Held SeatGeek for 14+ years; now a top-three global ticketing business, illustrating patience in compounding winners.

OpenAI and Anthropic are the Meta and Google of this era; trillion-dollar outcomes highly likely despite massive roadkill elsewhere.

Anthropic alongside OpenAI defines the generational platform layer; trillion-dollar trajectory despite broader AI roadkill.

GP as largest LP creates alignment for distributed proceeds (DPI) not management fees; discipline maintained despite pressure to raise larger funds.

Wix trading at 2.1B market cap on 2.1B revenue exemplifies irrational pricing that cannot be fought; don't fight the tide.

Multi-stage platforms returned well historically but post-2020 vintages show TVPI without DPI; sovereign wealth LPs prioritize IRR over multiples.

a16z's large funds returned well historically but post-2020 vintages show TVPI without DPI; sovereign wealth LPs prioritize IRR over multiples.

Ownership discipline (1-2% target) caused passes on Eleven Labs, Granola, StarCloud, Fractile — hundreds of millions in lost returns.

Cursor's valuation trajectory ($26B post, potential $60B exit) exemplifies the new reality where $1B+ valuations are effectively the new Series A price for elite AI companies.

Cognition's $26B valuation is cited as evidence that top-tier AI companies now command billion-dollar-plus entry prices, reshaping seed fund economics.

Mercor's $20B valuation is cited as evidence that top-tier AI companies now command billion-dollar-plus entry prices, reshaping seed fund economics.

All hardware (drones, cameras, sensors) is now commoditized; the defensible moat is the AI software stack that makes autonomous systems work — applied AI on cheap platforms.

AI-driven innovation cycles are compressing SaaS lifecycles; Airtable faces cannibalization before liquidity events can occur.

Snyk experiencing challenging growth rates similar to Airtable as AI accelerates competitive displacement in cybersecurity.

The SaaS selloff has thrown the baby out with the bathwater; deeply embedded, mission-critical SaaS (Veeva, Olo, Airtable) are hard to displace, and a basket of these discounted l…

Founder's deep vertical edge (uncle's HVAC roll-up) unlocks 100k+ mechanical engineer TAM; AI automates blueprint-to-quote workflow.

Founder Collective sold some Uber around $10B valuation (likely too early) but remained net long through IPO; illustrates the difficulty of timing exits in compounding winners.

Whoop's hardware path required long capital-intensive journey with high dilution; Suno's software speed minimized dilution.

Google's search-with-context capability and AI-native DNA put it in pole position versus Microsoft's second-rate AI execution.

Microsoft has done a crappy job on AI; its offering feels second-rate compared to top-tier models from OpenAI, Anthropic, and Google.

Enterprises will spend $100-200M on single model outputs for critical decisions (P&G, Coca-Cola, Nvidia, Visa, governments).

Frankel cited his role in Shield AI's first institutional check and argued that the United States needs defense AI because adversaries have powerful capabilities, highlighting his…

Photonic/optical chips will disrupt Nvidia's electrical architecture within 10 years, solving data-center energy constraints.

Thoma Bravo take-private at $2B after 17 years yielded modest IRR but extraordinary founder journey and board experience.

Deeply involved from seed; believes Suno's consumer music product can transition from creation tool to consumption platform disrupting Spotify/Apple Music.

Framework-driven discipline caused pass on Klaviyo despite loving founders; illustrates cost of valuation rigidity in winner-take-most markets.