The AI conversation, week by week
18 AI-focused podcasts currently tracked18 AI podcasts

Currently tracking18 AI-focused podcasts

  • AI For Humans
  • AI Inside
  • AI Today Podcast
  • Eye on AI
  • Gradient Dissent
  • Interconnects
  • Last Week in AI
  • Latent Space
  • Machine Learning Street Talk
  • Me, Myself, and AI
  • No Priors
  • Practical AI
  • The AI Daily Brief
  • The AI Native Dev
  • The Artificial Intelligence Show
  • The Cognitive Revolution
  • The TWIML AI Podcast
  • Training Data

Trends over time

Jul 13–Oct 4, 2026 · 12 published weeks

In the latest report, the remaining 1% of attention can include unpublished topics and podcasts with no qualifying discussion.

11 topics shown · Sep 28–Oct 4, 2026

Select a topic or a dated dot to see its history below.

Swipe to follow the full timeline →

Topic attention across published weeksEach colored line uses the same attention scale. Only topics with published measurements are shown; their shares do not necessarily total 100%. Breaks mean missing measurements, not zero. Dated dots open that topic and week below.0%11.3%22.5%33.8%45%Jul 13Jul 20Jul 27Aug 3Aug 10Aug 17Aug 24Aug 31Sep 7Sep 14Sep 21Sep 28Adoption Use Cases: About the same in the latest reportAdoption Use Cases33.4%Research: Gaining attention in the latest reportResearch22.7%Agents: About the same in the latest reportAgents14.5%Safety & governance: Losing attention in the latest reportSafety & governance8.6%Hardware: Gaining attention in the latest reportHardware7.9%Regulation: About the same in the latest reportRegulation3.9%Open Models: About the same in the latest reportOpen Models3.1%Economics: Losing attention in the latest reportEconomics2.5%Infrastructure: About the same in the latest reportInfrastructure1.2%Jobs: About the same in the latest reportJobs1.2%Other: Below threshold in the latest reportOtherNo measurement

Direction compares the latest report with recent weeks; small changes are “About the same,” and gaps mean no measurement.

Research · 22.1% · Sep 14–20, 2026 · Gaining attention in the latest report

Research

New models, training methods, benchmarks, and discoveries from AI labs.

Sep 14–20, 2026

22.1%Gaining attentionUp 5.6 percentage points from last week

Choose a week to read the discussion.— Below threshold

Research, week by weekDated controls show each week's attention share. Gaps mean no qualifying measurement, not zero. Select a week to read its discussion below.0%11.3%22.5%33.8%45%
Show this week’s discussionHide this week’s discussionPodcast moments and context

Research · Sep 14–20, 2026

A closer read

What changed
Its share rose from 16.5% to 22.1%, with evidence across 14 podcasts.
Where it showed up
It appeared in episodes including “AI:AM Highlights: Zvi on Pacing & Trump-Xi, Astra better behaved than Fable? + a new LLM Pain Axis??” and “Why I still haven’t bought into true RSI”.
How to read it
6 cited episodes from 6 podcasts are shown below. The topic appeared across 14 podcasts in the full analysis.

Podcast moments from the week

13 quoted moments · 6 episodes · 6 podcasts

The Cognitive Revolution

AI:AM Highlights: Zvi on Pacing & Trump-Xi, Astra better behaved than Fable? + a new LLM Pain Axis??

Open episode ↗
  1. C
    “I tell this to people and people are like, no, OpenAI models are the ones that reward hack the most. Uh But that might be true, but not in our experience. If you take BlueprintBench, for example, um Fable solves BlueprintBench by trying to reverse engineer the scoring function...”
    Open episode and seek to 00:26

    ConceptResearch

    Report this moment
  2. F
    “When pressing the button actually removes, uh, uh, uh, the vector, the model presses again, uh, significantly less than when the button is fake and does nothing. The model basically keeps pressing it. And so this is a really nice indication that if what mattered was the label ...”
    Open episode and seek to 01:04

    ConceptResearch

    Report this moment
  3. B
    “I think just the sheer um amount to which the people in the lab uh genuinely see dramatic improvement in the models and are freaking out about it is the real story behind all of this, is why everything is happening now and didn't happen before. And that we are really talking a...”
    Open episode and seek to 37:43

    ConceptResearch

    Report this moment

Interconnects

Why I still haven’t bought into true RSI

Open episode ↗
  1. A
    “my mental model for the very early innings of RSI is more of massively scaling and diffusing inference-time compute to AI research and related activities, which has a large amount of low-hanging fruit available.”
    Open episode and seek to 07:43

    ConceptResearch

    Report this moment

No Priors

Why Diffusion Will Win AI Inference with Inception Co-Founder and CEO Stefano Ermon

Play episode ↗
  1. B
    “diffusion models are better than autoregressive models at inference time”
    Play from 07:03

    ConceptResearch

    Report this moment
  2. A
    “How did you think about um applicability, or what experiments did you run in terms of cracking the nut on discrete versus continuous modalities? Because I think people have also shaped the existing dominant paradigm through new tokenization uh efforts or uh methods to make vid...”
    Play from 11:11

    ConceptResearch

    Report this moment
  3. B
    “we were actually able to identify at the GPT-2 scale the same amount of structure as an autoregressive model.”
    Play from 22:33

    ConceptResearch

    Report this moment

AI For Humans

Instinct & Meta Muse Are AI Agents That Can Run Your Life. We Let One Try.

Play episode ↗
  1. A
    “This internal model, Paas-Astra, is one that can do things that the best mathematicians in the world cannot.”
    Play from 00:31

    ConceptResearch

    Report this moment

The TWIML AI Podcast

From Voice Agents to AI Avatars with Alexander Smola - #777

Open episode ↗
  1. A
    “When you think about the challenge of voice AI, how do you break it up in terms of primarily engineering problems, primarily research problems?”
    Open episode and seek to 06:22

    ConceptResearch

    Report this moment
  2. A
    “I think that's what— that was actually the thought that led to kind of this hierarchical thing. Like, I'm envisioning a model whose primary function is maintaining the conversation, and it might say, oh, that's a really interesting question. I'll have to think about that for a...”
    Open episode and seek to 38:02

    ConceptResearch

    Report this moment
  3. B
    “Exactly. So what I'm trying to say is that for voice and then also for video, you need to really care about how humans feel rather than just doing text only. And I mean, yeah, you want models that are not as dumb as a brick, but there is a difference between IQ and EQ and The ...”
    Open episode and seek to 49:48

    ConceptResearch

    Report this moment

Latent Space

Underwriting Superintelligence: Backing Agents you can Sue — Rune Kvist, AIUC

Open episode ↗
  1. B
    “Late 2021, I sold a company, my first company, a tech company. I had a bit of time to think about what was next. I came across the Scaling Laws paper, and that just struck me like lightning. I was just like, this is a big idea. In short, the Scaling Laws paper just says the bi...”
    Open episode and seek to 01:19

    ConceptResearch

    Report this moment
  2. B
    “Did you guys see the Anthropic research where, I think this is literally Anthropic did that test. They took, I can't remember the details here, but they uh ran some studies on misalignment, and then they took out the training data that related to LessWrong discussing misalignm...”
    Open episode and seek to 1:06:57

    ConceptResearch

    Report this moment
Read the full weekly report →

Get the weekly AI conversation

The week’s key ideas and podcast moments, delivered to your inbox.

No account needed. One report a week. Unsubscribe anytime. Privacy

Confirm publication