The fundamental mistake they appear to have made was failing to research their market. Five minutes on Google would have shown them most clothing factories still use people to do the job they're trying to replace - that's in ideal conditions, completely controllable, and well-resourced ... and still the best solution is a person (with specialised tools).

They should have started by trying to replace that person. Make a $50,000 machine and sell it to businesses, then take what you learn doing that and make the home version.