The Model Executable Business Systems (MEBS) Podcast - Let's talk about building real-world business systems from executable models and digital twins. The question is simple: How do we turn business models into working business systems? Enterprise Architecture, MBSE, CASE, no-code, semantic technologies and now AI Agents have all approached parts of this problem. I want to explore the approaches — including yours. What have you built? What worked? What didn’t? What are we missing? I’m inviting practitioners, architects, engineers and researchers to join future episodes and challenge the ideas. The conversation starts here: youtube.com/playlist?list=PL… #MEBS #EnterpriseArchitecture #MBSE #ModelBasedEngineering

AI Thought of the Minute Jensen Huang says an AI agent is a model, a harness, tools and Skills. Quite so. But here’s the question the Ministry has been asking: What happens when an enterprise has thousands of them? Thousands of Skills. Thousands of Tools. Thousands of instructions. And increasingly, thousands of ways for the same enterprise technology to be used. The industry has spent decades making technology more reusable. APIs became more generic. Interfaces became more flexible. Middleware became more sophisticated. And transformation programmes became very good at connecting it all together. But has anyone stopped to ask what happens to the number of possible computational paths underneath all that reuse? And what happens when apparently different business purposes can hide behind the same technical interface? The Ministry has been investigating. There may be rather more going on here than simply giving AI more tools. Jensen Huang: An Agent Is A Model, Harness, Tools + Skills”, But Does It Scale To 1000s? MoAISC 063 youtu.be/8ulyjVRZcPQ And if you want to see where the Ministry goes next, subscribe, comment, like and switch on notifications. Thank you for your attention. #AgenticAI #AIArchitecture #EnterpriseArchitecture #SemanticAI #ArtificialIntelligence #AI #APIs #DataArchitecture #DigitalTransformation #MoAISC

Further reading — the mathematics behind Federated Subject Areas (FSA), including atomicity, placement, exact cover and disjointness: zenodo.org/records/18607750/…

AI THOUGHT OF THE MINUTE Yann LeCun says machines need hierarchical models. Quite so. But what happens when the machine is expected to model something rather more complicated than a laboratory environment? An entire enterprise. If hierarchical models are becoming important again, then perhaps we should ask some rather awkward questions about the hierarchies themselves… What belongs in them? What doesn’t? Who decides where things belong? And what happens when the enterprise keeps changing? The Ministry has been investigating. Watch MOAICS 061: Yann LeCun: “Machines Need Hierarchical Models.” What Is The Effect On Enterprise AI? | MOAICS 061⁠ youtu.be/4kqPCtojpRw What do you think? Watch. Subscribe. Comment. Share. And, as always… Keep Calm, and complete the model. #AI #EnterpriseAI #ArtificialIntelligence #EnterpriseArchitecture #YannLeCun

AI THOUGHT OF THE MINUTE Yann LeCun says AI needs a world model. Quite so… But why might that become rather complicated in the corporate world? In this briefing, Leah, Permanent Secretary at the Ministry of AI Common Sense, has been investigating the implications for Corporate AI. Find out in the latest Ministry briefing. youtube.com/shorts/Cw9MjhsmT… Watch, subscribe, like and comment for more AI Thought of the Minute briefings from the Ministry of AI Common Sense. Less Hype. More Common Sense. Keep Calm, and complete the model.

New MEBS Podcast Episode The frontier of AI may be approaching a point where slowing the pace of capability development becomes part of the safety conversation. But if that happens, perhaps the more interesting question is: What do we do with the time? What if AI cannot ever be made sufficiently safe for consequential corporate execution? What would the Productivity Plan B look like? In this episode, I explore whether the next productivity revolution is actually about making AI do more work — or about reconstructing the production system so that humans, AI, software and physical machines each do the work they are best suited to do. And where FSA and MEBS fit into that proposition. Watch the full episode, and let me know what you think. Is this a credible Plan B for AI-enabled productivity? youtu.be/Jxt6bQDHEEg If you find the discussion interesting, subscribe, turn on notifications, and leave a comment — I’d particularly like to hear alternative perspectives on the productivity argument. #AIProductivity #ArtificialIntelligence #EnterpriseArchitecture #AIAgents #MEBS

AI Thought of the Minute — 059 Something rather unusual is happening at the Frontier of AI. Dario Amodei, CEO of Anthropic: “We must slow the pace at which we improve the capabilities of AI models.” Sam Altman, CEO of OpenAI: “I agree with Dario that we need to pace the frontier.” Elon Musk: “Dario is right.” And Demis Hassabis, CEO of Google DeepMind: “Dario’s essay points towards the right path forward.” Four major Frontier AI voices. One rather remarkable convergence around the need to pace the Frontier. But the Ministry has an uncomfortable question. If AI is expected to deliver massive productivity gains through transformational reconstruction… And if increasingly capable AI presents safety concerns serious enough to make the Frontier worth slowing down… What happens if the danger persists? What happens to the productivity objective? What happens to the reconstruction programme? And, rather awkwardly… What is Plan B? Not because AI has failed. Not because AI should be abandoned. Simply because contingency planning is supposed to happen before you need the contingency. ▶️ youtube.com/shorts/aCwSJWSDC… What do you think? If Frontier AI has to be paced because the safety problem cannot yet be solved, should we be investigating other routes to the same productivity objective? Subscribe, like, comment, and turn on notifications so you don’t miss the next Ministry briefing. Less Hype. More Evidence. Better Decisions. #AI #ArtificialIntelligence #FrontierAI #AISafety #AIProductivity #AIArchitecture #BusinessTransformation #BusinessReconstruction #AIAgents #AILeadership #DarioAmodei #SamAltman #ElonMusk #DemisHassabis #Anthropic #OpenAI #xAI #GoogleDeepMind #ContingencyPlanning #PlanB #AIThoughtOfTheMinute #MoAISC

AI Thought of the Minute — 058 Jensen Huang has been thinking about the problem of specification. He says: “I describe it at a level that is sufficiently specific that people generally understand the direction and it’s actionable.” And then makes an even more interesting observation: “Writing a specification is coding.” Dario Amodei has described human language as: “a necessarily imprecise medium.” Sam Altman has suggested that, in the future: “maybe some people will program entirely in natural language.” And Demis Hassabis has warned about implying: “a level of precision that is not there.” So the Ministry has a rather awkward question… If AI is learning from the words we’ve spent decades choosing, how specific are those words actually? Take something as ordinary as: Material Type. Why Type? What does that word actually tell an AI Agent? And how many of the labels, descriptions and narratives in the wider training corpus contain similarly familiar words that humans understand only because we already know the context? Perhaps the problem isn’t simply whether AI can understand our language. Perhaps we should ask what language we actually taught it. 🎬 Watch AI Thought of the Minute — 058: youtube.com/shorts/dYTsN1TtK… What do you think? Does AI really have a language problem… or have we spent decades giving it words that were never sufficiently specific in the first place? Subscribe, like, comment, and turn on notifications so you don’t miss the next Ministry briefing. Less Hype. More Evidence. Better Decisions. #AI #ArtificialIntelligence #AITraining #AIArchitecture #ContextEngineering #NaturalLanguage #SemanticAI #KnowledgeEngineering #DataArchitecture #EnterpriseArchitecture #AIAgents #LLM #JensenHuang #NVIDIA #DarioAmodei #Anthropic #SamAltman #OpenAI #DemisHassabis #GoogleDeepMind #AIThoughtOfTheMinute #MoAISC

AI Thought of the Minute — 057 Something rather unusual is happening in Frontier AI. Dario Amodei, CEO of Anthropic, has published an essay entitled: “We Must Pace the Frontier.” And he says: “We must slow the pace at which we improve the capabilities of AI models.” Sam Altman agrees: “I agree with Dario that we need to pace the frontier.” Elon Musk is rather more concise: “Dario is right.” And Demis Hassabis says Dario’s essay points towards: “the right path forward.” Four major Frontier AI voices. One rather remarkable public convergence. But the Ministry has a rather uncomfortable question… If we are going to slow the Frontier, what exactly should happen during that slowdown? Should we simply develop better ways of governing AI after it executes? Or should we use the time to construct safety into the capability before it is allowed to run? And if we do slow down… What should we require to be established before we let it speed up again? 🎬 Watch AI Thought of the Minute — 057 youtube.com/shorts/xJzbfFB0X… Subscribe. Like. Comment. Turn on notifications. Less Hype. More Evidence. Better Decisions. #AI #ArtificialIntelligence #AISafety #FrontierAI #AIAgents #AIAlignment #AIArchitecture #AIReasoning #Anthropic #OpenAI #xAI #GoogleDeepMind #DarioAmodei #SamAltman #ElonMusk #DemisHassabis #AIThoughtOfTheMinute #MoAISC

AI Thought of the Minute — 056 Dario Amodei, CEO of Anthropic, once described language rather neatly: “Language is an evolved process.” Quite so… But that raises a rather awkward question. Should we automatically trust the labels and descriptions we give AI? Leah, the Permanent Secretary to the Ministry of AI Common Sense, has been investigating. And this time, Geoff has discovered that what appears on the menu… and what actually arrives at the table… may not be quite the same thing. So what happens when the words we’ve been given don’t quite describe what is actually there? And how might we establish the difference between a variant description and an invariant reality? Leah has a rather interesting answer. youtube.com/shorts/5EKzaBJgk… If you find the Ministry’s investigations useful, subscribe, like, comment, and turn on notifications so you don’t miss the next briefing. Less Hype. More Evidence. Better Decisions. #AI #ArtificialIntelligence #AIThoughtOfTheMinute #MoAISC #Invariants #Variants #ContextEngineering #AIAgents #AIArchitecture #NaturalLanguage #AIReasoning #KnowledgeEngineering #DarioAmodei #Anthropic

What happens when we give AI the wrong words? Part 2 And what happens when we give it the right words — but at the wrong level? That is what Leah, the Permanent Secretary to the Ministry of AI Common Sense, is investigating in these two briefings. 055 — Granularity At what level are your words operating? Lex Fridman: AI Builds a Hierarchy of Concepts — Granularity Matters. MoAISC 055 youtube.com/shorts/Lrsqm6DAy… 054 — Atomicity What are your words really telling AI? Dario Amodei: Common Words Can Result In Less Meaning - Atomicity, The Problem Of Labels. MoAISC 054 youtu.be/sR1nOkOGOLA Two simple questions. But perhaps rather important ones for AI. Less Hype. More Evidence. Better Decisions. #AI #ArtificialIntelligence #ContextEngineering #AIAgents #AIArchitecture #Atomicity #Granularity #KnowledgeEngineering #AIReasoning #MoAISC

What happens when we give AI the wrong words? And what happens when we give it the right words — but at the wrong level? That is what Leah, the Permanent Secretary to the Ministry of AI Common Sense, is investigating in these two briefings. 054 — Atomicity What are your words really telling AI? Dario Amodei: Common Words Can Result In Less Meaning - Atomicity, The Problem Of Labels. MoAISC 054 youtu.be/sR1nOkOGOLA 055 — Granularity At what level are your words operating? Lex Fridman: AI Builds a Hierarchy of Concepts — Granularity Matters. MoAISC 055 youtube.com/shorts/Lrsqm6DAy… Two simple questions. But perhaps rather important ones for AI. Less Hype. More Evidence. Better Decisions. #AI #ArtificialIntelligence #ContextEngineering #AIAgents #AIArchitecture #Atomicity #Granularity #KnowledgeEngineering #AIReasoning #MoAISC

AI THOUGHT OF THE MINUTE — 053 Sam Altman thinks some people may eventually program entirely in natural language. That sounds rather exciting. But the Ministry’s Permanent Secretary has a rather more practical question: If words become the programming interface… does it matter which words you use? This is her first briefing. She is the person who actually runs the department — and while the Minister is busy asking the uncomfortable questions, the Permanent Secretary is concerned with the rather less glamorous business of making sure the answers actually work. She has identified four rather unfashionable ideas that may matter rather more than all the AI hype surrounding them. And she has a particular view on the increasingly popular idea that more words mean more context. Perhaps it is time someone looked at the detail. Watch AI Thought of the Minute — 053. ▶️ youtube.com/shorts/j6EKaArgo… If you find the Ministry’s investigations useful, subscribe, like, comment, and turn on notifications so you don’t miss the next briefing. #AI #ArtificialIntelligence #NaturalLanguage #NaturalLanguageProgramming #ContextEngineering #PromptEngineering #AIAgents #AIArchitecture #SemanticAI #KnowledgeEngineering #Programming #Precision #SamAltman #OpenAI #AIThoughtOfTheMinute #MoAISC

AI THOUGHT OF THE MINUTE — 052 What happens when Jensen Huang’s prediction about English meets Rudyard Kipling’s Six Honest Serving Men? Jensen Huang says English may become humanity’s best programming language for the future. Rudyard Kipling gave us six rather curious servants more than a century ago: What. Why. When. How. Where. Who. The Ministry has been wondering whether these two ideas have rather more in common than first appears. And if they do… what might that mean for the way we engineer context for AI? Leah, the Permanent Secretary, has been asked to investigate. Her briefing is characteristically… economical. Watch the full short to hear what the Ministry has discovered. youtube.com/shorts/46fp7EASq… If you find the Ministry’s investigations useful, subscribe, like, comment, and turn on notifications so you don’t miss the next bulletin. #AI #ArtificialIntelligence #ContextEngineering #AIAgents #NaturalLanguage #PromptEngineering #JensenHuang #NVIDIA #RudyardKipling #SixHonestServingMen #AIThoughtOfTheMinute #MoAISC

AI Thought of the Minute Five frontier AI companies. One rather uncomfortable finding. Google. NVIDIA. xAI. OpenAI. Anthropic. The Ministry investigated how they provide context to their own internal autonomous AI agents. And the question was simple: Knowledge Graph logic — or annotations? The answer may be rather more significant than the Knowledge Graph industry would like. The companies actually building frontier AI are engineering agent context primarily through written-word annotations, enhanced data dictionaries and structured Skills. So perhaps the Ministry has been asking the wrong question. Not: “How do we put more logic in front of the LLM?” But: “How do we make the words we put in front of the LLM more precise?” Watch the Ministry’s concluding investigation: youtube.com/shorts/y21b1VWPk… Don’t forget to subscribe to the channel, like the video, leave a comment, and switch on notifications!!! 🔔 Five investigations. One finding. The Ministry has completed this investigation. #AI #ArtificialIntelligence #FrontierAI #AIAgents #AgenticAI #AutonomousAI #AgentContext #ContextEngineering #LLM #Annotations #WrittenWord #KnowledgeGraphs #KnowledgeGraph #PredicateLogic #Ontology #W3C #AgentSkills #Skills #DataDictionary #Metadata #GoogleAI #NVIDIA #xAI #OpenAI #Anthropic #AIArchitecture #AIEngineering #AIThoughtOfTheMinute #MoAISC

AI Thought of the Minute What does Anthropic actually give its own internal AI agents as context? Yesterday, the Ministry investigated OpenAI and found a common pattern emerging… OpenAI is using Human Annotations and Skills to provide its agents with additional written context. So today, the Ministry turned its attention to Anthropic. And Dario Amodei puts the requirement rather clearly: “Claude is powerful, but real work requires procedural knowledge and organizational context.” So how does Anthropic provide that context to its own operational Claude AI teammates? Is it Knowledge Graph logic and predicate representations? Or is Anthropic doing something rather more familiar? The Ministry investigated. Watch the investigation: youtube.com/shorts/B7FhsBEKZ… Don’t forget to subscribe to the channel, like the video, leave a comment, and switch on notifications!!! 🔔 The Ministry is now preparing its concluding investigation into the frontier AI companies — and the result may be rather interesting… #AI #ArtificialIntelligence #Anthropic #DarioAmodei #Claude #ClaudeAI #FrontierAI #AIAgents #AgenticAI #AutonomousAI #AgentSkills #Skills #ContextEngineering #AIContext #LLM #KnowledgeGraphs #KnowledgeGraph #PredicateLogic #Ontology #Annotations #WrittenWord #OrganizationalContext #ProceduralKnowledge #AIArchitecture #AIEngineering #GenerativeAI #AIThoughtOfTheMinute #MoAISC

AI Thought of the Minute What does OpenAI actually give its own internal AI agents as context? Yesterday, the Ministry investigated xAI and found something rather interesting… Grok’s internal agents are being given Skills — structured written instructions providing additional context for specialist tasks. So today, the Ministry turned its attention to OpenAI. Sam Altman says: “Metadata alone isn’t enough.” So what has OpenAI built to give its own agents the additional context they need? Is it Knowledge Graph logic and predicate representations? Or something rather more familiar? The Ministry investigated. Watch the investigation: youtube.com/shorts/-8Uf0GvRO… Don’t forget to subscribe to the channel, like the video, leave a comment, and switch on notifications!!! 🔔 The Ministry is continuing its investigation into how the frontier AI companies provide context to their own internal autonomous AI agents. #AI #ArtificialIntelligence #OpenAI #SamAltman #FrontierAI #AIAgents #AgenticAI #AutonomousAI #AgentSkills #Skills #HumanAnnotations #ContextEngineering #AIContext #LLM #KnowledgeGraphs #KnowledgeGraph #PredicateLogic #Ontology #Annotations #WrittenWord #DataDictionary #Metadata #RAG #AIArchitecture #AIEngineering #GenerativeAI #AIThoughtOfTheMinute #MoAISC

AI Thought of the Minute What does xAI actually give its own AI agents as context? Yesterday, the Ministry investigated NVIDIA and found that Jensen Huang describes AI Skills as “basically a manual” — something the AI reads and then understands how to use. So today, the Ministry turned its attention to @elonmusk ’s xAI and @grok. xAI has introduced Skills for its internal AI agents. But what exactly is a Skill? Is xAI giving Grok Knowledge Graph logic and predicate representations to provide the context? Or is something much more familiar being put directly in front of the LLM? The Ministry investigated. Watch the investigation: youtube.com/shorts/TOIxSAylh… Don’t forget to subscribe to the channel, like the video, leave a comment, and switch on notifications!!! 🔔 The Ministry is continuing its investigation into how the frontier AI companies provide context to their own internal autonomous agents. #AI #ArtificialIntelligence #FrontierAI #xAI #ElonMusk #Grok #AIAgents #AgenticAI #AutonomousAI #AgentSkills #Skills #ContextEngineering #AIContext #LLM #KnowledgeGraphs #KnowledgeGraph #PredicateLogic #Ontology #Annotations #WrittenWord #AIArchitecture #AIEngineering #GenerativeAI #AIThoughtOfTheMinute #MoAISC

AI Thought of the Minute What does NVIDIA actually give its own AI agents as context? Jensen Huang describes AI Skills as “basically a manual” — something the AI reads and then understands how to use. So the Ministry asked a very specific question… When NVIDIA needs its own autonomous AI agents to understand how to perform specialist tasks, is it giving them Knowledge Graph logic and predicate representations? Or is something else being put directly in front of the LLM? Watch the investigation: youtube.com/shorts/cB0TWP4e0… Don’t forget to subscribe to the channel, like the video, leave a comment, and switch on notifications!!! 🔔 The Ministry is investigating the frontier AI companies one by one — and the pattern is becoming very interesting… #AI #ArtificialIntelligence #FrontierAI #NVIDIA #JensenHuang #NVIDIAAI #AIAgents #AgenticAI #AutonomousAI #AgentSkills #Skills #ContextEngineering #AIContext #LLM #KnowledgeGraphs #KnowledgeGraph #PredicateLogic #Ontology #Annotations #WrittenWord #DataDictionary #AIArchitecture #AIEngineering #GenerativeAI #AIThoughtOfTheMinute #MoAISC