1. Opening
- Sequoia partners open with Don Valentine's framework: What's happening? Why does it matter? Why now? What should we do?
2. AI Market: Current State and Future
- Cloud hit $400B in revenue surpassing software. AI's starting point is much larger.
- AI threatens profit pools of both services and software.
- Technology diffusion is accelerating: "ChatGPT launched and the whole world noticed AI. 5.6 billion people are now connected."
3. How to Win in AI
- Past winners were mostly in the application layer; AI will be the same.
- But foundation models are now reaching deep into the application layer.
- Startup strategy: "Start from the customer. Focus on vertical or function specialization."
The 5% That's Different in AI Companies
- Real vs. Vibe Revenue: Check adoption, engagement, retention -- not just hype.
- Margins: Current COGS will decrease. Move up the value chain.
- Data Flywheel: Must answer which business metrics it moves.
- Nature Abhors a Vacuum: "The opportunity suction is deafening. Run at full speed."
4. AI Now: A Year in Review
- AI app engagement soaring (ChatGPT now at Reddit-level DAU/MAU)
- Applications: advertising, education, healthcare, voice AI, coding
- Claude 3.5 Sonnet launched AI coding mainstream
- Pre-training scaling slowing, but new breakthroughs in reasoning, synthetic data, tool use
5. Agents and Value Creation
- First generation of AI killer apps has arrived
- Vertical agents: "AI penetration testers outperform humans; AI network troubleshooters beat the best humans."
- Abundance era: "Code enters abundance first. 'Taste' becomes the scarce asset."
6. Long-term: Agent Economy
- From agent swarms to agent economy: agents will exchange resources, transactions, and trust.
- Three technical challenges: persistent identity, communication protocols, security.
- We need probabilistic (not deterministic) mindsets and management skills.
"Now is the time to run at full speed." "Customer trust is more important than the product." "AI is imminent. All conditions are in place."