Most software in other industries have clear definitions on what needs to be done and there's a much clearer path on how to implement it compared to game development.
This really just gives me the impression you aren't really considering the full scope of non-game software development. There's plenty of people working on stuff much more complicated and unpredictable than Microsoft Word knockoffs or iterative tax software. Anyone building operating systems or crafting machine-learning systems or bootstrapping whole new types of virtual infrastructure like they're doing over at Amazon is dealing with problems that are not yet well understood, with requirements that can shift dramatically, with engineering techniques that are adapted to whole new ends or created from whole-cloth.
That's just not true. General software is mainly about function, games are about entertainment and artistry.
This is just the games-specific version of the tech industry special-snowflake argument. Building entertainment software at a AAA scale isn't really that different from anything else: you still need a strong plan with good coordination and clear delineation about who can make what decisions to succeed; you still have a group of target customers to impress, and a process of iterative refinement if you don't meet their needs first time out; you're still building something whose primary purpose for existing is to capture customer revenue and whose creative elements are subsumed into that purpose. It's a field with a pretty high level of variance in interesting problems on the micro level but a very heavily circumscribed scope on the macro level outside the few people lucky enough to work on the really innovative stuff.
Like, are you really gonna tell me that the people sitting down to create the latest annual Call of Duty are going in blind to a process dictated purely by the unpredictable artistic muse? That game is getting scoped out tightly to hit a mandatory deadline, its design is getting driven by a brainstorming process for new market-distinguishing features, and it's getting built by experienced craftspeople who are applying tremendous skill to doing something they've trained to be really good at. It's more like launching a new annual fashion line than writing a novel or building conceptual sculpture for a gallery show.
If you really need a re-affirmation for how far from reality comparing (any)game development is to a standardized pipeline - look no further than the company who's made game-as-a-pipeline a science (Ubisoft), and their games are also by far among the most expensive and problem ridden productions of the industry.
I don't actually think this is a good characterization of Ubisoft. They don't have good centralized base technology, they rely on throwing huge manpower at games in an ad-hoc fashion rather than having clear global capacity planning, and they're one of the worst major publishers in the industry at post-launch support and the service-game model. EA and Activision-Blizzard are way ahead of them on this.