The AI Supercycle, California's AI Safety Bill Vetoed, and the Future of AR/VR | E2017

Episode Summary

Executive Summary: The episode argues that AI is the defining supercycle of the era, with broad effects on jobs, startups, and platform power. Jason and Alex celebrate SB 1047’s veto as pro-innovation, warn that open-source AI and vertical apps are the best startup strategy, question OpenAI’s burn-heavy economics, and explore whether AR/VR is finally approaching consumer readiness—while disagreeing on whether AR glasses improve or degrade daily life.

Main Topics: AI as the greatest supercycle (Priority: 5/5): Jason frames AI as a once-in-a-generation platform shift comparable to the PC, internet, cloud, and mobile, but broader in impact because it will reshape work, productivity, and human capabilities across nearly every part of life. SB 1047 veto and AI regulation (Priority: 5/5): The hosts break down California’s vetoed AI safety bill, arguing that model-level regulation would have burdened startups, chilled open-source development, and addressed hypothetical catastrophic risks while ignoring immediate real-world issues like job displacement and misuse. Open-source AI vs. closed platforms (Priority: 5/5): They argue startups should avoid dependence on closed AI providers like OpenAI and instead build on open-source models to preserve leverage, portability, and competitive pressure, especially as trust in platform vendors matters more over time. OpenAI valuation, burn, and infrastructure pressure (Priority: 4/5): The discussion questions whether OpenAI can justify its valuation while spending heavily on compute and data centers, suggesting its economics, capital needs, and reliance on big infrastructure partners could constrain future independence. AR/VR timing and the startup graveyard (Priority: 4/5): Using examples like Google Glass, HoloLens, Magic Leap, and early AR demos, they argue that the hardware and consumer readiness for AR/VR may finally be improving, but timing has historically killed many promising startups. Social norms, privacy, and wearable tech (Priority: 4/5): Jason strongly opposes always-on camera glasses and notification-heavy wearables, arguing they would worsen attention, manners, and public life unless society imposes stricter norms or bans in certain settings. Founder behavior, rule-bending, and startup moats (Priority: 3/5): The episode closes with a broader view that successful founders often bend rules without crossing into fraud, and that durable startup moats in AI will come from unique data, user experience, brand trust, and verticalization rather than raw model access.

Key Arguments: AI is a true supercycle because it will affect nearly every aspect of life, not just a single sector, and may drive major job displacement and productivity gains. Government regulation should focus on concrete misuse and social consequences, not speculative model-level controls that burden startups and open-source projects. SB 1047 was flawed because it targeted large frontier models with heavy compliance, penalties, and reporting requirements rather than specific harmful applications. Open-source AI is strategically important because it reduces dependence on closed platforms that can change prices, terms, or product access unexpectedly. OpenAI’s reported revenue growth is impressive, but its burn, infrastructure requirements, and need for massive data center investment raise questions about long-term independence. Startups should win in verticalized, niche, and trust-based products instead of trying to compete directly on foundation-model infrastructure. AR/VR may now be at a point where hardware quality and consumer interest are finally sufficient for real app development, but notification clutter and camera wearables could make the category socially undesirable. The strongest startup moats in AI will be unique proprietary data, superior UX, and trusted brand—not just access to a model API.

Data Points: OpenAI 2024 revenue run rate: $3.7 billion - Referenced as Reuters-reported current run rate in the valuation discussion OpenAI expected 2025 revenue: $11.6 billion - Used to argue the company could trade at about 13x next year’s revenue OpenAI valuation: $150 billion - Discussed in relation to expected revenue growth and capital needs OpenAI reported burn: $5 billion per year - Used to assess runway and compute costs OpenAI raise size: $5-6 billion - Referenced as the expected funding round to support ongoing losses and infrastructure expansion SB 1047 threshold: 10^26 integer or floating-point operations - One of the bill’s criteria for covered frontier models SB 1047 cost threshold: $100 million or more - Alternative threshold for models subject to the bill Vanta customer compliance speed: 2-4 weeks - Average time Vanta customers are said to achieve SOC 2 compliance Compliance cost reduction: up to 85% - Claim about savings from using Vanta CloudDevs talent pool: 10,000+ senior engineers - Used in the sponsor pitch for Latin American hiring CloudDevs cost advantage: up to 80% cheaper - Compared with local hiring for similar standards Quest 3S headset price: $300 - Cited as a potentially important price point for consumer AR/VR adoption Churn in startup AI usage study: 60-70% planning to move to open source - Referenced as a reported sentiment among ChatGPT builders Apple/Big Tech infrastructure spend: $10-20 billion - Estimated by the hosts as potential spending on goggles/data-center-related bets OpenAI employee count: 1,700 (mentioned as a last-known figure) - Used when estimating salary burn and operating costs

Pivotal Quotes: "We are on the precipice of the greatest super cycle in the history of business." — Jason: Jason explains why AI is the defining technology wave for founders and investors "The fear of like they're going to come and kill us is real... I do think the job displacement is the one that unions and individuals should actually realistically care about." — Jason: Jason distinguishes catastrophic sci-fi risk from immediate labor disruption "I think we've now reached the point when the hardware's gotten good enough. There's enough good apps." — Jason: Jason’s reason for believing AR/VR app development may finally be timely

Implications: Founders should prioritize open-source leverage, vertical products, and unique data over dependence on single AI vendors. Investors should watch infrastructure, burn, and regulation. AR/VR may be viable again, but only if it improves life without worsening attention and privacy.