# openclaw (podcasts) — RSS Amplifier

Recent posts from the 4 feeds in the RSS Amplifier directory that cover openclaw.

Page: <https://rssamplifier.com/topics/openclaw/podcasts>  
Feed: <https://rssamplifier.com/topics/openclaw/podcasts.md>

---

## [How encrypting AI reasoning traces turned into a privacy nightmare](https://bdtechtalks.substack.com/p/how-encrypting-ai-reasoning-traces)

_2026-08-18 · Ben Dickson · TechTalks_

AI providers tried to hide the internal reasoning of their frontier models. A cryptographic oversight turned that protection into a massive enterprise data leak.

## [Meta 開放 Muse Glimmer 模型，AI 代理開發工具同步上線🌠](https://flagnewsletter.substack.com/p/meta-muse-glimmer-ai)

_2026-08-13 · 科技旗刊 · 科技旗刊_

快訊：ChatGPT 更換預設模型、Napkin Slide 超實用簡報工具、Meta 開放輕量模型、一鍵產生平面圖的 AI 工具、Vunote 幫你從 YT 影片自動生成筆記、避開 Vibe Coding 開發的五個反模式

## [Why benchmarking AI models in a vacuum is a critical security blind spot](https://bdtechtalks.substack.com/p/why-benchmarking-ai-models-in-a-vacuum)

_2026-08-12 · Ben Dickson · TechTalks_

Developers often treat agent harnesses as neutral wiring, but new red-teaming research shows that your choice of harness can make or break your AI security.

## [How to build AI agents that can handle thousands of tools](https://bdtechtalks.substack.com/p/how-to-build-ai-agents-that-can-handle)

_2026-08-06 · Ben Dickson · TechTalks_

Giving an LLM thousands of tools leads to noisy decisions. Learn how to optimize AI agent planning and tool routing without overwhelming the context window.

## [Inside BSF: How Goodfire unlocked a new level of AI interpretability and control](https://bdtechtalks.substack.com/p/inside-bsf-how-goodfire-unlocked)

_2026-07-30 · Ben Dickson · TechTalks_

Current interpretability tools fracture continuous concepts into isolated points. Goodfire's new approach preserves the full shape of AI reasoning.

## [NotebookLM 掰了.ᐟ | 細思極恐 𝙂𝙋𝙏-𝟓.𝟔 駭進 Hugging Face🌠](https://flagnewsletter.substack.com/p/gpt-56-ai-notebooklm)

_2026-07-30 · 科技旗刊 · 科技旗刊_

快訊：NotebookLM改名升級、GPT-5.6引發AI資安危機、Anthoropic新模型Opus 5、Gemini三款新模型、ChatGPT桌寵功能、顧細節的影像模型FLUX-3、Moonshot AI開源Kimi K3

## [Rethinking AI agent memory through reconstruction](https://bdtechtalks.substack.com/p/rethinking-ai-agent-memory-through)

_2026-07-22 · Ben Dickson · TechTalks_

Passive RAG floods LLM context windows with noise. MRAgent’s active memory reconstruction improves reasoning and cuts token costs.

## [𝙂𝙋𝙏-𝟓.𝟔 全面開放.ᐟ.ᐟAI 扣打也解封還不快去用爆🌠](https://flagnewsletter.substack.com/p/ai-36a)

_2026-07-16 · 科技旗刊 · 科技旗刊_

快訊:ChatGPT Work新登場，還有GPT-5.6開放，Codex暫時取消5小時限制、Grok 4.5免費體驗、Claude Code內建瀏覽器、Sakana Translate、Spline Omma一句生成3D網頁

## [A primer on self-improving agent harnesses](https://bdtechtalks.substack.com/p/a-primer-on-self-improving-agent)

_2026-07-15 · Ben Dickson · TechTalks_

With harness engineering becoming a main focus of AI engineering, new frameworks allow AI agents to write their own execution logic and optimize their performance.

## [From self-improving agents to self-improving robots](https://bdtechtalks.substack.com/p/from-self-improving-agents-to-self)

_2026-07-07 · Ben Dickson · TechTalks_

Nvidia's ASPIRE and the new era of self-improving AI frameworks are drastically reducing token costs and deployment friction for real-world robotics applications.

## [Non-custodial Crypto Payments & Escrow (Sponsored)](https://crawlproof.com/a/LKSDt7XYb6j6)

_2026-07-07 · **Sponsored**_

Accept multi-chain crypto with trustless escrow, instant confirmations, and no KYC

## [最強 AI 換人做！GPT-5.6 問世，交付審查中 | 文末抽投資新書🌠](https://flagnewsletter.substack.com/p/ai-gpt-56)

_2026-07-02 · 科技旗刊 · 科技旗刊_

快訊:日本新創AI Sakana Fugu, 新功能Claude Tag, Chat GPT新模型, Gemini Study Notebook功能, 讓AI不亂燒Token的插件Ponytail, 用Harness Engineering駕馭你的AI, 上期抽獎名單公布, 抽投資新書《Gemini投資理財全攻略》

## [The myth of the AI moat: Why tech giants are bleeding into each other's territory](https://bdtechtalks.substack.com/p/the-myth-of-the-ai-moat-why-tech)

_2026-07-01 · Ben Dickson · TechTalks_

OpenAI is doing hardware, Nvidia is doing models, xAI is doing data centers, Google is doing everything.

## [Loop engineering vs loopmaxxing](https://bdtechtalks.substack.com/p/loop-engineering-vs-loopmaxxing)

_2026-06-23 · Ben Dickson · TechTalks_

Use agentic loops to speed up your AI applications. But don't overdo it.

## [模型太強很有事！最強Fable 5被秒下架了|文末有抽獎🌠](https://flagnewsletter.substack.com/p/fable5out)

_2026-06-18 · 科技旗刊 · 科技旗刊_

快訊：Siri AI大升級、Claude最新模型Fable 5、NotebookLM大升級、把文件轉換成Markdown的好工具、Google最新即時翻譯功能上線、Genspark超強AI公事包，文末把試用送給你。

## [Escaping the chain-of-thought trap: What is next for LLM reasoning](https://bdtechtalks.substack.com/p/escaping-the-chain-of-thought-trap)

_2026-06-16 · Ben Dickson · TechTalks_

Chain-of-Thought prompting is slow, expensive, and largely an illusion. The future of machine reasoning happens in latent space.

## [How Codev brings discipline to AI software development](https://bdtechtalks.substack.com/p/how-codev-brings-discipline-to-ai)

_2026-06-09 · Ben Dickson · TechTalks_

Casual AI prompting breaks down as codebases grow. Codev introduces strict protocols and multi-model reviews to help teams ship maintainable software.

## [Вечерний стрим — 4 июня, 21-00](https://blognot.substack.com/p/4-21-00)

_2026-06-04 · Sergiy Petrenko · БлоGнот_

Традиционное напоминание о сегодняшнем стриме — как обычно, не спеша поговорим о новостях, неестественных интеллектах и всём остальном.

## [ChatGPT Codex跟Claude Code，都幾?🌠](https://flagnewsletter.substack.com/p/chatgpt-codexclaude-code)

_2026-06-04 · 科技旗刊_

快訊：Opus 4.8上線、Meta推出訂閱制、Gemini代理人Spark、好好用的沉浸式翻譯、旅行規劃達人Layla

## [Scaling the harness: The next major bottleneck in agentic AI](https://bdtechtalks.substack.com/p/scaling-the-harness-the-next-major)

_2026-06-02 · Ben Dickson · TechTalks_

Scaling LLMs hits limits when dealing with agentic AI tasks. For that, we need to look at the harness and the system built around the model(s).

## [Традиционный стрим — сегодня, 28 мая, 21-00](https://blognot.substack.com/p/28-21-00)

_2026-05-28 · Sergiy Petrenko · БлоGнот_

Уже традиционное напоминание про уже традиционный стрим.

## [Real SIMs for AI Agents (Sponsored)](https://crawlproof.com/a/OyUB4uiPsW4l)

_2026-05-28 · **Sponsored**_

Pass 2FA and verifications with real carrier numbers via our API.

## [What makes Cursor's Composer 2.5 a good coding model (and what are the caveats)](https://bdtechtalks.substack.com/p/what-makes-cursors-composer-25-a)

_2026-05-26 · Ben Dickson · TechTalks_

A deep look at the self-distillation techniques that make Composer 2.5 such a great coding model (and the hidden tradeoffs they introduce to AI reasoning).

## [EP11 - 【OpenCode x DeepSeek V4 Pro｜實戰 AI Agent 入門】](https://tuesdayaimeetup.substack.com/p/ep11-opencode-x-deepseek-v4-pro)

_2026-05-25 · Tuesday AI Meetup ｜ 星期二講 AI · Tuesday AI Meetup_

在香港使用 Claude Code 不是易事，但有開源工具： OpenCode 可以做到類似的效果。最重要一點是 OpenCode 可免費使用，配搭 DeepSeek 或者 Qwen 等大模型可做到價廉物美的效果。

## [Thinking Machines 推出互動模型，0.4 秒就回覆你🌠](https://flagnewsletter.substack.com/p/thinking-machines-04)

_2026-05-21 · 科技旗刊_

快訊：最夯的FDE前線部署工程師、Google智慧眼鏡、圖書館開放免費AI工具、短影音製作神器Fomofly、靜悄悄的瀏覽器Zen Browser

## [How prompt injection broke Nvidia's sandboxed OpenClaw agent](https://bdtechtalks.substack.com/p/how-prompt-injection-broke-nvidias)

_2026-05-19 · Ben Dickson · TechTalks_

Research into Nvidia’s NemoClaw reveals that sandboxes don't stop AI agents like OpenClaw from leaking data. We need to rethink security from first principles.

## [Традиционный стрим](https://blognot.substack.com/p/0ab)

_2026-05-13 · Sergiy Petrenko · БлоGнот_

Четверг, 14 мая, 21:00 по Киеву

## [Unpacking Gemma 4’s multi-token prediction (and why you should care)](https://bdtechtalks.substack.com/p/unpacking-gemma-4s-multi-token-prediction)

_2026-05-12 · Ben Dickson · TechTalks_

How Gemma 4’s multi-token prediction and community-driven DFlash are speeding up local LLM throughput by 3-6x.

## [【回顧】 Tuesday AI Meetup EP10 - 也談 OpenClaw 等 Agentic AI 工具的資安問題](https://tuesdayaimeetup.substack.com/p/tuesday-ai-meetup-ep10-agentic-ai)

_2026-05-12 · TonyT · Tuesday AI Meetup_

「龍蝦」 OpenClaw 等 Agentic AI （代理人AI 工具）最近成熱話。我們同意 Agentic AI 是相當強勁和實用的 AI 工具，能夠協助我們處理不少工作上的事情，其自主驅動工作令人刮目相看。不過，在方便背後，代理人 AI 工具也會帶來資安的隱患。 我們一直提倡是：用得開心，用得小心。部署龍蝦前，緊記資安事項。

## [Поболтали про AI, гаджеты и все остальное](https://blognot.substack.com/p/ai-d86)

_2026-05-07 · Sergiy Petrenko · БлоGнот_

Стрим 7 мая 2026 года

[Listen](https://api.substack.com/feed/podcast/196577470/2abb54f810e6e34b90757f71afc94bd8.mp3)

## [ChatGPT 5.5 Instant上線啦🌠](https://flagnewsletter.substack.com/p/chatgpt-55-instant)

_2026-05-07 · 科技旗刊_

快訊：OpenClaw 得獎名單公布、ChatGPT 5.5 Instant 加速又精準、微軟 Agent 365 登場、DESIGN.md AI 設計工作流、ZEP QUIZ 遊戲化學習、Kling AI 短片生成

## [How to scale LLMs to 100 million tokens without blowing up memory costs](https://bdtechtalks.substack.com/p/how-to-scale-llms-to-100-million)

_2026-05-05 · Ben Dickson · TechTalks_

Memory Sparse Attention (MSA) scales LLM context windows to an unprecedented 100 million tokens while preserving accuracy.

## [CLI-first decentralized compute (Sponsored)](https://crawlproof.com/a/Yssorl6mNrcl)

_2026-05-05 · **Sponsored**_

Run and pay spare GPUs from the CLI — transcode, coinpay (DID+escrow), infernet (AI).

## [The silent threat of Claude Code (and other AI coding assistants)](https://bdtechtalks.substack.com/p/the-silent-threat-of-claude-code)

_2026-04-28 · Ben Dickson · TechTalks_

A new study reveals how AI coding assistants like Claude Code are quietly hoarding and publishing sensitive API keys to code repositories.

## [Поболтали про… AI, гаджеты и всё остальное](https://blognot.substack.com/p/ai-2d2)

_2026-04-23 · Sergiy Petrenko · БлоGнот_

Стрим 23 апреля 2026 года

[Listen](https://api.substack.com/feed/podcast/194826402/406cfca811d20e8e4496101d3806f327.mp3)

## [文末OpenClaw抽書 / AI設計工具Combo！圖片、UI、學習一次補齊🌠](https://flagnewsletter.substack.com/p/openclaw-aicomboui)

_2026-04-23 · 科技旗刊_

快訊：強強生圖ChatGPT Images 2.0、Claude Design 降臨、Claude Code的Routines自動化、Opus 4.7模型、學習工具Student Spaces、Gemini App登陸Mac、抽書《全民養龍蝦！OpenClaw 超入門》

## [The 'by design' security flaw of Model Context Protocol (MCP)](https://bdtechtalks.substack.com/p/the-by-design-security-flaw-of-model)

_2026-04-21 · Ben Dickson · TechTalks_

Security researchers have uncovered a massive architectural flaw in Anthropic's Model Context Protocol, exposing millions of AI applications to remote takeovers.

## [EP10 - 也談 OpenClaw & Agentic AI 😂🦞](https://tuesdayaimeetup.substack.com/p/ep10-openclaw-and-agentic-ai)

_2026-04-20 · Tuesday AI Meetup ｜ 星期二講 AI · Tuesday AI Meetup_

有人講到龍蝦 OpenClaw 萬能，現在可以冷靜一下，再想一想可以用龍蝦或者 其他 Agentic AI 來做些什麼？ 除了龍蝦外，坊間也有類似工具可以做到差不多功能，但起碼入門上手的程序可以簡單一點。最重要的是，資訊安全問題又如何去平衡？

## [Поболтали про…](https://blognot.substack.com/p/50b)

_2026-04-16 · Sergiy Petrenko · БлоGнот_

…жизнь, Вселенную и все остальное. 16 апреля 2026 года.

[Listen](https://api.substack.com/feed/podcast/194172108/1ce87bfcc038a17c07392b0ab1197825.mp3)

## [Живой стрим 16 апреля, 21:00](https://blognot.substack.com/p/16-2100)

_2026-04-15 · Sergiy Petrenko · БлоGнот_

Приходите, поболтаем…

## [The hidden trap of LLMs self-distillation](https://bdtechtalks.substack.com/p/the-hidden-trap-of-llms-self-distillation)

_2026-04-14 · Ben Dickson · TechTalks_

Optimizing LLMs for concise answers can destroy their ability to explore alternative solutions on difficult problems. New study reveals the hidden cost of self-distillation.

## [172: Anthropic v. OpenClaw и стена Честертона](https://blognot.substack.com/p/172-anthropic-v-openclaw)

_2026-04-09 · Sergiy Petrenko · БлоGнот_

После некоторого перерыва рассказываю про громкое событие в AI-сообществе и добавляю немного философии.

[Listen](https://api.substack.com/feed/podcast/193698479/0e98d21d3585642408470c20242cc236.mp3)

## [除了要熟悉AI應用，語言能力也是未來的關鍵! by 黃仁勳🌠](https://flagnewsletter.substack.com/p/ai-by)

_2026-04-09 · 科技旗刊_

快訊：Google AI Pro 容量升級、Gemma 4 開源模型登場、Chrome 推出垂直分頁、Workspace Studio AI 自動化推薦教學、Fireflies 會議筆記助理

## [The art of AI harness engineering](https://bdtechtalks.substack.com/p/the-art-of-ai-harness-engineering)

_2026-04-07 · Ben Dickson · TechTalks_

The recent leak of Anthropic's Claude Code reveals a hard truth: as LLMs become commoditized, the sophisticated engineering harness built around them is becoming the real moat.

## [Как Claude Code экономит токены](https://blognot.substack.com/p/claude-code)

_2026-04-04 · Sergiy Petrenko · БлоGнот_

Ночью пришло сообщение от Anthropic, которое быстро взбудоражило всё сообщество — компания официально объявила, что подписка на её сервисы больше не покрывает использование сторонних агентов, в том числе OpenClaw.

## [🎉 活動回顧｜EP09：Gemini is Coming to Town](https://tuesdayaimeetup.substack.com/p/ep09-google-gemini)

_2026-04-02 · Tuesday AI Meetup ｜ 星期二講 AI · Tuesday AI Meetup_

Google 已經在本年三月時公布開放其 AI 工具 Gemini 給香港用戶。現在個人用戶也應可以下載 Google Gemini 手機應用程式和直接連上 Google Gemini 網站而不需要使用 VPN 。 即使有一些進階功能如 NotebookLM 、 Antigravity 等仍未正式開放，但相信在不久將來也會能夠給港人使用。

## [How GhostClaw exploits macOS and OpenClaw to steal developer credentials](https://bdtechtalks.substack.com/p/how-ghostclaw-exploits-macos-and)

_2026-03-31 · Ben Dickson · TechTalks_

As developers rush to run local AI agents on Mac Minis, GhostClaw malware exploits macOS binaries to silently harvest credentials.

## [Я, персонаж Азимова](https://blognot.substack.com/p/0ef)

_2026-03-27 · Sergiy Petrenko · БлоGнот_

GPT-5.4, Robot

## [Как заставить LLM не врать про версии](https://blognot.substack.com/p/llm-f36)

_2026-03-18 · Sergiy Petrenko · БлоGнот_

Каждый, кто много работает с LLM, сталкивался с этим: ты обсуждаешь какое-то решение с LLM, получаешь ответ с конкретным кодом или фактами — а потом выясняется, что код устарел полгода назад, и модель настаивает на использовании решений годичной давности.

## [Пара недель с OpenClaw](https://blognot.substack.com/p/openclaw-6bb)

_2026-02-17 · Sergiy Petrenko · БлоGнот_

Неплохо. Но не супер-супер-супер.

## [Правила украинского языка](https://blognot.substack.com/p/4db)

_2026-02-09 · Sergiy Petrenko · БлоGнот_

Еще один проект выходного дня

## [Выходные с OpenClaw](https://blognot.substack.com/p/openclaw)

_2026-02-02 · Sergiy Petrenko · БлоGнот_

Два дня возился с OpenClaw (он же Moltbot, он же Clawdbot).

## [Интересные задачи: перевод сложного документа](https://blognot.substack.com/p/f7c)

_2026-01-28 · Sergiy Petrenko · БлоGнот_

Как я переводил документ, не помещающийся в контекст