Topic

AI and agents

Engineering, operating and evaluating AI systems, coding agents and model infrastructure.

Blog posts
168
Collections
03

What are you trying to do?

Choose a reading sequence, from the initial decision to implementation and verification.

Explore this topic

Start here

Latest in this collection

Stacked frames labelled TOOL CALLS and RESULT AI & Agents

How many tokens does Claude Code’s Code Mode save?

One job used 55,705 fewer input tokens and took 22 seconds longer. Try the terminal flag and a two-file check with a known answer.

Two stacked frames labelled PRETRAIN and SFT AI & Agents

Can you train MiniMind from scratch in two hours on an RTX 3090?

The stage table says 2.31 hours. Both trainers default to two epochs. Run the one-epoch recipe and check whether your examples retain any answer tokens.

Reviewed examples feed a compact classifier with automatic-decision and human-review paths AI & Agents

Small model fine-tuning: Airbnb’s production lesson

A specialist classifier needs more than a cheap inference call. Evaluate training data, accepted errors, fallback, full costs and the path to production.

A large euro symbol above segmented cost markers AI & Agents

Odoo AI credits: budget a reviewed quotation workflow

Budget the quotation queue before increasing volume. Separate Odoo credits, subscription uplift, initial review and correction work in a runnable synthetic ledger.

A shared starting point branches into 3 labeled paths AI & Agents

Odoo Helpdesk AI: classify, draft or escalate?

Separate ticket labels, internal reply drafts and human handoff. Use proposed support cases to test permissions and missed escalations.

A web evidence stream splits into bounded decisions in Wavect teal and violet AI & Agents

Parallel Decisions API: RAG for Decision Models in Practice

Turn current web evidence into bounded answers. A practical vendor-check example, with shared citations, optional confidence, batching costs and source limits.

Complete article directory

  1. How many tokens does Claude Code’s Code Mode save?
  2. Can you train MiniMind from scratch in two hours on an RTX 3090?
  3. Small model fine-tuning: Airbnb’s production lesson
  4. Odoo AI credits: budget a reviewed quotation workflow
  5. Odoo Helpdesk AI: classify, draft or escalate?
  6. Parallel Decisions API: RAG for Decision Models in Practice
  7. How to Validate Open Ontologies Derivation Certificates
  8. Does Scrapling Save Scraped Items in Checkpoints?
  9. TencentDB Agent Memory: Is SQLite Enough for Your AI Agent?
  10. Cue Turn Detection: Phone Numbers, Benchmarks and Limits
  11. RTK AI 1.2: Real Token Savings, Setup and Licensing
  12. Google SAM is now Agent Mesh: MCP tools across machines
  13. Anatomy for Claude Code: Interactive SVGs for Documentation
  14. EdgeQuake: PostgreSQL GraphRAG for MCP agent memory
  15. BitNet on CPU: 2B4T, real RAM needs and the 100B claim
  16. SuperPenguin: AI Cost per PR, Customer and Feature
  17. Claude Dashboards + Motion: From KPIs to Video
  18. OpenAI Intelligent UI vs OpenUI: Generative UI for Business Apps
  19. Wally by RunAnywhere: Cost per Accepted Coding Task
  20. Reducto for German Invoices: Measure Human Review Effort
  21. OpenAI Decisions API: Confidence, Refusals and Routing
  22. Claude Model Router: What Switches the Model, and When?
  23. Cloudflare Clef vs Jev: Pricing, Benchmarks and Migration
  24. Context Language Models vs Compaction: What to Pilot
  25. Caveman 3.0 for Claude Code: Local Input Compression, Recovery and Benchmarks
  26. LiteLLM Lens: Agent Trace Analysis with SQL and APIs
  27. SmythOS Studio Self-Hosting: Docker Setup, Costs and Limits
  28. DeerFlow 2.2: Docker Setup, Sandboxes and Memory
  29. CLM-8B Self-Hosting: vLLM, Action Cache and Verifiers
  30. AnyJev: LLM Calibration and Option-Order Bias
  31. LiteAgents SDK: Per-Turn Routing, Setup and Migration
  32. mcp-memory-service: Shared Memory for Claude Code and Cursor
  33. Claude Opus 5.5: Best Uses, Prompts and Effort Settings
  34. Laya and Jev for Business: 6 Practical Workflows
  35. Laya vs Jev: What the Benchmarks Mean for AI Startups
  36. Google AX: Suspend/Resume, Restart Costs and Self-Hosting
  37. WikiSkill: How Agents Evolve SKILL.md from Experience
  38. Tencent Octop Review: What Builders Actually Get for Free
  39. Wigolo Review: Local Web Intelligence for AI Agents
  40. Jev AI Review: Decision Models for Agent Workflows
  41. Supermemory: AI Agent Memory, RAG and Local Setup
  42. AI Agent Design Patterns: Start Simple, Verify Actions
  43. Claude Mods: Setup, Function Hooks and Security
  44. Voicebox: Local Voice Cloning, Dictation and MCP Setup
  45. Valyu’s 0.6B Multi-Agent Router: Results, Limits and When to Train One
  46. OpenAI Agents API Review: Migration, Costs and Data Controls
  47. Spotify shunt Review: Setup, Savings and Limits
  48. OpenBot Review: Self-Hosted AI Coworkers, Costs & Controls
  49. SwarmLLM Review 2026: Browser P2P LLM Inference Across Phones and Laptops
  50. Ramp Inspect Architecture 2026: Background Coding Agents at Scale
  51. Model Hardware Standard: Enterprise Guide to Physical AI
  52. Fonio AI Review 2026: Pricing, API, GDPR & Build vs Buy
  53. Mosaic (YC S26) Review: Shared Memory for Team AI Agents
  54. Ripwire Review 2026: AI Repo Context Without Embeddings?
  55. Atomic Multi-File Edits for AI Coding Agents: The Semaprax Lesson
  56. Phonely Alma Review: Is the Voice LLM Ready for Production?
  57. Utopia Review: Temporal Knowledge Graph for Enterprise
  58. AI Agent Knowledge Transfer: Use Frontier Models Once, Then Scale Cheaper
  59. claude-rotate: One Proxy for Multiple Claude Max Accounts
  60. NVIDIA PAIR Review: Local AI Routing and the AMD Gap
  61. Feynman Review: Is This Open-Source AI Research Agent Ready for Teams?
  62. Obscura Browser Review: Claims, Limits and Production Fit
  63. Claude Code Design System: 4 Parts for On-Brand UI
  64. Tencent Hy4 Preview Review: What Changes with the Full Hy4 Release?
  65. PageLM Review: Self-Hosting and Commercial Use
  66. ChatGPT Can Now Log In Without Seeing Your Password
  67. M6 Mac mini vs M5 Mac Studio for Local AI: Buyer’s Guide
  68. AI Agent Harness, Explained: The Reliability Layer Around an LLM
  69. FreeToken AI Review: Run Frontier MoE Models on Consumer GPUs?
  70. Darkbloom AI Review: Private Inference on Idle Macs
  71. Ox Alpha Revealed as GLM-5.3-Flash: What Changed
  72. Why Agent Edits Need Semantic Identity: Building SEMAPRAX in Rust
  73. LangChain Deep Agents Review: Is the Agent Harness Ready for Production?
  74. Pika Audio Models API Pricing: Is SFX Really 20x Cheaper?
  75. Thunder Compute’s $13M GPU Virtualization Bet: Enterprise Buyer’s Guide
  76. OpenViking Review: Filesystem Memory for AI Agents
  77. LLM-as-a-Verifier Explained: Architecture, Costs, and Production Fit
  78. AirLLM on 4 GB VRAM: How Layer-Wise Inference Really Works
  79. TrueForge Review: Is the Open-Source Agent Harness Production-Ready?
  80. Agent-Readable Websites: llms.txt, Markdown Mirrors and What Breaks
  81. Qwen3.8-27B: Self-Hosted Computer-Use Agents Without Exporting Screenshots
  82. Localized URL Slugs vs hreflang: Why We Keep One English Slug
  83. Can an AI Agent Use Your Product, or Only Read About It?
  84. Graft Review 2026: Do Agent Repo Maps Belong in Git?
  85. Codag for Coding Agents: Tool-Output Compression and Cost
  86. Smarter Token Usage: Measure Cost per Accepted Coding Task
  87. DeepSeek Harness Review: Is the Plugin Stack Production-Ready?
  88. Netflix’s vLLM and Triton Stack: 7 Production Lessons
  89. Alibaba OpenSandbox: Self-Hosting and Production Limits
  90. Transformers.js Browser AI: When Local Inference Belongs in Your Product
  91. Cloudflare Kitesurf Review: Cost, Limits and Production Fit
  92. GitHub Spec Kit Review: Is It Worth the Process?
  93. Internal AI Agent Marketplace: A 2026 Enterprise Build Guide
  94. How to Self-Host LiteLLM in Production: 2026 Guide
  95. AI-Ready Company Wiki: Architecture and Build Guide
  96. Is Linux the Best OS for AI Agents? A 2026 Infrastructure Guide
  97. Does Claude Watermark Text? The 2026 API Answer
  98. OpenKB Review: Knowledge Compiler vs RAG
  99. Unsloth Desktop Review: A Private Local AI Workstation?
  100. NeMo Switchyard 0.2: Agent Model Routing Without Training?
  101. Firecrawl AnyDoc Review: 14 Formats to Markdown
  102. MCP Cloud vs Manufact Cloud: MCP Hosting Guide
  103. Muse Glimmer 30B: Is Meta’s Local Agent Model Production-Ready?
  104. How to Make AI Writing Sound Human with Agent Skills
  105. NVIDIA NOOA Review: Are Object-Oriented Agents Production-Ready?
  106. OmniRoute AI Routing: Setup and Production Checklist
  107. Strix AI Pentesting: 30-Day Pilot and Buying Guide for 2026
  108. Hyperagent Review: Cloud AI Agents Without a Server
  109. Gemini Robotics 2: Whole-Body Control and the Pilot Decision
  110. Hark Handoff Review: The Agent That Actually Clicks
  111. Meta Muse Code Pricing: Is the Contributor Tier Safe for Client Code?
  112. pdf-inspector Review: Route PDFs Before OCR
  113. Agent Reach Review: Costs, Security and Real Limits
  114. Cloudflare Wallets for AI Agents: What Is Live?
  115. Local Multimodal AI Coding Assistant: Voice, OCR and Privacy
  116. DeepSeek V4 Flash 0731 on One AI PC: What Actually Works?
  117. YC QM Agent Review: Is Quartermaster Ready for Work?
  118. jcode vs Claude Code: Is the Rust Harness Worth Switching To?
  119. Lightpanda Browser for AI Agents: Production Guide
  120. Graph Engineering for AI Agents: When Does a Knowledge Graph Pay Off?
  121. Multi-Model AI Coding Agent Stack: A Team Buying Guide
  122. Is MCP Stateless Now? Your Server Migration Checklist
  123. MCP Is Not a Security Boundary: Protect Agent Data
  124. Fine-Tune Gemma 4 Free with Unsloth and Colab
  125. Can AI Agents Talk to Each Other? A Band Setup Guide
  126. Taalas HC1 Review: Is a Hardwired LLM ASIC Worth It?
  127. LeanCTX Technical Field Report: 64.1% Less Context
  128. llmfit Guide: Which Local LLM Fits Your Hardware?
  129. CLIProxyAPI: Run GPT-5.6 Sol Inside Claude Code
  130. Miso TTS Self-Hosted vs API: Cost, Latency and VRAM Reality
  131. Enterprise MCP Authorization Architecture
  132. Compress RAG Vectors to 2 Bits: Is Data-Oblivious Quantization Production-Ready?
  133. LMCache Reported 13.7x Lower Mean TTFT with a Shared KV Cache
  134. Lossless LLM Weight Compression vs 8-bit GGUF: What Is Ready for Production?
  135. Meterless Review 2026: Is This AI Agent Context Layer Ready?
  136. NVIDIA Nemotron 3.5 ASR: Is Free Self-Hosted STT Ready for Voice Agents?
  137. Claude Code vs OpenCode: Which Costs Less for a Team in 2026?
  138. Kimi K3 for EU Companies: API Cost, Data Risk, and a Pilot Plan
  139. Mesh LLM Review: Can One Large LLM Run Across Multiple Computers?
  140. Graphify Review 2026: Is a Codebase Knowledge Graph Worth It?
  141. Bonsai 27B Review: Can a 27B LLM Really Run on a Phone?
  142. Soofi S: Is Germany’s Sovereign LLM Ready for Business?
  143. AI Pilot Kill-or-Scale Scorecard: 12 Metrics to Check After 30 Days
  144. T3MP3ST Review 2026: Can It Replace a Penetration Test?
  145. Colibri Runs GLM-5.2 on Consumer Hardware. Here Is the Catch.
  146. Open Knowledge Format (OKF): The Enterprise Guide
  147. AI Agent Cost per Action: Why Agentic Workflows Blow Up Token Bills
  148. When Local Models Beat APIs: A Break-Even Calculator for EU Companies
  149. LLM Cost Calculator 2026: Cost per Task, Not Cost per Token
  150. Pxpipe Review: Can Images Cut Claude Code Token Costs 60%?
  151. Cheaper Per Token Can Still Cost More Per Task
  152. How to Use Claude Fable 5.2 in Claude Code
  153. Anthropic Says No Open-Weights Ban. The Cost War Is Still Real.
  154. What an Internal AI Assistant Actually Costs in the DACH Region (2026)
  155. LLM Gateways Compared 2026: LiteLLM vs OpenRouter vs Portkey vs RouteLLM
  156. Self-Hosting LLMs in the EU: When Open Weights Actually Pay Off
  157. Open-Weight LLMs 2026: DeepSeek, Qwen, Kimi, GLM, Llama and Muse
  158. The Bottleneck Was Never Intelligence. It Was Context.
  159. ChatGPT Enterprise vs Copilot vs Custom RAG
  160. MCP vs RAG vs Agent Skills vs Custom GPTs
  161. How to Cut LLM Token Costs in 2026
  162. AI Agent Pilot in 30/60/90 Days
  163. Permissions-First RAG over SharePoint, Confluence, Drive
  164. RAG Production-Readiness Checklist for EU Companies
  165. When Is an LLM Eval Worth Building? Cost, ROI, and Trusting the Judge
  166. Why AI Agent Projects Get Cancelled
  167. RAG vs Fine-Tuning vs Long-Context 2026
  168. LLM API Costs 2026. Architecture Shift
Inbox, without the noise

Get the next AI and agents field note

One concise email when we publish. No tracking pixels, and no inbox filler.