This approach does not scale up well. Adaptivities require complex projections between geometries with different coordinate systems. It is very slow. Never use adaptivity. Other thing to point out can be seen at 4:15. Your adaptive part has 4 flush constraints. First flush (surface to surface) removes 3 degrees of freedom (one rotation and two transtalions are left) The other one removes one rotation and one translation. Third one just pick a position along the last translation left. Adaptivity resulted in on aditional flush being required. Constrains are expensive to calculate. The program has to figure out where and in what orientation the part has to be. Less constrains = faster model. And more robust, as constraints can fail when editing the geometry. Much better solution is to create a multi-solid "skeleton" model and derive the components from it. This way you can use the very same sketch to make holes for bolts and such, negating the need for projections all together. Define the general shape and connections between parts in skeleton and leave the details (fillets, rounds, ...) in derived parts. I am using quote marks as Inventor does not have proper skeleton models. But one has to make due with what's available. This will save you constraints as the parts will already be in correct positions. Drawback to this is that the part is no longer placed at the origin. You can thank Autodesk for not including local coordinate systems within ipt files. For larger projects you can create a hierarchy of skeleton models to drive different subassemblies with the main skeleton governing the positions of subs-skeletons to each other. I did not watch any other of your videos. But from the presentation I dare to guess this is an introductory lesson on bottom up modeling. Please, just no. All the lessons that can be gained from this can be learned while working with content cetre. Top down should be introduced first as it is vastly supperior and previous knowledge of bottom up is not required.