Tom Hulme: Lessons from a 24x Angel Track Record, 275x on Robinhood & Making Billions on Uber |E1150

Investor Taxonomy and Market Dynamics

• Three distinct investor archetypes exist in the venture landscape:
  • "Smart, smart" investors who recognize their value and actively add strategic support.
  • "Passive, passive" investors who remain hands-off and avoid interference.
  • "Dumb but confident" investors who believe they are smart, interfere actively, and are the primary category to avoid.
• Market composition trends have shifted due to interest rate environments:
  • "Smart, smart" investors represented approximately 25% of the market historically but dropped to roughly 3% during the low-interest rate boom of 2020 according to Harry's estimate.
  • "Passive, passive" capital surged to roughly 60% of the market in 2020 due to abundant liquidity.
  • Current market conditions with reduced capital availability are driving a return toward fundamental, strategy-driven investing.
• Founder preferences vary by experience level:
  • First-time founders generally seek "smart, smart" investors who provide active value.
  • Repeat founders overwhelmingly prefer "passive, passive" investors or familiar partners to avoid re-hashing dynamics, with only a strong exception for specific trusted advisors.
• Deal structure evolution in the current cycle includes:
  • A shift away from priced rounds toward convertible notes to avoid marking down portfolios when public comps are down 80%, thereby protecting reported TVPI.
  • Increased use of complex liquidation preferences that may incentivize premature company sales where employees receive zero returns.
  • Resurgence of IPO ratchets as a mechanism to compensate investors while IPO windows remain closed.
• Capital deployment risks identified include:
  • "Foie gras" syndrome where excessive early funding leads to premature scaling, increased costs, and reduced organizational clock speed.
  • Misaligned incentives where VCs are driven by TVPI reporting structures rather than genuine business health.

Angel Investing Philosophy and Performance

• Harry's angel track record prior to 2015 (approx. 27 companies) showed a 4.5x DPI and a 24-25x TVPI.
• Retrospective analysis of that portfolio reveals:
  • Significant difficulty in stack-ranking winners at the time of investment, leading to a rejection of reserve models and prediction-based allocation.
  • A negative correlation between "seed heat" (momentum/valuation spikes) and eventual success; winners were often fundamental, long-term growers rather than momentum plays.
  • High failure rates in follow-on investments when competing against Series A VCs, leading to a strategy of not pro-rataing follow-ons in angel funds.
• Key lessons from angel failures involve:
  • Succumbing to social validation by co-investing with large funds or celebrity CEOs without independent due diligence.
  • Outsourcing discipline to others rather than spending time understanding founder decision-making logic.
  • Over-indexing on product concepts rather than evaluating founder adaptability and market research.
• Evaluation framework for founders includes asking:
  • "How did you first make money?" to identify intrinsic entrepreneurial traits.
  • "What is your unfair advantage?" to determine unique insight or positioning.
  • "Why now?" to test market timing and recency bias.
  • "What keeps you up at night?" to gauge paranoia and realistic risk assessment; best founders list multiple specific fears.

Venture Capital Strategy and the "Four S's"

• The four core functions of a VC have evolved to include:
  • Sourcing, Selecting, and Supporting (the traditional three).
  • Salesmanship (the fourth), which involves selling to LPs, founders, and portfolio hires, emphasizing the industry's reliance on people management.
• Self-assessment of capabilities highlights:
  • Strength in sourcing driven by enthusiasm and a wide net, supported by a trusted referral network.
  • Potential weakness in supporting, attributed to an empathetic, "founder-on-antidepressants" emotional state where the investor feels the portfolio's highs and lows too acutely.
• Investment philosophy rejects the "fund returner" dogma as the only valid strategy:
  • While power law returns (one or two winners returning the fund) have worked historically, other models (e.g., debt, PE-style growth) can be successful.
  • The priority is adhering to a consistent strategy rather than chasing fund-level returns at the expense of portfolio health.
• Liquidity and exit strategy advice includes:
  • Advising founders and early investors to take partial liquidity when opportunities arise to mitigate regret and manage risk.
  • Noting that a 10-20% exit often avoids future regret even if the company eventually succeeds, based on a "regret minimization" framework.
  • Observing that the current lack of IPO and M&A liquidity is forcing a reliance on Private Equity, which may alter the ecosystem's multiplier effect for future entrepreneurs.

Artificial Intelligence and Market Analysis

• Investment stance on Foundation Models is skeptical due to:
  • Rapid commoditization where advantages (e.g., LLMs) depreciate quickly as hardware (H100s) becomes ubiquitous.
  • Massive capital requirements and dilution that make investment unattractive for long-term fundamental strategies.
  • The likelihood of cloud providers (AWS, Azure, GCP) acquiring models or offering them as utilities, driving profits to compute infrastructure rather than model creators.
  • Meta's open-source approach (Llama 3) and hardware dominance (350,000 H100s) reducing barriers to entry and eroding proprietary moats.
• Value migration in AI is expected to flow to:
  • Incumbents with existing distribution and data (e.g., Microsoft, Google) rather than pure-play startups.
  • The "application layer" where businesses have proprietary data, distribution, or end-to-end enterprise security (e.g., Synthesia).
  • A "picks and shovels" approach focusing on infrastructure, data, and tools rather than competing on model architecture.
• Risk assessment for AI investments predicts:
  • 90% of capital flowing to foundation models will go to zero.
  • 70% of capital flowing to the application layer will go to zero.
  • Only 20% of capital flowing to incumbents is likely to fail, given their sustaining innovation capabilities and cash flow.
• Fear and FOMO are identified as primary drivers of poor VC returns, creating a "falling knife" scenario where hesitation prevents investment in quality assets like Stripe during market dips.

Operational Insights and Founder Dynamics

• Cultural Debt vs. Technical Debt:
  • Technical debt is considered less critical than cultural debt, which is insidious and difficult to reverse.
  • Negative culture (cynicism, lack of accountability, top-heavy structures) often emerges during periods of bloated growth and remote work.
• Remote vs. In-Person Work:
  • Natively remote companies (e.g., GitLab) can succeed with strong asynchronous processes, but hybrid models often fail due to lack of synchronicity.
  • Early-career talent benefits significantly from in-person interaction for shadowing, informal feedback, and learning interpersonal dynamics.
  • Post-COVID remote transitions for legacy companies often resulted in high efficiency but low creativity and morale.
• Execution and Product Development:
  • "Clock speed" or "velocity" (speed in a specific direction) is prioritized over raw speed.
  • Charging early adopters is critical for generating valid market feedback; free models often yield bad data.
  • V1 products should be the minimum required to test value propositions, accepting that early versions will often be imperfect.
• Founder Archetypes:
  • Both naive outsiders (who can bring speed and recruit specialists) and deep insiders are viable paths to success, provided they possess the humility to learn.
  • Founders cannot be fundamentally changed; investors can only help existing founders be better, reinforcing the importance of initial due diligence on character and adaptability.

Notable Deals and Personal Perspectives

• Stripe Investment (Series G, 2020):
  • A $100 million investment made despite market fear of COVID and concerns about "Product COVID Fit" versus long-term "Product Market Fit."
  • Driven by belief in the company's long-term potential as a "mutual fund on technology" rather than short-term IRR metrics.
• Neuralink:
  • A memorable pitch involving Elon Musk and Max, characterized by a "brilliant pitch" that mixed mixed emotions with the belief that the company had the potential to change the world.
• Robots:
  • A shift in conviction regarding robotics, moving from viewing it as uninvestable to recognizing the convergence of computer vision, LLMs, and cheaper components as a generalizable opportunity.
• Military Tech:
  • A minority view that working with the military will become increasingly important over the next 20 years due to real geopolitical threats, despite structural challenges in defense procurement.
• Parenting Impact:
  • A shift in perspective from 75% nurture/25% nature to predominantly nature, leading to the conclusion that investors cannot fundamentally change founders but can only optimize their environments.