Sign in
Interconnects podcast artwork

Interconnects

By Nathan LambertTechnology, ScienceEnglish

New episode weekly, latest 1 day ago, since Dec 2023

Save for your brand

Host this show? Claim it free to set your rates and verify your numbers. Claim this show

episodes in the feed
168
From the feed Oct 30
on air, since Dec 2023
3 years
From the feed Oct 30

About

Audio essays about the latest developments in AI and interviews with leading scientists in the field. Breaking the hype, understanding what's under the hood, and telling stories.
www.interconnects.ai

From the show's RSS feed.

Episodes 168 episodes in the feed

  1. Artificial Analysis v5 and the benchmarker's dilemma

    10 min

  2. Podcast: Kimi K3.1 Pro, the new cyber reality, and Muse musings

    46 min

  3. Meta shakes up the frontiers, open and closed

    11 min

  4. The Cyber Risk Discourse is Broken

    9 min

  5. Debating RSI, the US-China Gap, and Jaggedness with JS Denain of Epoch AI

    1 h 5 min

  6. The current balance of power in open models

    17 min

  7. Why I still haven’t bought into true RSI

    9 min

  8. One resignation turned the embers of AI fear into a wildfire

    10 min

  9. When will average people feel AI’s impact?

    8 min

  10. I wrote an AI textbook — how long until AI can do it better?

    12 min

Show 40 older episodes

  1. Open models recap: more on Kimi K3, Qwen 3.8, Xi's WAIC speech, distillation, the open-closed gap, and what's next

    49 min

  2. Kimi K3: The open-weights escalation

    20 min

  3. GLM-5.2 is the step change for open agents

    9 min

  4. Banning Open Source AI Would Be A Mistake

    6 min

  5. State of the blog, mid-2026

    6 min

  6. Frontier post-training recipe review with Finbarr Timbers

    56 min

  7. Claude Fable 5 and new AI safety fables

    12 min

  8. Farewell Ai2

    15 min

  9. Open and closed models are on different exponentials

    7 min

  10. Some ideas for what comes next, May 2026

    9 min

  11. Notes from inside China's AI labs

    16 min

  12. The distillation panic

    8 min

  13. My bets on open models, mid-2026

    6 min

  14. The inevitable need for an open model consortium

    5 min

  15. Claude Mythos and misguided open-weight fearmongering

    8 min

  16. Gemma 4 and what makes an open model succeed

    8 min

  17. Lossy self-improvement

    13 min

  18. GPT 5.4 is a big step for Codex

    6 min

  19. What comes next with open models

    18 min

  20. Dean Ball on open models and government control

    35 min

  21. Olmo Hybrid and future LLM architectures

    11 min

  22. How much does distillation really matter for Chinese LLMs?

    11 min

  23. Opus 4.6, Codex 5.3, and the post-benchmark era

    8 min

  24. Why Nvidia builds open models with Bryan Catanzaro

    1 h 7 min

  25. Thoughts on the job market in the age of LLMs

    10 min

  26. Arcee AI goes all-in on open models built in the U.S.

    1 h 12 min

  27. Get Good at Agents

    5 min

  28. Use multiple models

    7 min

  29. Claude Code Hits Different

    4 min

  30. Open models: Hot or Not with Nathan Lambert & Florian Brand

    37 min

  31. New Talk: Building Olmo 3 Think

    1 h 2 min

  32. Olmo 3: America’s truly open reasoning models

    10 min

  33. Why AI writing is mid

    8 min

  34. Interview: Ant Group's open model ambitions

    1 h 17 min

  35. 5 Thoughts on Kimi K2 Thinking

    7 min

  36. Burning out

    10 min

  37. How to scale RL

    13 min

  38. The State of Open Models

    47 min

  39. Thoughts on The Curve

    11 min

  40. ChatGPT: The Agentic App

    9 min

Sponsors

No sponsor reads found yet.

We read every episode's show notes for sponsor reads, promo codes and offers. None turned up in the episodes we have.

Audience

See who listens.

Signed in, you'll see what this show's listeners also love, when the show has a profile: brands, shows and interests. Taste data from Qloo.

Sign in to see it

Rates

No rate card yet.

Rates appear once the show claims this page and publishes them. Host this show? Claiming is free.