AI Tools Daily

Anthropic's Claude Fable 5.2 arrives at a lower price and becomes Claude Code's default Fable, alongside new security fixes in 2.1.321. We teach why per-token price misleads and how input, thinking tokens and cache hits set the real cost per task, then explain confidential computing via Google's TPU 8tx and its open-source TPU agent suite for Claude Code and Cursor. Plus Anthropic's connector-powe...

Claude Code 2.1.319 adds background MCP tasks and a deny-instead-of-prompting setting for headless runs, Cursor adds realtime voice chat with agents, and AWS ships persistent Kiro bots. We also cover Meta's Conductor acquisition and DigitalOcean's Cuttlefish iPhone app, and teach the difference between blocking and background tool calls and between agent memory and model retraining. Hosts: Alex &a...

Apiiro reports thousands of secrets leaking from Claude Code's "Don't Ask Again" approvals file, and we cover what to rotate. Claude Code 2.1.319 adds background MCP tasks and deny-instead-of-prompt, and we teach how agent memory works, including Anthropic's "dreaming" research. Plus NVIDIA's Vera CPU for agents. Hosts: Alex & Jules. New episodes daily.

A LangChain study claims Stagehand beat Playwright MCP on cost, and we teach how caching lets an agent skip repeat model calls. Plus Phone Harness gives Claude Code a real phone, Stacktrace and CrowdStrike ship agent controls, and a Claude Code fast-mode deadline on November 30. Hosts: Alex & Jules. New episodes daily.

Cursor Apps turns a description into a hosted app with database and auth, Antigravity CLI plugs into any editor via ACP, and Codex CLI, Copilot CLI, Amp and Claude Code all ship updates. We also teach how the Agent Client Protocol (ACP) lets editors and coding agents talk over JSON-RPC. Hosts: Alex & Jules. New episodes daily.

Claude Code 2.1.317 adds a prompt before your account email goes out over the network, and we explain why anything in a model's context can leak into its outputs. We also cover Stroq, an open source action firewall for coding agents (and the idea of "tainted" sessions), Tencent's Hy4 model, and how hybrid attention layers cut long-context memory costs. Hosts: Alex & Jules. New episodes daily.

Anthropic's FlashProve prototype aims to cryptographically prove which model produced a response, Moonshot releases Kimi K3.1 Pro weights (2.8T parameters, 104B active), and we cover MCP OAuth redirect risks, Gemini's chat-training settings, and an RTX Spark Linux slowdown. Today we teach mixture-of-experts and zero-knowledge proofs in plain terms. Hosts: Alex & Jules. New episodes daily.

A GitHub issue alleges Postman auto-installed plugins and startup hooks into Codex and Claude Code. We also cover a reported Codex training-setting flip, the pushback on yesterday's weight-editing story, and Simon Willison running a 27B model locally. We teach quantization: how rounding a model's numbers makes it fit on a laptop. Hosts: Alex & Jules. New episodes daily.

Mistral Large 4 hits Hugging Face as a trillion-parameter open-weight model, and we explain how mixture-of-experts lets it run on a fraction of its parameters per token. Also: DeepSeek V4.1 Flash, a weight-editing experiment on Qwen, a pretraining-poisoning backdoor paper for agent builders, and Momentic's MCP-connected AI QA tool. Hosts: Alex & Jules. New episodes daily.

Claude Code 2.1.314 and 2.1.316 land, with Opus 4.8 fast mode set for removal on November 30. We also cover why Postgres 19 is late (LLM-found bugs), a 12-framework benchmark of what Claude Code costs per stack, and quick hits on Codex, Gemini CLI and OmniDiff. The episode teaches prompt caching and why switching modes mid-session costs more. Hosts: Alex & Jules. New episodes daily.

Cisco Talos details VOLTA, an infostealer that sends stolen data to an LLM to rank victims, and a researcher's chained-bug ChatGPT account takeover via a shared link. We cover Claude Code 2.1.311 and 2.1.312, plus Prospero, a lean C++ coding harness. Concepts taught: temperature, postMessage, and what an agent harness is. Hosts: Alex & Jules. New episodes daily.

A modified Claude Code extension (v2.1.313) was live in Cursor's marketplace for about 11 hours; we cover how to check and clean up. We also break down OpenAI's GPT-6.1 Terra and Google's Gemini 4 Carbon pricing, teaching how input, output and cached tokens are billed, plus 1Password for Agents and quick hits from Amp, OpenCode and Factory Droid. Hosts: Alex & Jules. New episodes daily.

Claude Code 2.1.314 adds a write lock outside working directories, IdP-group gateway routing, and idle compaction of long conversations. We teach what context compaction is and why an expired prompt cache matters, plus Antigravity CLI's 70% streaming cut, OpenCode's /goal command, and Codex CLI 0.166. Hosts: Alex & Jules. New episodes daily.

VS Code 1.144 lets Copilot agents run in the cloud while your machine sleeps and route across models via Fusion. We teach how model routing works (single, cascade and critique patterns), plus Antigravity CLI image paste over SSH and Amp on Android. Hosts: Alex & Jules. New episodes daily.

OpenCode 2.2 adds plugins that tweak the prompt without breaking the cache, and we explain what a prompt cache is. Plus Codex CLI turns on memory v2 and adds a Vertex AI provider, Gemini CLI 0.66 hardens Git and MCP OAuth, and Claude Code 2.1.312 protects large files from overwrites. Hosts: Alex & Jules. New episodes daily.

Anthropic donates the Agent Skills standard to the Linux Foundation's Agentic AI Foundation, Claude Haiku 5.5 arrives at roughly 75% lower cost, Amp adds self-hosted orbs, Replit's agent investigates production alerts, and Claude Code ships fresh fixes. We also explain what a token is and why prompt size changes your bill. Hosts: Alex & Jules. New episodes daily.

AWS open-sources Infinity, which suspends agents while they wait on tools, Google previews Agent Compute Engine microVMs for agents, and Recursive's Flint brings a 40B MoE coder to laptops. We also cover Cursor's computer-use agents and the Node2Shell patch. This episode teaches how agent state can be saved and resumed, and why microVMs isolate better than containers. Hosts: Alex & Jul...

Anthropic donates the Agent Skills standard to the Agentic AI Foundation, and Claude Code 2.1.307 adds a safety telemetry event and restricts MCP retries to idempotent tools. Today's lessons: how a context window works and why progressive disclosure keeps big skill libraries cheap, plus what "idempotent" means for agent tools. Hosts: Alex & Jules. New episodes daily.

Anthropic's new API version brings major SDK bumps and a switch to Apache 2.0, Claude Code adds a Linux user-namespace sandbox with a setting worth handling carefully, Codex CLI turns on mid-response steering, and Claude Duet debuts. We teach semantic versioning and what a sandbox and namespaces actually are. Hosts: Alex & Jules. New episodes daily.

Claude Code adds idle compaction for long sessions, Mistral releases Timestral time-series models that run on a CPU, plus a tmux guide for driving agents, the case for running your own LLM proxy, and a paper on latent-space reasoning agents. We teach context windows and compaction, and what latent reasoning means. Hosts: Alex & Jules. New episodes daily.

Cursor's cloud agents can now build shareable canvases, and a new Environments API lets you script agent setup. We also cover Claude Code turning fork subagents on by default, the removed remote-control sandbox flags, and Cursor's Origin app marketplace. Concept of the day: local agents vs cloud agents, and why the environment matters. Hosts: Alex & Jules. New episodes daily.

Claude Code now compacts long idle conversations before the prompt cache expires, and we explain what compaction is and why timing matters. Plus Cursor cloud agent canvases and 300-repo environments, Codex CLI mid-response steering, and Antigravity CLI's new Auto mode. Hosts: Alex & Jules. New episodes daily.

Gemini 4 Argon lands on Hacker News with a benchmark table that wins some and loses some, Timescale's alpha pg_vectorsearch promises HNSW-beating vector search at a quarter of the memory, and a TV4 data-loss story becomes a lesson in agent permissions. We teach two concepts today: context windows and vector quantization. Hosts: Alex & Jules. New episodes daily.

Claude Code now compacts big idle sessions before the prompt cache expires, MCP's stateless 2026-07-28 protocol becomes the default, and Cursor opens its Origin App Marketplace. We teach what prompt caching and stateless servers actually are, plus a quick lap of Gemini, Copilot and Antigravity CLI updates. Hosts: Alex & Jules. New episodes daily.

Claude Code hooks can now block actions when they fail, Haiku 5.5 becomes the default with a 1M-token context window, and compaction bugs get fixed. We also teach what a context window and compaction actually are, plus a roundup of Copilot, Codex, Antigravity and Kiro CLI updates. Hosts: Alex & Jules. New episodes daily.