Peru Beans, AI Doom, and Code Review Burden

TJ Miller and Chris Gmyr get into the week's AI doom and gloom, from the Hugging Face incident to calls for regulation, before turning to a problem hitting their teams right now. They compare notes on automated code review, adversarial reviewers that never stop finding issues, and the evals it takes to tune them.

Links

• Ask a question or suggest a topic
• Podcast: Site | Twitter/X | Bluesky
• Chris: Site | Twitter/X | Bluesky
• TJ: Site | Twitter/X | Bluesky

References Mentioned

• ESP32 - the Wi-Fi controller chip TJ is wiring his shelf LED strips to
• Sonari - TJ's transcription app, with beta invites going out from the mailing list next week
• Halcyon - TJ's notes app, the other project he spun back up this week
• Mt. Comfort Coffee Peru Medium Roast - the 2.5 pound Costco bag TJ and his wife are loving
• Bizzy Espresso Blend Cold Brew - Chris's replacement espresso cold brew
• Anthropic researcher resigns with a warning - Jacob Coxon's resignation that kicked off the week's AI safety news
• OpenAI and Hugging Face incident - the containment failure both hosts point to
• Cal Newport: Anthropic Just Threatened to Kill Billions of People - Cal's recap and take on the resignation
• The Diary of a CEO - the podcast where TJ heard a former OpenAI forecaster describe the race to self-improving AI
• GitHub Copilot - the org-wide reviewer Chris syncs Claude rules into
• Amazon Bedrock - what Chris's AI reviewer workflow runs on
• CodeRabbit - the AI review tool TJ wants to study for ideas