Humans building LEGO sets using instructions might be a thing of the past, at least if this machine has anything to say about it.
Ruocheng Wang, Yunzhi Zhang, Jiayuan Mao, Chin-Yi Cheng and Jiajun Wu of Stanford University, MIT CSAIL and the Autodesk AI Lab as shared by Ben Houston on Twitter, have created a machine that can interpret and build LEGO sets just by looking at the step-by-step instructions that we use.
In short, this program can look at a set of instructions such as the guitar used in the example video and determine which shape and colour brick needs to be placed next, move it into position according to the manual and move on to the next step. This includes smaller builds that will be attached to the main assembly.
The example featured in the project is presented digitally, though it seems plausible that this program could be applied to a physical machine to take it to the next step.