Skip to content

Collections

Best Chain of Thought Episodes on Open Source AI

Open weights, open ecosystems, and what they cost in production.

Compare the costs and deployment choices behind open models with the ecosystem and trust questions around them. Fergal Reid describes Intercom’s move to a fine-tuned Qwen model; AMD’s guests discuss their open ecosystem, and Rizel Scarlett introduces Goose. The collection also includes the show’s early conversations about DeepSeek and enterprise trust.

  • 5 episodes
  • 6 guests
  • 4 hours of listening
  • 1 jump-to moment
  1. 1 EP 11 33 min Transcript How DeepSeek Changed the AI Race Overnight Atindriyo Sanyal, Galileo The week DeepSeek’s R1 reset what an open model was allowed to cost. Recorded while it was still landing, which is why it holds up as a record of the argument rather than a retrospective on it.
  2. 2 EP 49 54 min Transcript How Intercom Cut $250K/Month by Ditching GPT for Open Models | Fergal Reid Fergal Reid, Intercom Intercom was spending $250K a month on one summarization task and replaced it with a fine-tuned 14B Qwen model. Fergal Reid walks through the whole decision, the most concrete case on the show of running open weights in production. Jump to the moment · 14:42 · Testing Qwen beyond benchmarks →
  3. 3 EP 28 49 min Transcript AMD's Challenge to NVIDIA: The Open Ecosystem Bet | Anush Elangovan & Sharon Zhou Anush Elangovan & Sharon Zhou AMD’s Anush Elangovan and Sharon Zhou on betting a silicon roadmap against CUDA lock-in, and what it takes to make community contribution a working path instead of a press release.
  4. 4 EP 10 34 min Transcript AI, Open Source & Developer Safety | Block’s Rizel Scarlett Rizel Scarlett, Block Block’s Rizel Scarlett on Goose, the on-machine open source agent her team shipped, and how open source development and developer safety pull on each other.
  5. 5 EP 1 49 min Transcript The State of AI: Open-Source Models & Enterprise Trust | May Habib May Habib, Writer The show’s first episode, with Writer CEO May Habib on deploying generative AI at scale. It framed open versus proprietary as a question about enterprise trust rather than benchmarks.

More collections

All collections →

New here? Start with five → Browse by topic →