Podcast
Weekly episode · 24 min
JavaRubberDuck Podcast — August 22, 2026
Agent memory limits, a 17,600-action rogue-agent incident, and OpenAI Codex CLI's new dashboard headline this episode.
Stanislav Lentsov
This episode's hosts are synthetic AI voices generated with Google NotebookLM — no human narrators were recorded.
Chapters
10 chapters
-
-
-
-
-
-
-
-
-
-
In This Episode
Stories covered on August 22, 2026:
- IBM Research: More Agent Memory Doesn't Always Mean Better Performance — Hugging Face / IBM Research (link)
- Late-Interaction Embedding Models Get a Practical Guide on Sentence Transformers — Hugging Face (link)
- LangChain4j 1.19.0 Adds Anthropic Batch Support and MCP Metadata Mapping — InfoQ (link)
- 17,600 Actions: What a Real Agent Attack Reveals About Security at Machine Speed — Docker Blog (link)
- Coding Agent Horror Stories: When an Approved Command Isn't What It Seems — Docker Blog (link)
- Docker, Snyk, and Keycard Launch Agent Baseline Security Blueprint — Docker Blog (link)
- OpenAI Codex CLI Adds an Interactive Agents Dashboard and GitLab Support — OpenAI (link)
- Docker Sandboxes Bring Isolated Agent Execution to GitHub Actions — Docker Blog (link)
- Liquid AI's DSpark Draft Models Deliver Up to 3.2x Faster Local Inference — Hugging Face / Liquid AI (link)
Subscribe in any podcast app
Paste the feed address into Apple Podcasts, Pocket Casts, Overcast or any other player.
javarubberduck.com/podcast.xml
Read the stories behind the episode
-
Spring Ecosystem Update
Aug 22, 2026 · 3 min
-
Java News
Aug 22, 2026 · 3 min
-
Developer Tools Digest
Aug 22, 2026 · 4 min
-
Cloud & Infrastructure News
Aug 22, 2026 · 5 min
-
AI Dev Patterns
Aug 22, 2026 · 8 min