In the context of TrainAsONE, 'non-deterministic' describes how our AI scheduling system operates. A deterministic system always produces the exact same output from a given starting state. In contrast, a non-deterministic system explores vast numbers of potential pathways to reach your goal and may present different solutions even from the same starting point.
Because the TrainAsONE AI continually evaluates millions of potential combinations of workouts to find the most optimal path for you, triggering a recalculation can result in a slightly different sequence of runs. This doesn't mean the new plan is wrong; rather, it means the system has identified an alternative, equally effective, or slightly more optimal route to your target based on its latest mathematical computation.
You can think of this similarly to how modern Large Language Models (LLMs) like ChatGPT or Gemini operate. If you ask an LLM the exact same question twice, it will often give you two slightly different, but equally valid answers (ignoring hallucinations!). This is because it is non-deterministic — it dynamically generates the best response in the moment rather than retrieving a static, pre-written template. In much the same way, TrainAsONE dynamically generates the optimal training plan for you in the moment.
This behaviour is a hallmark of advanced artificial intelligence and complex modelling, allowing for maximum flexibility and continuous optimisation rather than rigid adherence to a single, static path.