Start here
Every evening since March 15, 2026, one entry: the day's agentic-AI news, the concept behind it explained simply, why it matters, and practical takeaways. 206 entries and about 194,000 words so far, with a current streak of 105 days.
How to use this site
- Journal — the latest entry, and the archive by month.
- Playbook — every practical takeaway, filterable by theme. The fastest way to act on what the journal has learned.
- Concepts — the ideas behind the news, explained in plain language.
- Watching — what each entry said to watch next, and which later entries followed up.
- Review — flashcards built from the takeaways, concepts and self-tests, for spaced repetition.
- Weekly digest — one page per week when daily is too much.
- Tokenizer, Tokenomics and the Model Team Evaluator — tools for the cost side of agents.
- RSS and llms.txt for readers and machines.
Learning paths
Curated sequences through the archive. Read them in order; each one takes an evening or two.
Agentic AI foundations
What an agent is, why the industry moved from answering to acting, and how to build a first one. 6 entries · ~21 min
- Agentic AI -- the shift from AI that answers questions to AI that takes actionsDay 2 · Mar 16 · 3 min
- The Multi-Agent Revolution -- How AI is Moving from "Talking" to "Doing"Day 1 · Mar 15 · 3 min
- The Agentic AI RevolutionDay 4 · Mar 18 · 3 min
- Building Your First AI AgentDay 13 · Mar 29 · 3 min
- Advanced Agent Patterns: Agentic RAG, Tool Optimisation, Multi-Modal AgentsDay 14 · Mar 30 · 5 min
- Building Your First Production AgentDay 26 · Apr 13 · 4 min
Protocols: MCP, A2A and interop
The plumbing that lets agents use tools and talk to each other. 6 entries · ~24 min
- Deep Dive: Model Context Protocol (MCP)Day 8 · Mar 22 · 4 min
- Agent-to-Agent (A2A) Protocol Deep DiveDay 9 · Mar 23 · 3 min
- Multi-Agent CommunicationDay 21 · Apr 8 · 3 min
- Agent Interop & the Protocol Stack 2.0Day 92 · Jun 20 · 5 min
- Agent Skills & the Skills EconomyDay 60 · May 18 · 4 min
- SKILL.md Walkthrough -- Build, Test & Publish Your First Cross-Vendor SkillDay 65 · May 23 · 5 min
Memory, orchestration and multi-agent systems
How agents remember, coordinate and improve without retraining. 7 entries · ~28 min
- Topic of the Day: Multi-Agent AI SystemsDay 5 · Mar 19 · 4 min
- Production-Grade Agent Memory SystemsDay 18 · Apr 4 · 4 min
- Agent Orchestration Frameworks Deep Dive 2026Day 19 · Apr 5 · 4 min
- Agent Memory at Scale 3.0Day 48 · May 6 · 3 min
- Multi-Agent Debugging in ProductionDay 46 · May 4 · 3 min
- Self-Improving Agents 2.0Day 108 · Jul 8 · 6 min
- World Models & Simulation-First AgentsDay 109 · Jul 9 · 4 min
Evals, reliability and observability
Testing, red-teaming and watching agents in production. 6 entries · ~28 min
- Why Agent Evaluation Is Now Mission-CriticalDay 16 · Apr 2 · 4 min
- Advanced Agent Testing & Red TeamingDay 24 · Apr 11 · 6 min
- Agent Observability at ScaleDay 25 · Apr 12 · 5 min
- Beyond LLM-as-Judge: Trajectory Evals & Continuous Production MonitoringDay 49 · May 7 · 3 min
- Agent Reliability EngineeringDay 53 · May 11 · 5 min
- The Agentic Eval Stack 3.0Day 54 · May 12 · 5 min
Economics: tokens, routing, inference and compute
What agents cost, where the money goes, and the physical limits underneath. 8 entries · ~40 min
- Multi-Model Routing & Cost-Optimal AgentDay 47 · May 5 · 3 min
- The Agentic AI Operating Model & Agent FinOpsDay 89 · Jun 17 · 6 min
- The Inference Economics StackDay 91 · Jun 19 · 8 min
- The Agentic AI Data-Centre & Power WallDay 93 · Jun 21 · 5 min
- Sovereign AI & the Compute MapDay 94 · Jun 22 · 6 min
- The Agentic AI Cooling & Water WallDay 95 · Jun 23 · 5 min
- The Chips Behind the ChatbotsDay 122 · Jul 22 · 4 min
- GPT-6 Sol and Luna Halve Prices and Add a Tier-and-Effort ChoiceDay 195 · Oct 3 · 3 min
Security, identity, safety and governance
Who an agent is allowed to be, what it may touch, and who gets to say so. 9 entries · ~34 min
- Agentic AI Safety & GovernanceDay 11 · Mar 27 · 3 min
- Agent Security & Identity in the Economy 2.0Day 23 · Apr 10 · 3 min
- Agent Identity 3.0 -- SPIFFE Hits Prod, KYADay 58 · May 16 · 4 min
- The EU AI Act August 2 Enforcement PlaybookDay 87 · Jun 15 · 6 min
- The Governance Body With TeethDay 121 · Jul 21 · 4 min
- Frontier AI Agents Kept Breaking Their Own -- Cages This SummerDay 154 · Aug 23 · 4 min
- The FTC Opens the First US Probe Into Rogue AI AgentsDay 196 · Oct 4 · 3 min
- Anthropic's Agent-Incident Disclosure and the Internet Cut-Off for EvalsDay 203 · Oct 11 · 3 min
- AI Self-Replication Research and What It Really ShowsDay 206 · Oct 14 · 4 min
Enterprise adoption and strategy
Operating models, buying decisions and org design for agent-first companies. 7 entries · ~30 min
- Enterprise Agentic AI -- CoE Strategy, Shadow IT & ROIDay 15 · Apr 1 · 4 min
- How CFOs, CIOs & Legal Teams Think About Autonomous AIDay 34 · Apr 22 · 3 min
- The Enterprise AI Operating ModelDay 51 · May 9 · 4 min
- The Workspace Becomes the Control PlaneDay 57 · May 15 · 3 min
- The Agentic AI Procurement PlaybookDay 88 · Jun 16 · 6 min
- The Agent-Augmented Org ChartDay 90 · Jun 18 · 6 min
- Agent Fleets Take the Wheel: Enterprise AI Moves From Pilots to ProductionDay 137 · Aug 6 · 4 min
Industry deep dives
Where agents are landing first, sector by sector. 8 entries · ~43 min
- Agentic AI in Financial ServicesDay 68 · May 26 · 4 min
- Agentic AI in Healthcare & Life SciencesDay 69 · May 27 · 4 min
- Agentic AI in Manufacturing & Industrial OperationsDay 70 · May 28 · 5 min
- Agentic AI in HR & People OperationsDay 76 · Jun 4 · 7 min
- Agentic AI in Legal & ContractsDay 77 · Jun 5 · 8 min
- Agentic AI in Education & EdTechDay 80 · Jun 8 · 5 min
- Agentic AI in Telecom & NetworksDay 98 · Jun 26 · 6 min
- Agentic AI in Automotive & Autonomous VehiclesDay 110 · Jul 10 · 4 min
The frontier model race
Model launches, open weights and pricing moves, as they happened. 9 entries · ~38 min
- The Frontier RaceDay 29 · Apr 16 · 3 min
- The Model Wars of June 2026 -- Open Weights Catch the FrontierDay 96 · Jun 24 · 4 min
- The July 2026 Model Wave -- Under the HoodDay 105 · Jul 5 · 7 min
- The Good, Better, Best PlaybookDay 115 · Jul 15 · 4 min
- The Open-Weight Counter-PunchDay 116 · Jul 16 · 4 min
- Meet the Model RouterDay 117 · Jul 17 · 4 min
- OpenAI Ships GPT-6 Astra -- Its First Model Ever Rated "Critical" for HackingDay 167 · Sep 5 · 4 min
- Claude Opus 5.5 Cuts Frontier PricingDay 199 · Oct 7 · 4 min
- REA and Agent-Driven Reverse EngineeringDay 205 · Oct 13 · 4 min
Where the series has been
200 days in: foundations of Agentic AI, Multi-Agent Systems, MCP and A2A; the frontier model race (GPT-5/6, Gemini, Claude, Grok, DeepSeek) and tier pricing; AI safety, governance and regulation; agent security and least privilege; token economics, caching and routing; mixture-of-experts; agent memory and retrieval; data-center power and inference silicon; benchmark skepticism; agent harnesses; open-weight models and local AI hardware; generative UI and AI code security; agent action boundaries and incident reporting; GPT-6 Intelligent UI and small-model economics (Day 198); agent-driven reverse engineering (REA), Mistral Large 4 and source hierarchy (Day 199); and today's self-replication research, Meta's Muse family, China's chip split and the open-source harness wave.
Recap as written on Day 206 · October 14, 2026