This is AGI: Sequoia AI Ascent 2026 Keynote

Macro Trends and Market Scale

• Market Size Potential: The total addressable market (TAM) for AI services is projected at approximately $10 trillion, a figure noted as "conveniently round" but indicative of immense scale compared to the $650 billion software TAM of the previous cloud transition.
• Vertical Comparison: The U.S. legal services market alone ($400 billion) represents the same magnitude as the entire historical software market, highlighting the vertical penetration potential of AI.
• Velocity of Adoption: The AI wave is characterized as the "fastest" yet, with the "white space" of companies reaching $1B+ revenue filling rapidly due to the confluence of cloud, mobile, and AI shifts.
• Aggregation of Waves: Current AI capabilities are the additive result of decades of evolution in silicon transistors, networks, the internet, social media, cloud, and mobile devices.

Technological Paradigm Shifts

• Communication vs. Computation: The presentation distinguishes AI as a revolution in computation (information processing) rather than communication (information distribution), noting that the "floor keeps moving" daily as the foundational technology changes.
• Three Major Inflection Points:
  • November 2022 (ChatGPT): Revealed the power of pre-training.
  • O1 Model (Recent): Introduced a second scaling law focused on inference-time compute and reasoning.
  • Claude Code / NOPUS 4.5/4.7: Demonstrated the capability of "long horizon agents" capable of persistence and error recovery.
• AGI Definition (Commercial): While avoiding technical definitions, AGI is functionally defined as the ability of an agent to dispatch a job, recover from failure, and persist until completion.
• Productivity Leap: The market is shifting from "faster horses" (10–40% productivity gains) to "cars" (10–40x productivity gains) that fundamentally alter organizational structures and work nature.
• Agent Capabilities:
  • Reasoning/Planning: Models now sustain progress on complex tasks for hours rather than minutes.
  • Tool Integration: Agents utilize existing SaaS tools (terminal, Slack, web search) as "arms and legs."
  • Self-Improvement: Early examples exist of agents autonomously improving research capabilities (e.g., achieving GPT-2 level in two hours).

Strategic Framework for Founders (The "MAD" Model)

• M (Moats): In a "revolution of computation," durability comes from wrapping around customers ("customer-back approach") rather than the rapidly changing underlying technology; customer needs change slower than model capabilities.
• A (Affordance): Successful products must offer immediate usability ("paths of least resistance") without requiring explanation, contrasting with raw foundation models like "Claude Code" which lack inherent affordance for average users.
• D (Diffusion): A widening "diffusion gap" exists between the rapid rate of capability creation and the slower rate of enterprise adoption, creating significant opportunity for application-layer companies.
• Market Access: The rapid pace of new capabilities means "no lead is safe," allowing new entrants to overtake incumbents regardless of current market position.

The Era of Agents (2026 Outlook)

• Shift from Assistive to Agentic: The market has moved from 2022's "cute but useless" agents to 2026's functional agents capable of long-horizon tasks (e.g., coding, tax fraud detection, generative media campaigns).
• Agentic Scale: A progression from "autocomplete" (assistive) to "agentic development" (human managing agents) to "background agents" (async) and finally "dark factories" (fully autonomous, human-review-free production).
• Economics of Scale: Services are becoming the "new software" because agents are infinitely scalable via compute, lower maintenance than humans, and cheaper per task (tokens vs. salary) despite currently lower general intelligence.
• Timeline Compression: Agents enable "compressed timelines," turning projects that previously took years into ones completable in days (e.g., rewriting 8 million lines of code in six weeks).
• Future Commerce: Emerging scenarios include agent-to-agent commerce, negotiation, and autonomous swarms for tasks like cybersecurity.

Societal and Philosophical Implications

• Cognitive Industrial Revolution: Just as the Industrial Revolution shifted 99%+ of physical work to machines by 2026, AI is projected to shift 99.9% of cognitive work to machines within the near future.
• Alien Design: AI-driven optimization will produce outcomes (chips, buildings, antennas) that are unintuitive to human minds but significantly more efficient (e.g., NASA's evolutionary algorithm antenna).
• Emergence of a New Science: The current "tinkering" phase of AI will likely yield a fundamental science (comparable to thermodynamics) within two decades to formally master AI and consciousness.
• Reimagining Value: As AI democratizes high-skill tasks (e.g., PhD-level insights), human value will shift from execution to the "human connection," relationships, and the unique ability to adapt and find meaning in the process.
• Historical Parallel: The reaction to AI mirrors the shift from realistic art to impressionism/cubism post-photography; the purpose of human work will redefine itself to focus on "what the heart and soul sees" rather than mechanical reproduction.