/
© 2026 RiffOn. All rights reserved.

Get your free personalized podcast brief

We scan new podcasts and send you the top 5 insights daily.

  1. Grit
  2. The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos
The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos

Grit · Sep 7, 2026

Arena CEO Anastasios Angelopoulos on the AI leaderboard, the rise of Chinese open-source models, and why the belief in API lock-in is a myth.

Stanford's Inflexible Leave Policies Cause Faculty Brain Drain to Startups

Stanford requires faculty who start companies to return after a two-year leave, an inflexible policy that forces top academics like Fei-Fei Li to leave permanently. This creates a "logistical nightmare" for departments and drains the university of its most impactful, real-world talent.

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos thumbnail

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos

Grit·2 months ago

Arena's $100M Revenue Engine Is Private Model Evaluations, Not Its Public Leaderboard

While known for its public AI model leaderboard, Arena's core business is a paid "evaluation" product for AI labs. Labs pay Arena to analyze dozens of their private, internal model checkpoints weekly, providing the critical performance insights that drive Arena's rapid revenue growth.

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos thumbnail

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos

Grit·2 months ago

Chinese AI Models Are Surpassing US Counterparts Through Innovation, Not Just Distillation

The "distillation narrative"—that Chinese AI labs only copy American models—is being proven false. Kimi, a Chinese model, ranked #1 on Arena's web development leaderboard, showing they are pushing the frontier beyond US labs and signaling a potential "flippening" in AI dominance.

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos thumbnail

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos

Grit·2 months ago

Most AI Benchmarks Are Flawed Because They Test Abstract Puzzles, Not Real-World Utility

Many AI benchmarks focus on arbitrary, synthetic tasks (like a "pelican riding a bicycle" SVG test) that don't reflect real user workflows. This creates a disconnect where models top leaderboards but fail at practical jobs. True value is measured by observing users getting their work done.

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos thumbnail

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos

Grit·2 months ago

High API Spend Creates an "Easy Come, Easy Go" Dynamic, Undermining Vendor Lock-In

Contrary to the assumption of strong vendor lock-in, companies spending heavily on AI APIs are highly motivated and able to switch providers. If an open-source model can perform 90% of the work at 10% of the cost, this presents a significant structural problem for incumbent API providers.

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos thumbnail

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos

Grit·2 months ago

AI Will Break Capitalism by Decoupling Labor from Capital, Rendering Price Signals Meaningless

Capitalism works because capital and labor are coupled; people earn money via labor, creating scarcity that makes price a meaningful signal of value. AI breaks this by creating value without labor. This decoupling could make traditional economic models obsolete and force society to redefine how value is assigned.

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos thumbnail

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos

Grit·2 months ago

Implicit User Actions Like Query Reformulation Are Better AI Quality Signals Than Explicit Feedback

Instead of relying on explicit feedback like upvotes, Arena measures AI quality through implicit user behavior. Actions like reformulating a prompt multiple times, downloading a generated file, or merging a code suggestion are powerful, organic signals that a user is (or isn't) getting their job done.

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos thumbnail

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos

Grit·2 months ago

US Lags in Open Source AI Because the "Give it Away" Model Clashes with American Corporate Economics

The US is behind in open-source AI development because of a fundamental business model problem. American companies struggle to justify spending billions on training a frontier model only to release it for free. Chinese companies can pursue this strategy due to different corporate dynamics and state influence.

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos thumbnail

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos

Grit·2 months ago

AI's Improvement Rate Is Accelerating, With Generational Leaps Growing Larger, Not Smaller

Contrary to expectations of diminishing returns, the pace of AI development is accelerating. Arena's data shows that performance gaps between new model generations (e.g., GPT image 2 vs 1.5) are sometimes the largest in history, suggesting progress is speeding up rather than plateauing.

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos thumbnail

The AI Race Has a Leaderboard | Arena CEO Anastasios Angelopoulos

Grit·2 months ago