It should be different because software is different from physics objects like food, it works differently and people interact with it differently.
Imagine you were making cookies for your friends, and Nabisco happened to get one of the cookies somehow. Should you be responsible if they decide to use the ability, conferred from having one cookie, to create an infinite number of copies of that cookie (without checking it for defects or even really investigating it much) and start selling them across the country? I’d say probably not. But our legal framework doesn’t cover it, because physical objects don’t work like that, unlike software.
We could also look at software as more like a recipe. If Nabisco decided to copy my cookie recipe, they’d unambiguous be responsible for checking it and making sure it wasn’t actually poisonous. We also have cook books, I guess if you put a poisonous recipe in a cookbook you’d have some responsibility. But this is all manageable because cooking recipes are pretty short and easy enough to verify, and the foot-guns of the hobbyist cooking field are mostly well known.
A closer recipe analogy is probably — Nabisco probably gets their bulk supplies from a network of suppliers, who have to manage things like contamination levels and fitness for a given purpose. But now we’re hitting the point where the analogy is at least as complicated as the software supply network; I’m sure there are lots of shared responsibilities and regulations in that network, and it is all professionalized and for-profit.