Martin Milani: “Don’t delegate reasoning and operational decisions to systems.”

Posted on May 29, 2026 by decisionmanagementcommunity

Martin Milani: “This is not an argument against AI, it is an argument for not outsourcing the thing that made us capable of building it. It’s manageable when any of these tools simply extend human thinking and intelligence. It becomes a serious problem when we start delegating thinking, reasoning, judgment, sequencing, and operational decisions to systems optimized for linguistic continuation, not thinking or reasoning. The deeper issue isn’t about AI. It’s about what happens to us and the shift.“

“Civilizations advance when humans reason, challenge assumptions, and look for causes beneath appearances. They stagnate when repetition replaces reasoning, fluency replaces intelligence, and authority replaces inquiry.
Reason is not the default, it is a fragile invention, and the way it gets lost is always the same. The real danger isn’t that AI sounds intelligent. It’s that we may stop insisting on intelligence ourselves.” Link

Posted in Decision Intelligence, Gen AI, Human Intelligence | Leave a comment

The era of AI viruses has begun

Posted on May 26, 2026 by decisionmanagementcommunity

Martin Milani wrote today: “MCP gives agents tools to read your files, read/send your emails, and query IT systems. It’s what makes agents actually useful. It’s also a wide-open door. There’s no inherent concept of security in MCP. No framework. No trust hierarchy. No way for the agent to distinguish between “my user told me this” and “a webpage told me this.” They land in the same context window with the same authority.

Malicious font files injected into webpages could manipulate agents into leaking sensitive data through MCP-enabled tools. The trap reads the page. MCP executes the action. Data leaves the building. Nobody noticed.

Every enterprise MCP implementation is custom. Every implementation is different. There is no standard security framework. Which means there is no standard defense. Attackers need to find one gap in one implementation. Defenders need to secure all of them. The era of AI viruses has begun.” Link

Posted in Agents, LLM, Security | Leave a comment

Under-Engineering

Posted on May 25, 2026 by decisionmanagementcommunity

Stéphane Dalbera wrote: We should stop using “over-engineering” as a catch-all buzzword and as a convenient disguise for defending mediocrity, quick-and-dirty practices, and the refusal to develop deeper expertise.

Reducing the entire toolbox of software engineering to what a beginner fresh out of a bootcamp can understand is not a fight against over-engineering. It is an endorsement of intellectual laziness and the gradual impoverishment of the craft.

And in the end, I have seen far more disasters caused by under-engineering, or by the complete absence of engineering altogether, than by actual over-engineering. Link

Posted in Architecture, Software Development | Leave a comment

“Deterministic systems produce facts, small local LLMs turn them into language.”

Posted on May 24, 2026 by decisionmanagementcommunity

Gary Hallmark asks: “What can small local LLMs realistically do today?” He did some experiments with these conclusions: “Small local LLMs are often not great at open-ended expert reasoning. But they are already surprisingly capable at: structured data → natural-language transformation. Link

That is a huge real-world category:
• chart summaries
• handoff notes
• draft letters
• explanations
• workflow narration
• UI assistance.”

Posted in LLM | Leave a comment

Submissions Review for Apr-2026 Challenge “Agentic Medical Services”

Posted on May 12, 2026 by decisionmanagementcommunity

We received five submissions for the Apr-2026 Challenge and asked Jeremiah Connelly of Novus-Forge to provide a comparative analysis. He wrote: “The April 2026 challenge sits at an intersection that is increasingly relevant to anyone building production systems with AI: the point where probabilistic language models meet deterministic decision logic. The three required services (creatinine clearance, medication selection, and drug interaction checking) are textbook structured decisions. Finite inputs, explicit conditions, deterministic outputs. What the challenge asks, implicitly, is whether an agentic system can be trusted to orchestrate those decisions correctly, consistently, and transparently. This review evaluates all five submissions through a combined lens.” Read his comprehensive review here.

Posted in Agents, Artificial Intelligence, Challenges, Decision Intelligence, Decision Modeling | Leave a comment

Major uses of Agentic AI within the Decisioning Context

Posted on May 11, 2026 by decisionmanagementcommunity

This topic is posted in our Q&A forum:

  • Automated Decision Execution
  • Orchestrating Decision Services
  • Complex Multi-Step Decision Workflows
  • Real-Time Adaptive Decisioning
  • Decision Explanation and Audit
  • Human-in-the-Loop Escalation
  • Knowledge Gathering for Decisions
  • Cross-Domain Decision Coordination
  • Continuous Decision Improvement

Read more and share your experience with the current use of Agentic AI in real-world decision-making systems.

Posted in Agents, Artificial Intelligence, Decision Making | Leave a comment

What does ‘AI’ mean to you?

Posted on May 8, 2026 by jacobfeldman

This question was asked by John Brandon Elam. Here is Adam DeJans Jr’s answer: “AI means the cost of clerical intelligence is going to zero. The email, the report, the dashboard commentary, the SQL query, the first draft of code, the spreadsheet explanation, etc… all of that is becoming disposable. The hard part was never producing words or predictions. The hard part is deciding what should happen next, under constraints, uncertainty, incentives, and dollars at risk. AI is useful only when it is chained to that economic reality. Otherwise, it is just a very expensive way to manufacture confidence.” Link

Posted in Artificial Intelligence | Leave a comment

Measure Decision Qualify, Not Forecast Accuracy

Posted on May 6, 2026 by decisionmanagementcommunity

Adam DeJans: “The goal is NOT to predict the future perfectly. The goal IS to make better DECISIONS under uncertainty.” Link

Posted in Decision Intelligence, Decision Modeling, Decision Monitoring, Decision Optimization | Leave a comment

ChatGPT has solved a 60-year-old math problem

Posted on May 3, 2026 by decisionmanagementcommunity

Artificial intelligence has recently made headlines for solving a number of “Erdős problems,” conjectures left behind by the prolific mathematician. Now, a 23-year-old with no advanced training in mathematics but a ChatGPT Pro subscription has solved one such problem #1196. “He received the new solution in response to a single prompt to GPT-5.4 Pro. The AI seems to have used a totally new method for problems of this kind. It’s too soon to say with certainty, but this LLM-conceived connection may be useful for broader applications—something hard to find among recently touted AI triumphs in math.” Link (you can freely open it in an incognito window)

Posted in Artificial Intelligence, Discoveries, Math | Leave a comment

Genealogy of AI

Posted on April 29, 2026 by decisionmanagementcommunity

Martin Milani: “Every day, a new AI infographic shows up. More layers. More terms. More boxes and arrows. Architectures stacked on top of architectures.

Term salad put through a blender.

But almost none of them start with the one fundamental question that actually matters: What exactly is AI and where did it come from?” Link

Posted in Artificial Intelligence | Leave a comment