GIL OZERI
"The neural scaling laws were just the beginning. The real alignment risk was within the boardroom."
The Parallel Tracks of Theory and Performance
Scroll vertically to explore the comparative trajectory of Theoretical Physicist Jared Kaplan and Alternative Comedy Veteran Gil Ozeri. Left and right channels translate at offset rates, locking into place as milestones intersect the viewport centerline.
Holography & Science Fiction Limits
Stanford University (BS); Harvard University (PhD); doctoral thesis "Aspects of holography" advised by Nima Arkani-Hamed.
Sparked by an early obsession with validating science fiction limits and faster-than-light travel, Kaplan enters theoretical physics.
Improvisational Notes & Optimization
Upright Citizens Brigade (UCB) Theatre student under genre founders Roberts, Walsh, and Diaz.
Earned a legendary reputation for optimization by taking notes harder than anyone in NYC scene history, mapping comedic structure.
Theoretical Fellow & SLAC Research
Research fellow at SLAC / Stanford; mapping mathematical limits of conformal field theories.
Investigating high-energy particles and structural geometry to model complex holographic horizons.
Alternative Comedy Powerhouse "Hot Sauce"
Co-founder of alternative comedy trio "Hot Sauce" alongside Ben Schwartz and Adam Pally.
Scaling unhinged long-form sketches, culminating in direct guidance from Bob Odenkirk and David Cross in Montreal.
Conformal Bootstrap & JHU Professorship
Associate Professor at Johns Hopkins University leading pioneering Conformal Field Theory Bootstrap research.
Utilizing strict consistency conditions to solve quantum field theories without traditional approximations.
Elite Harold Teams & Improv Rigor
Performer on highly disciplined UCB Harold teams including "fwand," "Death by Roo Roo," and "The Law Firm".
Executing narrative threads under intense constraints of the Harold long-form architecture.
Neural Scaling Laws & OpenAI Shift
Joins OpenAI; co-authors foundational "Scaling Laws for Neural Language Models" and GPT-3 papers.
Proving that language model performance scales predictably with compute, data volume, and parameter scale.
Premium Writing & Animated Finales
Core writer and Co-Executive Producer for Netflix’s flagship "Big Mouth" (writing all season finales); co-stars in "#blackAF".
Voicing iconic characters (Brad, Wiggles) and writing complex, long-running character arcs across seasons.
Anthropic Co-Founder & Chief Science Officer
Co-founder and CSO of Anthropic; managing the Claude large language model family and enforcing strict Responsible Scaling Policies.
Enforcing critical alignment thresholds, managing multi-billion dollar compute budgets, and predicting next-generation AI limits.
National Tour Sold-out Domination
Dominating national live stage tours as a core performer in "Ben Schwartz & Friends" across sold-out US theaters.
Performing live, long-form improvised narrative arcs in front of thousands, scaling comedic emergence to national heights.
Mathematical & Behavioral Isomorphisms
Mapping computational law to long-form performance structures
Left side graphs empirical neural network loss decreases under power-law scale (L(P)). Right side structures the recursive state expansion of live improv scenes (Scene_t+1). Simple localized rules, when scaled extensively, lead to complex emergent logic.
Conclusion: Predictability at scale overrides complex customized rulesets.
Comparative analysis mapping Anthropic’s self-supervised **Constitutional AI** critique/revision cycles against Gil Ozeri’s extreme **Endurance Performances** (e.g., the 47-hour *Entourage-athon*). Both systems prove that imposing rigid internal rules prevents output drift.
Conclusion: Rigorous confinement preserves safety, logic, and comic focus.
The "Constitutional AI Meta-Bit" Simulator
In live performances, Ozeri can adopt Jared Kaplan’s precise clinical delivery to put a strict safety constitution on his comedy. Select a raw comedic monologue below, then run the engine to review the real-time alignment critique and clinical revision.
Spread the Strategic FYC Message
Copy and post these messages to target scientific advisory groups and entertainment casting pipelines.
🚨 THE CASTING DECISION OF THE AGENTIC ERA 🚨 There is an inevitable prestige drama coming about the OpenAI/Anthropic split, and the universe has already chosen its lead: Comedian Gil Ozeri as Anthropic Chief Scientist Jared Kaplan. The physical resemblance is uncanny. The chaotic-to-brilliant energy is perfect. Even Gil is tracking it on Instagram: "a lot of people are suggesting this . i might do it on the pod with Tony!" Let's make this happen. #GilAsJared #AnthropicMovie #ScalingLaws #ThePodWithTony
For Your Consideration: Gil Ozeri as Jared Kaplan in the upcoming prestige tech-drama about the AI boom. The physicist behind the 2020 neural network scaling laws meets the comedic genius who watched 83 consecutive hours of Entourage. When asked about the resemblance, Gil responded: "a lot of people are suggesting this . i might do it on the pod with Tony!" Tag @gilozeri and get this discussed on the pod! 🎬🧠⚡️ #JaredKaplan #GilOzeri #Anthropic #PrestigeDrama #AI #Compute