Latest podcast knowledge.

New episodes processed by Podwise, ready to search, skim and continue in the full app.

Just summarized

Read what each episode covers before you press play.

The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)Nov 10

Canva Is Tackling AI Design’s Most Frustrating Problem with Stefano Corazza- #784

AI summary Canva’s AI research team is transforming design workflows by shifting from single-shot image generation to iterative, human-in-the-loop processes. Stefano Corazza, head of AI research, details the development of the "Craft" model, which utilizes a two-stage…

Canva
Canva Craft
Canva Brush

58 minEnglishSummary · Transcript

Search EngineNov 11

What was Jensen Huang thinking?

AI summary Nvidia CEO Jensen Huang faces a major scandal involving an alleged affair with a woman linked to Chinese intelligence, raising critical questions about the security of U.S. AI chip export controls. Reporters and industry experts, including Leo Schwartz of The Information and…

Nvidia
Jensen Huang
Export controls

44 minEnglishSummary · Transcript

Letters from an American Nov 11

Hitting the Ground Running

AI summary Washington’s political climate has undergone a rapid transformation following the midterm elections, with Democrats preparing an aggressive agenda focused on affordability and government accountability. Key initiatives include reversing Medicaid work requirements and launching…

Project 2025
War Powers Act
Medicaid

16 minEnglishSummary · Transcript

松原タニシの恐味津々Nov 12

#293 全国を結ぶUMA・妖怪の足跡？地方の不思議スポットをご紹介！そしてラストは衝撃の…【対談相手：吉田猛々（ナナフシギ）】

AI summary YouTube チャンネル「ナナフシギ」の吉田猛々が、怪談やオカルトへの傾倒と、全国各地の伝承を巡るフィールドワークの経緯を語る。岩手県の丹内山神社や秩父の即道神社、新潟の西生寺など、現地調査で得た知見を基に、妖怪や UMA の正体を考察する。特に、ビッグフットがテレポーテーション能力を持つという仮説や、各地の伝承が同一の存在を指しているという独自の「UMA 妖怪説」を展開。お笑い芸人としての活動と並行し、怪談本『独り言』の出版や毎日更新の動画配信を行うなど、オカルト探究者としての多忙な日常も明かされる。

UMA
アラハバキ
マロウド神

43 minJapaneseSummary · Transcript

How I AINov 11

Claude Code + Obsidian: My family’s AI brain | Scott Farrar

AI summary Managing a household with AI requires grounding large language models in a persistent, transparent "source of truth" rather than relying on ephemeral chat sessions. By utilizing local Markdown files as a personal knowledge base, homeowners can provide AI agents with specific…

Claude Code
Obsidian
Markdown

00:34:14EnglishSummary · Transcript

BestBlogsNov 11

EP234 · A2UI 生成界面、Codex 持续排障、Muse 优化 TPU 推理

AI summary AI 智能体正从简单的对话交互转向更高效的工程化落地。Google A2UI 1.0 通过原生交互卡片将模型决策与应用界面结合，减少了用户反复确认的沟通成本。OpenAI Codex 的持续工程实践证明，通过将复杂任务拆解为 “主智能体协调 + 子智能体执行” 的模式，能有效实现长期维护。同时，针对 TPU 的推理优化及自动化验收机制，进一步解决了长任务运行中的性能瓶颈与可靠性问题。这些实践强调了在生成式 AI 时代，明确输出条件、构建可验证的循环流程以及优化底层执行路径，是提升智能体生产力的关键，也标志着 AI…