I think there are multiple factors involved:

Places like twitter and reddit compete traditional forums. (those still have #screenshotsaturday - i still think we really want this back to have more life here)
U-engines cause a decrease on technical expertise, as it's simply no longer needed that much.
Some people think games are already dead. They have no idea, but i would agree to the saying ‘Going mainstream does not increase quality’.

Imagine you would use U too. You would be already done with your game! :O

But would it still be your game? : )

Regarding expertise, it's not really decreased, but focused. Epic hires all the experts.
It looks like the typical career of engine devs is like that:
Work on some local studio, e.g. to make the biggest and coolest RPG of all time. Deadlines, crunch, release to early, dissatisfaction.
It's meh, so after the games release, they leave their big local studio to found their own smaller indie studio. To do everything right! Without greedy management affecting their dream of the perfect game. :D
Then they run out of money, and investors just replace their greedy management from before. Again deadlines, crunch and failure.
It's meh again, so they take the job offer from Epic to try that.

Meanwhile, their former big local studio runs out of expertise, so they have to proudly announce switching over to using UE5. Although their big city game was technically much more impressive than the actual UE5 city demo :D

I don't think Epic is the bad guy here. They just do usual business, like all the others too.
But that's what i mean with ‘mainstream? → things will go downhill’.

So we better keep trying, even if the current moment does not look like the peak of gaming.

Calin said:
I think one of the reason for this happening is the desire to keep up with the standard of the day.

Sounds more like an excuse than the actual reason. Looking at CP2077 or Tomb Raider, none of them feels technically behind of the UE4 standard we have to compare them to.
So there can be only two reasons:
Giving Epic a cut costs less than paying in house engine team. This argument also includes artists and designers, as newcomers likely already have experience with UE, but not with your in house custom engine.
Or they can't manage their own complexity, because complexity increases but available expertise decreases.

It's probably a combination of both. But if the future is that all games run on one of two remaining engines, we will see even more stagnation and lack of innovation, only accelerating the sinking of the AAA titanic.

Calin said:
Basically the weight is shifted from the 50% technical side 50% aesthetics to the mostly aesthetics.

Obviously. To maintain technical progress, you still only need a small team of engineers, compared to the army of artists needed to create the content of huge modern games.

This completely transforms what games are meant to be. If we have one system programmer for 10 designers and 100 artists, the technical side is just toned down. No way to prevent this. It degrades from innovation lead towards fixing bugs and making things easier.
Now we can argue what's more important - tech or art. If we did, we may arrive at two perspectives:
Technical progress is almost completed. So from now on, game design no longer requires engineering skills, because we already have all functionality and mechanics we need. So making good games is just a matter of creativity of thought, and the final goal is games to become a piece of true art.
Or: Games are not meant to become art at all, because they have a function. Thus they are primarily a result of engineering, and artists are only needed to create pleasant visuals, background story, etc. Engineers invent Portal and Gravity Guns, while artists just invent character editors where we can draw our own tattoos and make woman with beards.

I'm more from the latter camp, even i made much more money with artwork than with programming. But ofc. the proper ratio is somewhere in between.

Though, i don't think opinions are relevant here at all. The main problem is that content creation simply causes too high costs. I have no doubt that's the major problem of the industry and the origin of all other issues we have discussed.
Because i don't believe in a way back to lower the standards, i think the only way is work on more efficient workflows. Better tools, procedural generation, eventually sharing content libraries, like Epic tries with Quixel.

We have to admit that we did not invest enough in content generation. The failure is the same everywhere: Hobby devs, indie devs, AAA companies. We worked on BSP trees, SSAO, visibility culling, cool physics… all kinds of runtime tech.
But we did not work enough on tools.

Calin said:
U engine has everything an amateur game/engine would need to have so it doesn`t act as a stimulus to build anything, it`s all already there.

Haha, that's actually a bad argument, but probably the true and honest reason some guys are left working on their own tech. :D

However - it may be a bad argument, but sometimes… a Portal Gun pops out. XD

Calin said:
A production game needs polished graphics both levels and characters and a 100% smooth user experience while exploring the levels. My guess is U engine doesn`t come with that.

Epic already gives you Meta Humans for free, plus a library of Quixel models. That's a lot of help and effort on their side, even if Quixel still feels super tiny in comparison to Shutterstock.
Regarding smooth experience, it seems out of the box character controllers are still way behind top notch stuff as seen in Dishonored or the recent Halo.
But maybe they improved this already. I've tried third person template in UE5, and it felt pretty good. But hard to judge without gameplay.

Their offer surely is hard to resist. ; )