I guess it depends on what people define it as. But I've seen many devs build custom harnesses of sorts, and spend lots of time on that to wire it all together, to then really not achieve much with them. Examples being tools or stuff that got superseded once mcps became a thing. Or spend hours making gigantic folder of skills that specified loads of stuff that the next generation model just could infer. Or build a gigantic parallel agent suite when the bottleneck wasn't the agents in the first place.
None of those things sound like anything I have in mind for a software factory. If anything they have the anti-factory ethos in mind.
The meaning is in the name. It is a factory for software. You have actionable work items that are put in a queue (assembly line) to be worked off (along dedicated stations). Then you have a quality assurance process at the end that rejects faulty parts.
And the goal of the developer is to optimize the process instead of the end result because you can regenerate the end result from scratch as many times as you want, not too different from a compiler.
If agents have excellent single shot performance then it would be best to lean in on it and build cattle software instead of wasting human time with AI written pet software.