Every episode, newest first.

1. Oct 10Anthropic’s agents go off-script, Cloudflare opens Clef-omni.4:09
2. Oct 9OpenAI’s Ultrafast Sol, Anthropic’s free open-source scans, Xiaomi’s RL recipe.5:29
3. Oct 8Claude Haiku 5.5, GPT-6 in ChatGPT, OpenAI’s math withdrawals, a Category 5 influence op.7:01
4. Oct 7OpenAI’s math dump, Anthropic’s cyber tiers, SynthID for everyone.4:53
5. Oct 6Google maps outbreaks with Earth AI, Mistral previews a trillion-parameter model.3:27
6. Oct 5OpenAI watermarks text for the EU, vLLM restarts faster.3:32
7. Oct 4Gemini’s free tier narrows, ThinkingBox grades agents by outcomes.2:52
8. Oct 3Anthropic’s engineer academy, OpenAI’s GPT-6 guide, Aleph Alpha’s Kolibri.3:06
9. Oct 2Claude Code mods, Claude-shaped science, Cloudflare’s Clef.2:57
10. Oct 1Gemini 4 Argon, a distillation campaign, watermarked proteins.3:04
11. Sep 30GPT-6.1 Sol, GLM-5.3’s cyber test, OpenAI’s dots.3:02
12. Sep 29Sonnet 5.5 ships, GPT-6.1 Astra is held back.3:04
13. Sep 28Nine loops, agents in bounds, Meta goes enterprise.3:12
14. Sep 27Closing a gap takes more than an alert.2:58
15. Sep 26Inside Panther Lake. Inside an agent’s memory.2:56
16. Sep 25From a moving image to a useful action.2:51
17. Sep 24Agents need more than a model.2:57
18. Sep 23Predict the world. Choose what to compute.3:03
19. Sep 22Make progress visible.2:50
20. Sep 21Look beyond the model.3:03
21. Sep 20Put AI claims to the test.3:01
22. Sep 19Shared access. More capacity.2:46
23. Sep 18Better coordination. Better use of memory.2:50
24. Sep 17Seven things worth your attention.2:59
25. Sep 16Make AI easier to put to work.2:55
26. Sep 15Better systems, not just better models.2:57
27. Sep 14The constraints beyond the model.3:06
28. Sep 13Less hardware. Better questions.2:58
29. Sep 12Take the work into the real world.2:56
30. Sep 11Build leaner. Reach real users.2:48

Use a podcast app? Follow the podcast feed instead.