Yeco Flash Mini-IT: The Italian Micro-LLM That Challenges Silicon Valley at Zero Cost
118M parameters, trained for under €15, ~200 tok/s on a consumer CPU. Yeco Flash Mini-IT is our proof of concept that sovereignty beats scale.
A reflection on systemic safety and human error in critical infrastructure, dedicated to Domenico.
Marco Nasi
Founder & CEO
Systemic Safety & Human Error in Critical Infrastructure
The YecoAI team extends its sincere condolences to the family of little Domenico. The recent events regarding the transport of the organ intended for him highlight a fundamental challenge that spans medicine, aviation, and engineering: the management of human error in…
The YecoAI team extends its sincere condolences to the family of little Domenico. The recent events regarding the transport of the organ intended for him highlight a fundamental challenge that spans medicine, aviation, and engineering: the management of human error in critical processes.
While grief is the immediate and natural response, for the technological community, this event must serve as a stark reminder of our responsibility. When systems fail, the focus often lands on individual actions, but a robust engineering perspective requires us to look at the systemic design that allowed those actions to result in a catastrophic outcome.
Human error is an inevitable variable in any operational context. In safety-critical environments—whether it's organ transport or autonomous vehicle navigation—protocols cannot rely solely on human vigilance. A system that depends entirely on a person "not making a mistake" is a fragile system.
The goal of modern safety engineering is to build resilient architectures where single points of failure, including human judgment errors, are intercepted before they propagate.
The integration of AI and automated systems into logistical and medical workflows offers a path to reduce this fragility. We are moving towards an era where "double-checking" is not just procedural, but algorithmic and physical:
This event demonstrates that critical systems must be designed to reduce human error. As an industry, the mandate is clear: we must continue to develop technologies that act not just as tools, but as guardians. The complexity of biological and logistical systems requires an equally sophisticated layer of safety assurance.
Our thoughts remain with Domenico's family.
118M parameters, trained for under €15, ~200 tok/s on a consumer CPU. Yeco Flash Mini-IT is our proof of concept that sovereignty beats scale.
Ender-1 is our proprietary 3B model trained on EnderDevelopment data (free users, policy-compliant). It is in beta and rolling out on EnderDevelopment today.
278M-parameter neural model for PII detection and anonymization. Native Italian, 6 European languages, average F1 0.966.