Skip to content

Listening guide

Podcasts like Machine Learning Street Talk

Tim Scarfe talks with researchers and builders about machine learning, intelligence and its limits. Here are six MLST episodes, each with a Chain of Thought conversation to play next.

  • 6 episode pairings

I host Chain of Thought. We have no affiliation with Machine Learning Street Talk: no partnership, no arrangement, nothing either of us gets from this page.

What Machine Learning Street Talk is

Machine Learning Street Talk began in April 2020. Tim Scarfe runs the show, and Keith Duggar appears regularly. Its Apple Podcasts description covers current affairs in AI, cognitive science, neuroscience and philosophy of mind; recent episodes are technical conversations with researchers and builders.

Recent subjects include voice agents, formal proof, agent evaluation, tabular models, robotics and speech recognition. The show links to ReScript transcripts from some recent episode notes.

At a glance

Machine Learning Street Talk and Chain of Thought
DetailMachine Learning Street TalkChain of Thought
HostsTim Scarfe, with regular appearances by Keith DuggarConor Bronsdon
First episodeApril 2020November 2024
FormatTechnical conversations with researchers and buildersInterviews with engineers, founders and operators shipping AI in production
TranscriptsReScript transcript links in some recent episode notesA full transcript on every episode page
Where to listenMLST websiteEpisode archive and podcast apps

Details for Machine Learning Street Talk come from the sources below, checked 14 October 2026.

Where Chain of Thought fits

Chain of Thought follows AI infrastructure from the point of view of people shipping it. Conor Bronsdon asks about model, data and product decisions in specific systems. MLST gives researchers and builders room to examine the technical ideas and limits underneath those decisions.

Which one to start with

Start with Machine Learning Street Talk to spend time on the mechanics of a model, method or research result.

Start with Chain of Thought to hear what happens when a team turns an AI capability into a service people use.

You do not have to pick one. The pairings below are where the two shows meet.

  1. Then play

    Chain of ThoughtEP 7055 minTranscript 11 Cameras, Dozens of Mics: The AI That Reads the Room | Tormod Ree, Neat Tormod Ree, Neat Tormod Ree of Neat follows the audio question into a meeting room, where edge models combine microphone and camera signals to decide how to respond.

  2. Then play

    Chain of ThoughtEP 4055 minTranscript After Code Gen: What Graphite Is Building for the Post-AI Dev Stack | Greg Foster Greg Foster, Graphite Greg Foster of Graphite follows the verification question into everyday software work, where AI-generated pull requests have moved pressure onto review.

  3. Then play

    Chain of ThoughtEP 5743 minTranscript Every AI Agent Has an Evaluation Gap | Alex Ratner, Snorkel AI Alex Ratner, Snorkel AI Alex Ratner of Snorkel AI takes up the measurement problem, describing the gap between agent benchmark results and the work enterprises need agents to do.

  4. Then play

    Chain of ThoughtEP 4353 minTranscript Beyond Transformers: How Liquid AI Is Rethinking LLM Architecture | Maxime Labonne Maxime Labonne, Liquid AI Maxime Labonne of Liquid AI follows the model-design question into language models, explaining why his team changed the architecture to meet speed and device constraints.

  5. Then play

    Chain of ThoughtEP 4751 minTranscript Gemini 3 & Robot Dogs: Inside Google DeepMind's AI Experiments | Paige Bailey Paige Bailey, Google DeepMind Paige Bailey of Google DeepMind follows physical AI into another developer ecosystem, discussing robotics work alongside the Gemini and Gemma model choices.

  6. Then play

    Chain of ThoughtEP 2041 minTranscript Breaking the Language Barrier: Smartling's AI Translation Pipeline | Olga Beregovaya Olga Beregovaya, Smartling Olga Beregovaya of Smartling follows the language question into translation, where English-centered models and domain-specific errors still need attention.

Sources

Facts about Machine Learning Street Talk on this page were checked against these pages on 14 October 2026.

More listening guides

New to our show? Start with five →Machine Learning Street Talk ↗