Machine learning, cognitive services, and a changing relationship with software.
08 articles
Newest first
After two years of AI helping us build the product, this year we shipped it inside it: automated log analysis for safe storage devices, a troubleshooting assistant for service engineers, and support triage. The one rule that made it possible, the evaluation set that keeps it honest, and what we got wrong.
The coding agents could read our repository. The support team asked the obvious question: can the assistant look at the device's logs itself? The Model Context Protocol is how we said yes. What MCP is, how an OpenAPI description becomes an MCP server, and the security rules a safe-storage company needs first.
A benchmark design for seven tasks across React, C#, Delphi, SQL and support work: how to score outputs, record repeatable runs and compare OpenAI and Anthropic models. Measured results are not yet published.
Coding agents are not just for single-language start-ups. How my teams in Bari, Sofia, Kyiv and Gothenburg use them across React, C#, Avalonia, firmware and a Delphi codebase that arrived on a NAS: what works per stack, what we never let them do, and why the pipeline is the real safety net.
In 2022, ChatGPT became the talk of the town because of its capacity to produce text and respond to complex queries in a manner that almost seems human. It has grown extremely famous in a matter of…
Luxury brands like Louis Vuitton and Gucci are everything a shopaholic would want. These brands have been classified as the cream of brands that have kept them as the most wanted items. Thus, with the…
Machine learning is gradually becoming the driving force for every business. Business organizations, large or small trying to seek machine learning models to predict present and future demands and do…
The news about wildfire raging in Red Forest reminded me of my visit to Chernobyl this February – and of even more distant past when I first learned about this place.