art. lots and lots of art. i have the art books for both The Last of Us, and The Last of Us Part 2, and there is so, so much art, and I know from experience having worked on the art book for our game, so much art didn't make the cut for the art book (for our book, i was specifically in charge of capturing environmental shots, while others on the book chose various pieces of art from across development to include). because art isn't a one and done kind of thing, a lot of time will be spent working with the various character and environmental artists to hone in and nail down on the look of the game. that kind of thing is either going to happen during the back half of year 1, and carry into year 2 of development, and it takes time with the back and forth until the art design is locked down.
once that's done, you can start bringing in 3D modelers for characters, environments, weapons, props, vehicles, etc, etc, and creating those assets takes a very long time, especially if you need a lot of them. We all know how dense and detailed Naughty Dog character models and environments are. All that shit has to be modeled, rigged for animation or physics, and then, working with the game design, level design, art directors, etc, then placed into the environments that are being worked on by the environmental modelers.
but there's also a majorly important step that i should have put before the art and 3d modeling paragraph: prototyping. once your concept and design/mechanics are figured out, you're going to want to prototype those things as quickly as possible before you put your art department to work on creating assets. you want to make sure that your gameplay designs are actually worth building on. prototyping, and noodling with that prototype until it feels right (and game feel is one of the single most important things to truly nail in your game). prototyping requires a lot of collaboration between the designers, creative directors, engineers, and programmers, and maybe a 3D modeler and animator that can whip up some janky as stand ins for characters and enemies.
depending on the complexity of your game, prototyping can take a long time. there's a benefit to working on sequels of established, and successful IP, because you have a strong foundation to build on, so you may be able to roll over into producing the necessary assets once the concept, game design and game flow is decided. For a new IP, though, if it's going to be different from your previous projects (if you are an established studio like ND), you're more than likely building all of those things from the ground up, with perhaps the game engine being the one thing you can carry over, and upgrade and optimize for the needs of the new project.
with prototyping done, and the concept being greenlit by leadership as being viable, you can in earnest start directing the team on creating all the various components of the game that need to be worked on. Script, full level design, audio design, UI concepts/iteration (this often changes a lot as a project develops, and more mechanics that may need some kind of UI element are nailed down), engine tuning and optimization for the specific project, programmers writing and implementing code based off of the design doc/plan, etc, etc. by the time the game reaches this stage, you're probably already at year 3 of dev, and it might feel like you barely started, though the game has already had a lot of significant work put into it thus far.
so your year 3 is largely spent building on the foundation established by the prototype, and the various departments are now working very, very closely to start putting it all together. the more complex the game (like an open world title, or an RPG), is going to require a lot more time, simply because of the scope of the game, and making sure that all of the planned systems and mechanics are taking shape as intended. this is such a critical stage for the game, because by this point, the team has already invested a shit ton of time, resources, and budget to get this far. having to pivot or drastically alter the design of the game in some way could be disastrous for the team, and if the project is ultimately seen as unworkable, they'll have to start the process all over again with a new concept.
the point is, making games at the level that a studio like Naughty Dog or Guerrilla Games work at, is going to take way more time these days than it did in the past, and for a combination of factors, one of the largest being asset generation, as well as implementation of the various gameplay systems.
the "fun" part of game development, ie, the time where your game starts looking and feeling like a "proper" game, doesn't usually happen until much later in the development process, as more assets and systems start to be implemented (ie, combat systems, enemy AI systems, weapon functionality, etc, etc), which also tends to excite and fire up the team as things slowly start coming together.
even with development tools getting better, most of them aren't able to solve the myriad of things that can, and will, go wrong during game development.
the reason why so many big publishers are salivating at the prospect of AI tech advancements is because of the potential to reduce dev times, but also costs. meaning, the better these tools get at doing what human's do, the less need to keep so many humans employeed, so more layoffs and incredibly talented people desperately looking for work.
it's a double edged sword, i think. better tools could make our jobs "easier," but it also runs the risk of making our jobs obsolete, or only needing to be done by a small number of people, when before, it may have required many. and a smaller team with more efficient tools isn't a guarantee that development will be any shorter than it already is.
i know i've brought it up before, but Prince of Persia: The Lost Crown, a rather straightforward 2.5D Metroidvania game still took Ubisoft Montpelier 4 years to develop (a likely reason for the $40-50 price tag it had at launch; the credits for that game took me almost 10 minutes to get through, that's how many people worked on that game). It's not some AAA mega blockbuster tentpole. it's not 100 hours long. it's not pushing the graphical envelope (though it's still a very nice looking game in my opinion), yet it still took 4 years, and hundreds of people working on it to complete.
something of the caliber of The Last of Us Part 2, or Horizon: Forbidden West, is going to take much longer, even with better dev tools, even if it was another entry in those series.
it's just really challenging to make games, and even moreso these days, when there is so much competition for player's time, energy, and money. if your game doesn't stand out, whether it be visually, some interesting gameplay or narrative hook, or, ideally for many gamers, all of the above, it's going to be hard to be successful.
higher fidelity graphics are certainly one way to get attention, but like you were saying before, graphics have reached a really, really satisfying level of fidelity. So then it boils down to content. But thanks to the PS3 era, and the PS4 era putting a strong emphasis on "value for your dollar," most games are trying to cram in as much content as possible so they aren't seen as not having enough value for your money, but creating and implementing that content takes time.
One of the things that impressed me the most about Horizon Forbidden West was that aside from the gorgeous visuals (easily one of the best looking games of the generation), was how so much of the side and extra content was more than just filler. Like, the side quests were so in depth they could have easily been mistaken for parts of the critical path. From dialogue, to cinematics, to narrative beats, etc, it was more than just, "Side quest: kill X enemies. Okay, you're done with that? Cool. Check box marked off." and there were so many side quests that were that in depth. that required so much work from GG, and was such a step up from Zero Dawn that I couldn't help but be blown away by it. The Witcher 3 (which is what I think Forbidden West reminds me the most of), is another example of that depth of content beyond just quantity of content.
sorry for the tangent, but, yeah, the competition to stand out enough to sell enough copies of your game to continue to make more games is a large part of what's driving these teams to pursue grander and grander scopes for their games, which is, in turn, increasing dev times. and, unfortunately, a lot of gamers aren't eager to spend $70 on a more streamlined, focused experience, even if it did boast the most impressive visuals in the world (i'm not sure how well Hellblade 2 sold, but man, is it a stunning game). Hell, even if the smaller, more streamlined game actually does release at a lower price (like Hellblade 2, or Spider-Man: Miles Morales), many gamers still don't see it as "value for dollar."
Because that's such an arbitrary, subjective measure, it's hard for devs to strike that balance. Spider-Man 2 has less content than Spider-Man 2018, but the content that's there is more in depth than in SM2018 (especially the side quests, but even activities like the enemy base camps have more meaning and meat to them narratively, providing further context for the overarching narratives), yet some folks were still disappointed that it was a shorter game compared to the original. Miles Morales sold less than Spider-Man 2018, despite being cheaper, more polished, but shorter than SM2018. point is, it's very unfortunate that in the gaming industry, price, visuals, and amount of content have all been wrapped up in a way that it's hard to balance all of that with rising dev costs, and longer dev times.
it's a tough nut to crack.
i know i personally don't mind shorter games that look and play well, but i don't speak for the gaming community as a whole, and definitely not the millions of gamers that don't frequent enthusiast gaming forums, and largely just see a cool ad online or in a movie theater, and decide it looks interesting enough for them to buy. unless a review happens to mention the short game length, or not feeling like there's enough content for your money. then they're more inclined to be like, "Eh, i'll wait for a sale then," which doesn't really help the devs much in an industry where more and more importance is put on how well these games do out the gate, and not necessarily how long the legs are for non-live service games. especially new IP that desperately need to build an audience as quickly as possible up front compared to an existing IP that has a built in audience, where a sequel will likely perform very well out the gate, but also have legs.
Like Ghost of Tsushima is a nice example of a new IP that managed to find its audience on PS4 (thanks to the healthy install base of the PS4 in 2020), was well received critically, and then got new life breathed into it when it released on PC 4 years after release. i'm also sure that outside gaming industry factors like the popularity of Shogun also gave a little boost of interest in Ghost and Rise of the Ronin, so, great timing on Sony/Nixes, and Sony/Team Ninja's part with the releases of those lol