Pretty much.
AI: You don't have to use plastic! Silicone, graphite, ceramics, glass, woods, and aluminum can all be used as substitutes and often have more desirable physical properties for specific applications.
CEO: I hear you, I really do, but scientists already recommended this and we've already done numerous analyses that have all concluded that it'd be too costly to implement and would leave us with products that aren't competitive.
CEO: ...Could you figure out how to increase our gross margins by suggesting changes to these designs?
AI: Sure! We can start by replacing those braze-on threaded nuts with a plastic clamp. I suppose that lag bolt could be plastic as well.
For me it's the difference between a preponderance of evidence suggesting such, and something being applied and proven until any doubt is removed.
For example, I was trying to find studs in drywall recently (last house was plaster and lathe), and looking at things Socratically, I could use a stud finder but I might be drilling into conduit or a pipe. So I was like "I can use magnets to hit drywall screws to try to confirm the presence of a stud", and it seems reasonable, but I've never attempted it in practice, and there could be all sorts of things a magnet could hit, since I've no experience with drywall, how close a steel pipe could be, any of that. So it's a belief. It'd be rather arrogant of me to accept this as a reliable method without testing this method, drill through a pipe and wind up with egg on my face.
So, I tested this by getting two magnets to stick vertically, then measured 16" out, got 2 more magnets to stick vertically, kept doing that until I hit half a dozen spots, all 16" apart. Drilled a pilot hole, felt resistance and the smell of wood, drilled a couple more.
I think somewhere between mounting a flat screen to fixing 3 closet shelves it became knowledge, not sure exactly when, but all the doubts were removed and it never blew up in my face. I can just waltz in a room and sink a bunch of holes in the right spot now without being skeptical of some electronic stud finder.
I guess what I mean to say is that testing something and having it consistently work and be reproducible is what leads to knowledge imo