Fabrica AI’s Post

Most people don't know this, but Singapore spans 2 climate zones. The temperate zone of an office and tropics everywhere else. Problem arises, when you develop the robot in the office, and then first come to a construction site. Initial deployments of any outside robot have to be finetuned. We've been lucky to have Woh Hup (Private) Limited as a research partner for our point-solution, the Grouting robot. The access to their constuction sites is invaluable. What did we learn so far? Main differences: - Higher humidity and temperature results in changes to grout behavior - Grout settles faster (by 20 to 100 %, depending on type) - Different sites both use different materials and have different requirements on the finish of the grout Thankfully we've been testing it on site in some capacity from the very early prototype. We have a quick-load system. It enables us to flip between nozzles in seconds. The other day I came to the site with 16 different nozzle variants, tested them all within an hour, collected feedback and had the "winner" design the very next day that satisfied the exact requirements at the specific project. We approach Fabricator development the same way. One cannot hope to develop a platform for 100x speed up of robotics development without constant contact with reality (aka deployment). The only difference is, to develop research heavy deep-tech idea like automation of robotics development, learning on a weekly basis from an external robotics company won't cut it. You have to codevelop the core with Hardware engineers and roboticiscs at hand. And the best thing? Customers of the point-solution (the Grouting robot) will subsidize the data and research engine.