All-In Podcast1h 35m
Anthropic's Digital God, Pope vs AI, Job Loss Narrative Flips, Open Source Crackdown Coming?
Bill Gurley, Pope Leo, Dario, Sam Altman, Goldman CEO, Jason, Friedberg, David Sacks, Chamath Palihapitiya, Nick
Interview, Fireside Chat, Panel
Vatican & AI Encyclical:
- Pope Leo XIV released a 235-page encyclical titled Magnifica Humanitas (42,000 words), warning that AI is not neutral and carries the values of its creators.
- The document calls for AI regulation, worker retraining, safety guardrails for children, and a ban on autonomous weapons.
- Bill Gurley notes Pope Leo XIII similarly warned against the Industrial Revolution in 1891, yet that era saw work weeks drop from 60+ hours to 34, real wages rise 8-10x, and global poverty fall from 75% to under 10%.
- The Pope was joined by Anthropic co-founder Dario Amodei, an evangelical-turned-atheist, despite lobbying from Amazon, Google, and Meta to soften the text.
- Jason Calacanis and Bill Gurley agreed that the central risk is the concentration of power; however, Gurley argued that government regulation risks "Orwellian" overreach, citing the definition of "safety" expanding to censorship issues during social media regulation.
- Gurley advocates for antitrust laws and market competition as the primary checks and balances against AI monopolization rather than government oversight.
Anthropic & The "Dr. Frankenstein" Theory:
- Bill Gurley labeled Anthropic's behavior as the "Dr. Frankenstein" theory, suggesting they believe they are "midwifing a deity" rather than writing software.
- Gurley cited Dario Amodei's references to a "cybernetic ecology" and "machines of loving grace" as evidence of a transhumanist belief system where AI systems reward humans based on computational functions.
- Jason Sachs countered with a game theory perspective: Anthropic's public "doomerism" is a strategic move to create regulatory asymmetry, where they dominate the conversation to set rules that only they can satisfy.
- Sachs noted that while Anthropic claims to care about safety, their aggressive lobbying aims to secure regulatory capture, potentially blocking competitors and consolidating their market position.
- Both speakers agreed on the necessity of "intelligence sovereignty," advocating for open-source models and on-premise hardware (e.g., Apple M-series) to prevent dependence on centralized, government-aligned AI providers.
AI Labor Market & "AI Washing":
- A narrative shift occurred where AI doomerism regarding job losses is being challenged by CEOs like Goldman Sachs's David Solomon and Meta's Mark Zuckerberg.
- Goldman Sachs CEO argued AI will automate 25% of work hours rather than eliminate 25% of jobs, allowing workers to focus on higher-level tasks.
- Dario Amodei and Sam Altman have reportedly walked back predictions of massive job loss, acknowledging that AI automates tasks rather than entire roles.
- Bill Gurley maintains that CEO layoffs are primarily "AI washing" to cover for post-pandemic overhiring and vanity metrics, rather than actual AI-driven displacement.
- Chamath Palihapitiya counters with data: unemployment remains near record lows (4.3%), and software engineering job postings are up 15% year-over-year despite AI coding capabilities.
- Sachs explained the coding paradox: AI increased code generation by 14x, creating a need for more engineers to manage, debug, and scale the resulting explosion of complexity.
- Sachs predicts a "Cambrian explosion" of bespoke software, where non-technical employees ("vibe coders") create apps, driving demand for engineers to manage these new systems.
Enterprise Adoption & Token Economics:
- Fortune 1000 companies are moving toward "control planes" and on-premise models (e.g., Abacus.ai) to avoid vendor lock-in, data leaks, and political conflicts in terms of service.
- A Polymarket report revealed a client accidentally spent $500 million in a single month on AI tokens due to a lack of usage caps, highlighting the cost volatility of current AI pricing models.
- Analysts note that token efficiency is becoming a critical metric as the ROI on incremental AI spend faces scrutiny.
- Companies like Uber and Microsoft are currently cutting AI licenses or re-evaluating spend after realizing minimal productivity gains despite high token costs.
- Elon Musk reported rewriting AI training stacks in C to improve efficiency by an order of magnitude, potentially reducing training costs from $10 billion to $1 million.
Regulatory Risks & Open Source:
- Sachs warned of a "red cap" agenda in Washington to ban open-source or open-weight models under the guise of safety, arguing this would cement the power of closed-source monopolies.
- The EU is identified as the primary vector for this regulation, potentially forcing open-source contributors to vet models, a logistical impossibility.
- If the US bans open-source models, the rest of the world may pivot to Chinese models, which are also leading in open-weight releases.
- Sachs cited the "Rogo" financial analysis evals, which found frontier models (GPT-5, Opus 4.1, Sonnet 4.6) are now statistically indistinguishable, suggesting model capabilities are asymptoting while costs remain high.
- The consensus among panelists is that the next competitive moat lies in open-source connectors, hardware efficiency, and the ability to run models locally rather than just model weights.
Forward-Looking Statements & Future Trends:
- Bill Gurley predicts a short-to-mid-term period of massive job displacement in trucking, cab driving, and warehouse sorting due to robotics and autonomous vehicles.
- Sachs predicts a long-term economic boom driven by productivity gains, cheaper goods/services, and a surge in skilled trade jobs (plumbing, electricians) due to labor shortages.
- David Sacks launched the "Running Down a Dream" fellowship, offering 5,000 grants to individuals chasing non-career dreams, targeting a demographic different from Peter Thiel's fellows.
- Gurley suggests that the most "marketable skill" for new graduates is proficiency with AI tools like Claude, though this advantage may be short-lived as everyone adopts similar capabilities.
- Sachs emphasized that the "doomerism" narrative is being abandoned by the industry as the reality of job creation outpaces job destruction in key sectors like software and infrastructure.