top of page


Understanding AI: From Machine Learning To Agents
The term artificial intelligence has become increasingly difficult to use precisely. A predictive model, ChatGPT, an image generator, a recommendation algorithm and an autonomous coding agent can all be described as AI, despite operating in very different ways. Technically, artificial intelligence also has a much older and broader academic definition that includes approaches such as rule-based systems, search and planning. In everyday business use, however, AI increasingly re
Dennis Hulsebos
Sep 218 min read


The Price Of A Moving Machine
Why artificial intelligence has become one of the hardest things in enterprise technology to budget AI is not one stable product with one stable price. It is a stack of models, reasoning settings, tokens, routing, service tiers and credits. Each layer can change independently, turning annual procurement into management of a moving computational market. For decades, enterprise software procurement rested on a stable unit: the licence. Products evolved, but a seat still bought
Dennis Hulsebos
Aug 246 min read


The Cost Of Thinking
Why AI Agents Make Deterministic Code More Important
Dennis Hulsebos
Jul 205 min read


The Wrong Question: Choosing An AI Coding Platform Is Not About The Model
Most comparisons between AI coding tools begin with the wrong question. They ask which model writes the best code, which assistant performs best on benchmarks, or which platform produces the cleanest pull request. Those questions matter, especially for developers who need to trust the output. But they are not enough for a business decision. Organisations rarely adopt technology because it is technically impressive. They adopt technology when it fits into how people already wo
Dennis Hulsebos
Jun 254 min read


Redefining Completion in Modern Software Engineering
Many organisations now operate with multiple software development models. One model follows a traditional engineering structure focused on stability, formal validation, and controlled releases. Another model uses AI-native workflows integrated into development environments, where generative systems assist or even lead coding tasks. Both models use the term done to describe completion, but the meaning varies significantly. The difference arises from when uncertainty is absorbe
Dennis Hulsebos
Mar 124 min read


When AI Starts Acting, Chat Stops Being the Interface
For a brief moment, chat felt like the natural endpoint of human–AI interaction. It was intuitive, flexible, and forgiving. You asked a question, the system responded, and you decided what to do with the answer. As long as AI remained an instrument for producing text, chat worked remarkably well. That logic breaks down the moment AI stops responding and starts acting. Once an AI system is capable of changing records, triggering workflows, deploying code, or modifying producti
Dennis Hulsebos
Feb 194 min read


The New Logic of Intelligence
Gemini 3 and the Move From Curation to Connection In recent weeks, Gemini 3 has sparked a quiet but meaningful shift in how we think about artificial intelligence. The conversation is no longer focused only on scale or multimodality, or benchmark performance. It is increasingly centred on something more subtle: the emergence of models that behave less like pattern recognisers and more like systems that can genuinely reason.
Dennis Hulsebos
Dec 11, 20253 min read


Garbage In, Intelligence Out: Why Reason Beats Curation
Humans learn slowly. We spend years studying, specialising, and narrowing our field of vision until we can say we truly know something. A good physicist often knows little about law, and a great lawyer rarely writes code. This limitation shapes how we think about intelligence. We divide people into generalists and specialists because our minds are bounded. There is only so much we can hold and process. When artificial intelligence arrived, we extended this same mental map to
Dennis Hulsebos
Nov 20, 20255 min read


Echoes of Thought: Memory as the Key to Human Intelligence
What makes intelligence? Is it the ability to solve puzzles, recall facts, or predict outcomes? A compelling view is that intelligence emerges from the union of two core abilities: the use of language and the use of memory. Language gives shape to thought. A sommelier can detect and describe subtle differences in wine not only because of sensory skill but also because of vocabulary. Words sharpen perception, making distinctions visible that might otherwise go unnoticed. Memor
Dennis Hulsebos
Oct 7, 20254 min read
bottom of page
