Giancarlo F.

intelligent behavior is also associated with the ability to find the best combinations of possible "moves", not just a single correct step. In the Kubrick experiment, I try to combine natural language, data and constraints in programming. here I explain the combinatorics behind it: https://gfrison.com/patents/bridging-natural-language-with-data-programming-combinatorial

Accidental complexity slows down developers and limits agentic AI. Kubrick cuts it way down using relation algebra, logic, and combinatorial ideas to enable reliable agentic programming and true AI-human collaboration. From my MSc work, now moving to open-source. Presenting at PX/26 (Munich, Mar 16-20). Thoughts? https://gfrison.com/2026/pull-down-programming-complexity-kubrick

I wrote a short piece on adaptive programming systems and how they might support human–AI collaboration. While it’s not strictly about KGs, it is related on themes of representation, adaptability, and interaction. https://gfrison.com/2025/adaptive-programming-systems Curious how this might connect (or not) with ideas in the KG space.

🚀 When Creativity Meets Precision: Hybrid AI for program synthesis. What if LLMs could think like engineers, combining creativity with the rigor of logic? Large language models are reshaping how we code, automate, and problem-solve. Yet even their brilliance has limits: 🔥 Hallucinations and struggles with constraints. Meanwhile, symbolic AI suffers of scalability. But what if there’s a third way? 🔍 In this exploration, I unpack Hybrid AI as a fusion where LLMs imagine solutions and symbolic systems ground them in logic. The result? https://gfrison.com/2025/hybrid-ai-for-generating-programs

Justify Recommendations with Semantic Technologies - Justify Recommendations with Semantic Technologies 5 minute read E-Stores loose sells due to the negative biases of consumers. While salespeople give proper reasons to change consumers misbelieves, it is problematic to address those issues in an online shop. I have proposed a method to combine semantic graphs with logic programming and symbolic machine learning to restore consumer’s confidence. The digital agent detects what are the problems users might have and offers them explanations and valid arguments for not worry about, or why a given recommendation is more suitable than others. https://gfrison.com/2023/08/02/argumentation-recommendation-ecommerce-knowledge-graphs/