Conference Presentation, Panel, Fireside Chat
AI: Hope, Hype, and Headwinds | Milken Institute Global Conference 2024
- Los Angeles aims to establish itself as a global hub for creativity and innovation by hosting the World Cup in 2026, the Super Bowl in 2027, and the Olympics and Paralympic Games in 2027 and 2028 respectively, drawing millions of attendees from over 200 countries.
- The city is executing a $30 billion capital improvement program designed to create world-class experiences and connect the airport to the Metro rail, with a parallel goal of achieving 100% clean energy by 2035.
- Generative AI is characterized as a fundamental technological shift comparable to the early internet, expected to reinvent every daily product and process while enabling organizations to invest in growth trajectories similar to multiple Amazons.
- The technology landscape anticipates a step function change in problem-solving complexity and accessibility, with compute capacity following Moore's law via GPU advancements and the Blackwell chip delivering 30 times more training performance than A100s at one-fourth the energy cost, expected in the fall.
- AI integration is predicted to rewrite job functions, requiring companies to re-architect departments and work processes, while education systems may need re-architecture as current assumptions become obsolete.
- Specific enterprise applications include Salesforce Einstein AI for summarizing customer interactions, automating documentation matching, and acting as real-time co-pilot assistants, alongside Amazon's use of AI for code generation, testing, and pull request reviews.
- Regulated industries such as healthcare, life sciences, and finance are positioned to rapidly adopt AI to detect oncology targets, analyze historical insurance policies for risk, and provide doctors with complete patient history, potentially leapfrogging digital-native competitors.
- The emergence of autonomous AI agents is imminent, with capabilities to predict outcomes, execute actions, and refine workflows independently, though security constraints and guardrails are being developed to manage deployment risks.
- Market adoption is currently focused on ancillary business tasks and incremental revenue or cost savings of approximately $1 billion annually per customer, with projections indicating a shift toward core business functions over time.
- Forward-looking timelines suggest the technology is currently in the high-growth section of an S-curve, with a hockey stick inflection point expected over the next 12 to 18 months and the middle of the high-gradient phase reached in 24 to 36 months.
- Future model capabilities, such as GPT-5 or 6, are anticipated to demonstrate advanced reasoning abilities comparable to Putnam competition levels, while synthetic data generation offers potential for training but carries risks similar to synthetic CDOs.
- Risks associated with widespread AI adoption include disinformation, discrimination, data privacy/security breaches, and workforce displacement, necessitating a massive expansion in cybersecurity.
- Employment dynamics may shift toward larger, higher-performing software development teams due to increased efficiency and ROI, though the ratio of AI contribution to human input could reach 90-10 or 95-5, creating potential visibility issues.
- Housing and homelessness challenges in Los Angeles are expected to be addressed by AI speeding up permitting processes, while general societal problems such as foster care placement and medical issues previously unsolvable are targeted for resolution.
- Open source models are identified as a capital-efficient business strategy, with expectations for an open source GPT-5 and the possibility that current seed companies will pivot to Series A funding as they find product-market fit.