AI Observability Will Become More Important Than The Model Itself
The conversation around AI in enterprise environments has focused for years on foundation models. More parameters, larger context windows and stronger reasoning capabilities appeared to define competitive advantage. However, as organisations deploy AI into production, the problem is changing radically. What truly matters is no longer just the model itself, but the ability to understand […]
Data Entropy: The invisible problem of AI
Most AI projects fail slowly and silently. The issue is rarely a lack of computational power or the wrong choice of model. The real problem emerges when data begins to deteriorate over time and the system gradually loses coherence without anyone immediately noticing. This phenomenon could be described as data entropy. In modern enterprise environments, […]
Synthetic Data in Enterprise AI Without Privacy Risks
The rapid growth of artificial intelligence is forcing organisations to rethink how they train models without compromising sensitive information. Sectors such as healthcare, insurance, industry and legal services require massive volumes of data while simultaneously complying with increasingly strict regulations around privacy and data governance. In this context, synthetic data has become a strategic alternative […]
Prompt Injection: the SQL Injection of the AI era
Prompt Injection represents an evolution of SQL Injection within the context of generative artificial intelligence. In both cases the underlying issue arises from combining untrusted information with mechanisms capable of altering system behaviour, although in language models this boundary is far less explicit. In large language models there is no strict separation between data and […]
Kagent and the Evolution of Autonomous Infrastructure
The management of Kubernetes clusters has moved from static configuration to systems that require faster and more precise decisions. Kagent addresses this challenge by integrating AI agents directly into the Kubernetes control loop. This architecture allows the system to go beyond executing scripts and instead use a reasoning loop to adjust infrastructure based on real-time […]
Claude Code leak reveals how modern AI systems are built
A leak of internal Claude Code code has made it possible to observe how one of Anthropic’s most advanced AI systems is structured. Beyond the incident itself, what matters is what it reveals about the real architecture behind modern AI tools. The code shows that LLMs alone are not enough to reliably solve complex tasks. […]
Memory is the real bottleneck of AI
In Artificial Intelligence, most of the focus is on models, their size, the number of parameters, and the computational power they require. However, in enterprise environments, the real limit is often found in memory and how the system manages information. The first barrier appears in short-term memory, known as the context window. This mechanism determines […]
The Difference Between Probabilistic AI and Context-Aware AI
In modern artificial intelligence not all systems work in the same way. Some simply respond, generating text from probabilistic models that calculate the likelihood of the next token based on the previous context. In technical terms, the model does not search for information but estimates a probability distribution across all possible words and selects the […]
Building the Multimodal Digital Worker
For years, AI specialized in a single modality, it understood either text or images. Today, multimodal models (like Gemini or GPT-4) are capable of processing language, vision, audio, and other data simultaneously. This evolution is key because it reflects how we interact in the real world. Humans do not process words or images in isolation; […]
Human Pose Estimation
Human Pose Estimation is a computer vision technique used to localize anatomical keypoints in images or video. These points represent joints and body landmarks, enabling the construction of a skeletal model based on 2D or 3D coordinates. Detection Architectures Top-Down Approach In this approach, individuals are first detected using bounding boxes, and subsequently, the pose […]