I work in biomedical informatics, genetics background, and your friend is technically correct and MONSTROUSLY WRONG. Bacterial populations that develop these r

I work in biomedical informatics, genetics background, and your friend is technically correct and MONSTROUSLY WRONG. Bacterial populations that develop these resistances are certainly self limiting to a small extent - the resistance makes them less fit, less survivable and powerful, than other bacteria that lack the resistance trait - this means that bacteria WITHOUT the resistance out-compete bacteria WITH it - That, in turn, means that when we take away antibiotics from the bacterial growth environment, the population with the resistance dwindles greatly or even disappears - BUT And this is a GIGANTIC BUT! - often the mechanism that causes bacteria to stop expressing the resistance trait doesn't actually remove the genetic material from the individual bacterium, and even more, doesn't remove the trait entirely from the population. The end result of this is that you have a bacterial population that isn't DIRECTLY resistant to a given antibiotic NOW, but that if you introduce an antibiotic, you will almost immediately have a bacterial population resistant to that antibiotic. Through mechanisms like turning resistance genes back 'on' that were previously 'off', genetic material communication (essentially just bacteria passing genes back and forth amongst themselves, many bacteria are frighteningly efficient at this!) - we get the end result that's the functional equivalent of a a town full of regular folk who can go to a bunker in the event of an emergency and immediately transform into the hyper-militia from hell. Like a population full of terrorist sleeper cells, only with far more destructive capability than any terrorist group!