Mensagens

Publicação em destaque

Conflitos institucionais: ATRP, FPA e o montanhismo federado

Post da série História do Trail Running em Portugal: os conflitos institucionais da modalidade entre o atletismo federado (FPA), o montanhismo federado (FCMP/FPME) e a associação sectorial (ATRP) — com a proximidade do autor declarada e as lacunas assumidas.

The Inversion, Confirmed: What Two Months Did to The Great Inversion

Imagem
Series Navigation The Great Inversion  |  Fifty Years of Paradigm Shifts  |  The Unreliable Agent  |  The Burning Question  |  The Physical App Store  |  Current: The Inversion, Confirmed A Two-Month Follow-Up The Inversion, Confirmed What two months did to The Great Inversion — and the one rung of the ladder I never saw coming. By Luís Matos Ferreira · June 2026 · Licensed CC BY 4.0 Cover image for this follow-up essay. On the evening of 12 June 2026, the best model I had access to stopped working for me — because of where I was born. I had spent three days with Claude Fable 5. It launched on the 9th as Anthropic's most capable public system, and for those three days it was the sharpest collaborator I had ever put in front of a specification. Then, at 5:21pm Eastern, the US Commerce Department handed Anthrop...

From the Software App Store to the App Store of the Physical World

Imagem
Series Navigation The Great Inversion  |  Fifty Years of Paradigm Shifts  |  The Unreliable Agent  |  The Burning Question  |  Current: The Physical App Store  |  The Inversion, Confirmed  |  Watch videos Physical AI Companion The Physical App Store: How Robot Skills Extend The Great Inversion into the Real World From mobile apps to AI skills to robot skills: the distribution model is crossing from software into embodied behavior. May 2026 · A companion piece to The Great Inversion Cover image for this companion chapter. Companion article context: The Great Inversion argued that when implementation gets cheaper, engineering effort shifts toward specification, supervision, and accountability. This companion extends that logic into robotics: when behavior becomes installable as a skill, the hard part shifts again, from...

The Unreliable Agent: Why Guardrails Are Not Guarantees

The Unreliable Agent: Why Guardrails Are Not Guarantees Companion Essay Back to Main Article Series Navigation The Great Inversion  |  Fifty Years of Paradigm Shifts  |  Current: The Unreliable Agent  |  The Burning Question Agentic Reliability Companion The Unreliable Agent: Why Guardrails Are Not Guarantees A companion section for The Great Inversion, focused on the operational and security limits of natural-language guardrails in agentic systems. There is a comforting myth forming around AI agents: that if we give them better prompts, stricter system instructions, more detailed guardrails, and a checklist of forbidden actions, they will behave like obedient junior engineers. They will not. The uncomfortable truth is that AI agents are not deterministic programs executing rules. They are ...