can be legal. Speakers warn of privacy, safety, discrimination, and fraud risks and call for federal privacy legislation. Generative AI accuracy and overconfidence: The hosts discuss errors in Google Bard’s James Webb Space Telescope response and a Bing demo that invented or misstated financial figures. They argue that polished, rapid responses can make false information especially easy to trust, so AI outputs need verification. A tougher startup fundraising environment: IVP partner Tom Loverro predicts a difficult fundraising period in late 2023 and 2024 as valuations normalize, investors become more selective, and companies approach cash constraints. He cautions founders not to rely on headline amounts of VC dry powder or assume past valuations will hold. Founder survival and financing practices: Loverro advises founders to start fundraising well ahead of needing cash, cut spending to the core business, accept that market pricing may mean a down round, and favor multiple bids. He also explains how liquidation preferences and other deal structures can alter outcomes for founders, employees, and investors. Boards, investors, and operating discipline: The interview stresses regular investor communication, engaged governance, and selecting board members with conviction, preparation, time, and good chemistry. Companies should balance growth with unit economics and experienced operators, while ensuring employees support the shift from growth at all costs. Potential advantages of a downturn: For companies that survive, reduced competition and hiring pressure may improve access to talent, lower some customer-acquisition costs, and enable greater focus on customers and core products. Employees who remain may gain responsibility and advancement opportunities.

Key Arguments: HIPAA does not protect health information in every context; when people share sensitive details with apps outside covered healthcare entities, those companies may be able to sell or disclose it with few legal constraints. AI chatbots can present fabricated details in a convincing format. The Bing examples included financial metrics that did not appear in the source report, so users should verify important claims against primary documents. Large amounts of venture dry powder do not guarantee funding: investors may slow their pace, avoid lower-conviction deals, or be unable to resolve complicated cap-table and valuation issues. Founders should begin fundraising months before cash runs low, because a short runway weakens negotiating leverage and can raise questions about judgment and planning. A lower valuation can be preferable to running out of cash; founders should prioritize company survival and let market demand, including competing bids, help establish price. Good governance is practical support, not merely control: prepared, engaged board members can ask hard questions and help companies navigate difficult decisions. The downturn can reward focused, mission-driven companies by easing talent competition and forcing clearer attention to product-market fit, customers, and sustainable economics.

Data Points: Companies reportedly willing to sell health-data bundles: 11 - A Duke University Sanford School of Public Policy research team found 11 companies willing to sell bundles that included sensitive health-related information. Market capitalization lost after Bard error: About $100 billion - The hosts said Google’s stock fell after Bard gave a false answer about the James Webb Space Telescope during its unveiling. Bing-reported Gap operating margin: 5.9% (reported); 4.6% including impairment and 3.9% excluding impairment (figures cited from the report) - A blog post cited in the episode said the chatbot’s 5.9% figure did not appear in the company’s earnings document. Early-stage founders with fewer than 12 months of runway: Four in five of 450 surveyed founders - A January Ventures survey cited by Loverro found that most surveyed early-stage companies had less than a year of runway. Views of Loverro’s startup-extinction thread: 2.7 million in less than a week - The host cited the reach of Loverro’s viral thread warning of a difficult period for startups. Recommended fundraising lead time: Start about 9 months before cash is needed; raise at least 6 months before it is needed - Loverro’s guidance is intended to give founders time to meet investors, negotiate, and pursue alternatives. Downside liquidation preference example: 2x - Loverro explained that a 2x liquidation preference could return $10 million to an investor who put in $5 million in certain downside sale scenarios. Historical data-broker list of reported rape sufferers: More than 1,000 names offered for $80 - The hosts cited testimony from 2013 about an advertised list that was later removed.

Pivotal Quotes: "Don't make dry powder your fundraising strategy." — Tom Loverro: He argues that the amount of capital nominally available to venture firms is not a reliable fundraising plan for startups. "The most dangerous mistake is to make it a high wire act and just really be out there without much room for error." — Tom Loverro: He warns founders against waiting until they have little runway left to seek funding. "The real takeaway should be that none of this is ready for prime time." — Rachel: She summarizes the discussion of generative AI systems producing plausible but inaccurate claims.

Implications: Users should treat sensitive health apps and AI-generated claims cautiously, while policymakers face pressure to close privacy gaps. Startup founders should plan financing early, emphasize durable economics, and carefully vet investors; disciplined companies may find downturns create strategic advantages.