The person holding the settings, not the equipment, decides quality
Cooking robots and farm robots are grouped together under a single problem: how to turn a skilled worker's tacit knowledge into data and put it into a machine. A cooking robot trained on more than 3,800 collected wok-tossing motions was soundly beaten in a blind tasting against a master with 50 years of experience, but the bar the industry had set was not top quality; it was a reproducibility standard under which reaching 90% of a person's level with consistent taste counts as success. The most telling scene involves a grill robot: with the same ingredients and the same equipment, changing only the temperature setting produced a completely different result. In other words, what decides success or failure is not the hardware but the involvement of the skilled person who holds the parameters. In the second half, the 3D image-based teat-detection algorithm and the crop-scouting robot that statistically analyzes fruit size, color and count as distributions by range are effectively built the same way as a vision inspection and measurement pipeline. However, figures such as 80% of kitchen jobs being replaced or 4-fold market growth are given without sources.
Key points
- Splitting the 3,800-plus motions into condition-specific profiles is the core of turning tacit knowledge into data
- Redefines the pass mark not as "beating a person" but as "holding 90% with no variation, 24 hours a day"
- The strongest evidence is the A/B scene in which only the settings changed, on the same equipment with the same ingredients, and the results diverged
- Equipment rollout plans must include authority to tune parameters and the involvement of skilled workers
This is a note: my own write-up after reading several sources. It has no source link; events checked against a source are kept as records.