The jump from 63 models in early 2024 to 668 by late 2025 sounds impressive, but are we certain that 'tracking' automatically informs and populates a daily research dataset directly?
If we hadn't already spent time focused on this observation, would we automatically assume raw tracking data instantly becomes a validated dataset?
In hiring, I track hundreds of CVs, but that doesn't mean those raw numbers are immediately ready for strategic decisions; we need to clean, verify, and structure the information.
Simply watching something grow doesn't guarantee its immediate utility in a precise dataset without clear data processing and quality control steps.
It's a common sunk cost fallacy to believe that because we've observed something, it's inherently useful in its raw form.