Built the data pipeline foundation and the staged AI pipelines on top of it. Runs model evaluation on EuroHPC supercomputers: measuring the smallest model each production task requires.
Key findings: Weekly registrations of domains containing "agent" spiked to more than 3x their March 2025 baseline in early 2026, and were still running at roughly 1.7x by the end...
Read →Can a general-purpose open model assess brand threat without any task-specific training? We tested four model configurations on 25 public domain disputes and found that they already recognize many of the right signals. But their readings change when you change which facts are presented first.
Read →Midterm report from EuroHPC Leonardo. We tested models across four architecture families. Each failed differently, and the search is narrowing.
Read →Continuing the work we started in our previous post. What We Did These Two Weeks Leonardo was in a planned maintenance window for most of the past two weeks. With...
Read →Not every spike in newly registered domains is a real trend. This post explains the independence test behind DomainKits Keyword Trends: how different registrants, nameservers, timing, and usage patterns help separate real registration momentum from one-operator noise.
Read →Two weeks ago, we announced our access to the EuroHPC AI Factory Playground program. Today we want to share what we have accomplished and what we have learned. First, we...
Read →We are excited to announce that Lyalpha GmbH has been granted access to the EuroHPC Joint Undertaking's AI Factory Playground program. This opportunity allows us to explore AI training at...
Read →