UpNext AI podcast

Legora’s Legal AI, Weight-Space Programming, and WebHarbor’s Browser-Agent Tests

11/3/2026
0:00
7:52
Rewind 15 seconds
Fast Forward 15 seconds

Legora CTO Jacob Lauritzen discusses confidential legal AI evaluations and Tabular Review; Rikiya Takehi, Jayden Teoh, Alex Zhang, and Omar Khattab report early weight-space programming experiments; and WebHarbor offers repeatable environments for browser-agent tests. Also covered: Anthropic’s Claude Code game-recreation rule, Simon Willison’s LLM release, OpenAI’s bug bounty, Marvell’s Lumentum deal, and Oracle’s AI infrastructure financing.

  • Legora’s approach to evaluating legal agents without access to customer documents, and Lauritzen’s account of changing legal billing.
  • Early experiments in reading and editing neural-network weights directly, including a reported repair to Qwen’s reasoning loops.
  • Claude Code’s instruction against recreating specific existing games.
  • WebHarbor’s website environments for training and evaluating graphical-interface agents.
  • Simon Willison’s LLM tool release; OpenAI’s higher bug-bounty payouts; Marvell’s planned Lumentum acquisition; and Oracle’s debt and stock offerings.

Sources:
- Legora interview: https://latent.space/p/building-ai-for-lawyers-jacob-lauritzen-cto-of-legora/
- Weight-space programming: https://rikiyat.github.io/blog/weight-space-programming/
- Claude Code game rule: https://arstechnica.com/ai/2026/11/claudes-new-rules-against-game-cloning-wont-stop-the-ai-copying-boom/
- WebHarbor: https://arxiv.org/abs/2611.03634v1
- LLM release: https://simonwillison.net/2026/Nov/2/llm/
- OpenAI bug bounty: https://the-decoder.com/bug-hunters-can-now-earn-up-to-172000-from-openai-as-company-ramps-up-security/
- Marvell and Lumentum: https://www.theinformation.com/briefings/marvell-buy-lumentum-115-billion-all-stock-deal
- Oracle offerings: https://www.theinformation.com/briefings/oracle-launches-20-billion-stock-bond-sales

More episodes from "UpNext AI"