Nvidia Has Done the Big Bad

Lawful Masses with Leonard French
Lawful Masses with Leonard FrenchJun 2, 2026

Why It Matters

The ruling narrows the "Cox" safe‑harbor for AI developers, signaling that providing piracy‑enabling tools can trigger direct and contributory infringement liability, which could reshape data‑sourcing practices across the industry.

Key Takeaways

  • Judge Tigar denies Nvidia’s motion to strike BitTorrent references.
  • Direct and contributory infringement claims against Nvidia survive the dismissal.
  • Nvidia’s scripts automate pirated book downloads, deemed infringement tools.
  • Supreme Court’s Cox v. Sony decision shapes AI copyright liability.
  • Ongoing discovery could set precedent for AI training data acquisition.

Summary

The video dissects a landmark copyright case in the Northern District of California where author‑plaintiffs sued Nvidia for training large language models on a corpus of pirated books. Judge John Tigar rejected Nvidia’s bid to dismiss the core allegations, allowing both direct infringement and contributory infringement claims to proceed while only striking the vicarious claim.

Key facts include Nvidia’s use of the open‑source "Books 3" dataset—derived from the BitTorrent tracker Bibliotic—and the company’s provision of scripts that automatically download and preprocess the entire collection for corporate customers. The court focused on those scripts, finding they serve no legitimate purpose beyond accelerating illegal acquisition, and refused to treat BitTorrent as a neutral tool. The ruling leans heavily on the Supreme Court’s recent unanimous Cox v. Sony decision, which narrowed the safe‑harbor for services that facilitate infringement.

Leonard French highlights the "dolphin painting" analogy used by the judge, emphasizing that removing BitTorrent references would be akin to striking paintbrush mentions from a painting case. He also references parallel litigation—Anthropic’s $1.5 billion settlement and Meta’s Kadri case—illustrating how the industry is testing the boundaries of fair‑use defenses for AI training data.

The decision opens discovery into Nvidia’s acquisition and distribution mechanisms, setting a potential precedent for how courts treat AI infrastructure providers. If plaintiffs succeed, AI firms may face heightened liability for the tools they supply, reshaping compliance strategies and possibly prompting broader licensing negotiations for training corpora.

Original Description

A recent federal court ruling has brought a fascinating twist to a copyright infringement lawsuit, challenging a multi-trillion-dollar tech company's attempt to remove all mentions of bittorrent. The judge's colorful analogy about 'dolphin painting' highlights the core issue in this file sharing case. This decision could reshape how p2p protocols are viewed in legal battles.
Special thanks to my wife Kayleigh for speed painting the dolphins on canvas.
Want to do something about it? postcardstocongress.org — code Just5Dollars sends postcards to all three of your federal reps for $5.
Support the channel: patreon.com/ljfrench
00:00 - Nvidia has done the big bad
01:17 - WHAT JUST HAPPENED
02:27 - THE COX CURVEBALL
04:10 - THE PIPELINE
06:18 - NVIDIA'S BET
09:05 - THE PAINTBRUSH
10:40 - WHERE THIS LEAVES THE INDUSTRY
13:41 - CLOSING
14:27 - Please Support Lawful Masses

Comments

Want to join the conversation?

Loading comments...