跳至主要內容
黑五

鎖定六折,往後每次續費都是六折黑五 · 續費也六折黑五 · 六折

12 月 6 日截止領取六折 →

最新播客 知識。

Podwise 處理的新節目,隨時可搜尋、瀏覽,並在完整應用中繼續。

剛剛摘要

在按下播放前,先讀讀每集講了什麼。

Learning English Conversations

The English We Speak: In for a treat

AI 摘要 The idiomatic expression "in for a treat" describes a situation where someone is about to experience something highly enjoyable or special. While often associated with food—such as a freshly baked cake—a treat encompasses any infrequent, pleasurable activity like a professional…

  • In for a treat
  • Luddites
  • ChatGPT

3 minEnglish摘要 · 逐字稿

AI Engineer

2025 in LLMs so far, illustrated by Pelicans on Bicycles — Simon Willison

AI 摘要 The rapid evolution of Large Language Models over the past six months has introduced over 30 significant releases, complicating evaluation through traditional benchmarks. A novel, informal approach—prompting models to generate SVG code of a pelican riding a bicycle—reveals…

  • LLM (Large Language Model)
  • SVG (Scalable Vector Graphics)
  • Hugging Face

19 minEnglish摘要 · 逐字稿

AI Engineer

Software Engineering Is Becoming Plan and Review — Louis Knight-Webb, Vibe Kanban

AI 摘要 Software engineering is shifting from active code writing to a role defined by planning and reviewing AI-generated outputs. This transition necessitates a strategic approach: investing time in comprehensive upfront planning minimizes the need for iterative, time-consuming…

  • AI Tinkerers
  • SweBench
  • GitHub Copilot

20 minEnglish摘要 · 逐字稿

AI Engineer

How Transformers Finally Ate Vision – Isaac Robinson, Roboflow

AI 摘要 Vision Transformers (ViTs) have superseded convolutional neural networks in computer vision despite their lack of inherent inductive bias. This shift is driven by massive ViT-specific pre-training techniques, such as masked autoencoders, which allow models to learn necessary…

  • Transformers
  • Convolutional Neural Networks (CNNs)
  • Vision Transformer (ViT)

17 minEnglish摘要 · 逐字稿

AI Engineer

The AI Bill Is Too Damn High: Tokenomics 101 — Eugene Cheah, Featherless AI

AI 摘要 Optimizing AI token economics requires a systematic approach centered on measurement, right-sizing, and process refinement. Organizations often face runaway costs by applying high-performance models to every task, failing to distinguish between agentic workloads—which are…

  • Tokenomics
  • RWKV
  • MOE (Mixture of Experts)

55 minEnglish摘要 · 逐字稿

科技这碗饭

谷歌AI风云15年:诸神黄昏——发明未来的人为何错过未来?

AI 摘要 谷歌在过去十五年里从 AI 领域的先行者逐渐陷入被动,其核心矛盾在于研究导向的文化与商业化需求之间的冲突。谷歌大脑与 DeepMind 作为两大顶尖研究团队,虽在 AlphaGo、Transformer 及 AlphaFold 等领域取得突破,但长期缺乏产品落地,导致在面对 OpenAI 的 ChatGPT 冲击时反应迟缓。内部组织架构的频繁调整、算力分配的内耗以及对 AI 安全与伦理的过度谨慎,进一步削弱了谷歌的先发优势。随着 Gemini 的推出及团队的强行整合,谷歌被迫从 “研究优先” 转向 “产品优先”,这一转型不仅导致了 Jeff Dean…

  • AlphaGo
  • Transformer
  • AlphaFold

02:24:18Chinese摘要 · 逐字稿

更多新節目

  1. You’re About to See the Real AI Winners Stand Up

    The Compound and Friends

    摘要1 h 28 min
  2. 摘要00:27:30
  3. 摘要5 h 11 min
  4. 摘要27 min
  5. 摘要1 h 15 min
  6. 摘要
  7. 摘要1 h 15 min
  8. 摘要5 min
  9. 摘要9 min
  10. 摘要36 min
  11. 摘要55 min
  12. 摘要12 min
  13. 摘要8 min
  14. Tenstorrent COO Keith Witek on AI-powered autonomy

    New York Stock Exchange

    摘要5 min
  15. 摘要7 min
  16. 摘要7 min
  17. 摘要00:55:56
  18. 摘要00:32:01
  19. No Priors Ep. 31 | With Cerebras CEO Andrew Feldman

    No Priors: AI, Machine Learning, Tech, & Startups

    摘要30 min
  20. 摘要29 min
  21. Cerebras CEO Andrew Feldman Talks Earnings

    Bloomberg Talks

    摘要8 min
  22. Cerebras CEO Andrew Feldman Talks IPO

    Bloomberg Talks

    摘要9 min
  23. Oral History of Jim Keller

    Computer History Museum

    摘要1 h 45 min
  24. 摘要00:42:27
  25. Selects: The Soul Train Episode

    Stuff You Should Know

    摘要47 min
Podwise

每一集新節目 都有摘要。

在 Podwise 追蹤你的節目,在決定收聽前先讀每集新節目。