This is a game from 2009. What happened?

Let's remember the last time video games made crazy progress in graphics. It wasn't that long ago. A damn 21 years ago. I don't know how many of you played the first Far Cry, but when I played it for the first time, it just blew my mind. It was the most beautiful game of its time. Even now, if I show you a real photo of a tropical island and a screenshot from this game, you-

That's really good. For 2004, that's crazy good.

You will not be able to tell the difference.

But the most amazing thing was that the first Far Cry was actually- I'm gonna open this up on YouTube

just so I can hear it more easily. Tropical Island was that the first Far Cry was actually the first game created on the Crytek engine. In other words, it was essentially just a test project, a trial run of a new engine by developers, the capabilities and features of which were still unknown. Yeah. And yet this engine was still able to produce a game that set a new graphical standard for shooters for years to come. But can we see something similar when it comes to modern engines? Unreal Engine has been around for many years.

A lot of projects have been made on it, but developers still can't master it. Even five years after the release of this engine, we still get unoptimized games that don't work well even on top of the line PCs. Have developers really not learned how to create games on Unreal Engine 5 after all these years? Or is this a problem with all the modern engines and technologies?

Why has the technical state of games deteriorated and why are we not seeing- It is crazy that you can look at Black Flag, and like this game came out in 2014, right? And this game came out last year, and Black Flag just looks, I mean, this is my opinion, infinitely fucking better. Like, way fucking better. Any graphical revolutions! And most importantly, why were developers able to do 15 years ago, what modern developers are unable to do? I wonder. Dr. Freeman I presume. Damn. On November 19th 2004 when Half-Life 2 was released on Steam,

the server simply couldn't handle it. Hundreds of thousands of people tried to activate their copies at the same time. Damn. For Valve, this was not just a game release, but a test of an entire technical chain, which ended with the sale of the game through digital distribution and began with the creation of their own game engine. Source Engine looked like something from the future at the time. For the first time physics lighting and animation worked not as separate systems but as a single process. I don't think that I can really explain how crazy it is that like I grew up

like anybody that's like my age and especially people that are a little bit older Isn't it insane that we started with Super Mario World or Super Mario Brothers? And now we're looking at things like, uh, you know, Expedition 33. Oh my god Like and like nowadays like people don't have that same- like, and I- and you remember the big jumps, don't you? Like, the big jumps, like, uh, you know, fucking- even Super Mario Brothers 3 was a big jump. And then you had, uh, Super Mario World, Yoshi's Island,

Super Mario RPG was like, what the fuck? And then Ocarina of Time comes out and then you have and and this was the big one for me for him It might have been Far Cry for me final fantasy 10. Oh my god like that game was like a that was like a console level of like Difference it was ridiculous final Fantasy X for its time was insane.

And then you just had massive improvement over massive improvement over massive improvement. And again, it makes sense why they've declined, right? Because of diminishing returns and you know, there's just the way that things are. But man, growing up with that was just fucking insane.

Well known Havoc physics was not called by the engine. It was sewn into the render. So objects reacted not according to the script, but rather according to the real mass and momentum. That is why the physics in the game for which Havok was responsible felt so progressive, smooth and realistic.

Even dynamic light affected dust particles in the sky in Half-Life 2 was rendered as a separate layer that took into account atmospheric absorption. A small detail, but it did mark the beginning of the era of realistic global illumination. And at the same time, the game ran smoothly on video cards that today would not even run Steam.

Zeus was so flexible that Valve ended up using it for another 15 years. They used it to build Portal, Left 4 Dead, CSGO, and even Half-Life 3. The engine survived an entire era, simply because its architecture was originally designed for scaling and technical development of video games, not just marketing. A few years after the release of Half-Life 2,

Crytek tried to repeat the explosive effect they got from the first Far Cry. At that time, the team developed their own engine just for demonstration purposes.

CryEngine 1 was supposed to show that an open- Look at that. Back then, this was just like, it's fucking ridiculous.

But environment could look like a rendered level, but the result exceeded the goal. People bought the game, not just for the plot or gameplay, but to test whether their PC could handle it. And because of the bright screenshots of a tropical island on the back of the disc box.

After that, Crytek completely rewrote the engine. And in 2007, Crysis was released on Cry Engine 2. This engine was the first to combine an automatic LOD system with streaming objects and textures, which allowed the engine to switch detail levels on its own without scripts. In simple terms, the same object in the game had several versions.

Look at that. Look at that. What year did this game come out at? I actually don't know. Like if this video game got released today, people would say that the graphics for it pass.

2007. Oh, my God.

A high polygon version for close distances and a simplified version for far distances. In Crysis, this process worked automatically for the first time. The engine calculated the distance to each object in the frame and replaced it with a model of the appropriate complexity in real time. The engine calculated the distance to each object in the frame and replaced it with a model of the appropriate complexity in real time. This meant that the player could see tens of thousands of objects.

You're telling me this was 2007? There's no way this was 2007. Wait, so, so...

This is so sad. What?

How? Oh my god. How have we devolved? This is the, you realize that this is the plot of Idiocracy, right? Just keep that in mind. Trees, rocks, buildings, while maintaining a stable FPS.

And yes, if it weren't for this system, Crysis would not have run on any of the top PCs of the time. As graphically advanced as this game was for its time, CryEngine made it possible. At the same time, it had dynamic texture streaming which freed up memory in real time as the player moved around the map. For 2007, this was unthinkable, so Crytek essentially implemented manual load management

which allowed the game to remain playable.

Well, almost.

Water in this game also looks phenomenal.

Meanwhile, Rockstar was working to prove the opposite, that an open world could be compact and stable rather than cumbersome. The render where engine used in GTA three Vice City and San Andreas was an invisible revolution. And if the city was divided into hundreds of 64 by 64 meter sectors

and the engine only loaded those that were in the camera's field of view, that they're effectively doing that even with servers now. I think that there's a server technology that Star Citizen's using that's similar to that. That's why the player could drive through the entire city without a loading screen on 32 megabytes of RAM. Back in the day like World of Warcraft not having loading screens between zones was like a huge settling point and

this is one of the things that like a lot of younger people like let's say you're like 21 right now this entire paradigm might seem weird to you but like back then people bought games for the graphics because something like that was just completely unthinkable and if you want to know how recent it was, think about the fact that the Ten Commandments, you know, the old movie with Charlton Heston,

won an Oscar for special effects. And go back and watch it. And think, that's how fast things have evolved. For the PS2, it was almost magic. The same principle was adopted by Ubisoft and Bethesda, and it was with RenderWare that the era of the seamless world began.

By the way, this engine was super versatile. It worked well on dozens of platforms and several generations, and the games created on it are still a pleasant memory of a warm childhood for most people. These engines were not better than modern ones in the general sense. They simply worked according to a clear technical task, and allowed games not only to function normally but also to develop both technically and in terms of gameplay.

Developers didn't just take the ready-make universal tools of these engines, but adapted them to different tasks and technical challenges, thereby constantly expanding the boundaries of what a video game can show a player. But years go by and now everything looks a little different in the industry. How did engines created for progress turn into a barrier? Why do new games, instead of pushing the boundaries, often repeat the same technical mistakes? And most importantly, how did it happen that with more powerful hardware,

we get less stable games? I think that there's like 10 reasons why. I think about half of them are technological and half of them are systematic. Things like, for example, hiring practices and then also like, you know, like how fast people have to make video games now

Well, no you can't say DEI is all of this I do think DEI is a component, but I think it's unfair to say that that's like a big reason I think one of the other big reasons is that it's the same as like what happened to Marvel after endgame Did Marvel forget how to make something as good as Thanos in Endgame? No, they didn't forget. They just simply don't care. They don't care to invest that much time and money into it because it doesn't matter.

Today it has become much easier for developers to release photorealistic bombs onto the market. And I mean literal bombs, so poorly optimized that your PC explodes. Did people have a problem with Wu-Tang? I didn't. The latest releases have different authors, but a common problem. But you can't just say that Unreal 5 is a bad engine. Rather, it's a system that has grown to such a level of complexity

that it's almost impossible to master. Engines used to be created for specific games or types of games, but now, when the market is overflowing with studios and projects of varying complexity, something universal is needed. And Epic has turned Unreal into a full-fledged ecosystem. Not just a tool, but rather an environment in which studios exist without even noticing that they have become a part of it. Unreal Engine doesn't just give studios technology,

it takes away their very engineering instinct. Everything is too convenient. Modern developers don't need to build an LOD system because there's Nanite. They don't need to work with lighting because there's Lumen. They don't need to think about how to reduce the load because the engine will do everything itself. But it only does that in theory. In practice, everything breaks. So you're saying so what he's saying basically is that

Unreal five has allowed developers to vibe code. And the problem is that there's gaps in the code. I think that's effectively what he's trying to say. Main technical reasons are real time shader compilation. So I don't know about it.

It's at the optimize new systems and most importantly, an architecture The main technical reasons are real-time shader compilation, so-called traversal stutters,

unoptimized new systems, and, most importantly, an architecture that still carries code from Unreal Engine 2. Even today, UE5 performs some simulations on a single processor core. Epic CEO Tim Sweeney publicly acknowledged this problem, saying, Many studios build games for top-of-the-line PCs and put off optimization for weaker ones until later.

And that's the problem. And you have two examples, like, I mean, for what it delivers visually, I think Crimson Desert is a good example of optimization, but it's not my best example. My best example is Kingdom Come Deliverance 2. I heard that people were playing that on like 1080 or 2080 TI's. That was insane. And I do hope that like, I wish, this is gonna sound stupid, I wish that Jeff Keighley would add in a category in the Game Awards for the best optimized game. I, I, I, maybe I should, I should DM, Jeff, I've got an idea for your show.

I've got an idea. I've got an idea for an award for you. Like that would be so good. And I feel like that would at least create a small incentive for people to design and at least have there be some acknowledgement

for that incredible level of optimization. He didn't mention the main thing.

Epic itself created a culture where later...

Yeah, Monster Hunter Wilds, remember how good that was?

...became part of the design. When the engine grows faster than the industry can master it, the result is that everyone gets used to technical chaos. And this chaos is profitable. Every game on Unreal is another project tied to the Epic ecosystem. Studios receive grants, integrate Epic online services, I don't know, pay tribute to Epic

games and ultimately become dependent on the chain built by one corporation. That's why in 2025 we see a new type of developer. Not an engineer, but simply a user of the engine. A person who doesn't create technology or even adapt it to their needs, but simply presses the enable manatee button. And when the game doesn't work well, developers don't look for a solution, but wait for a patch from Epic.

This actually kind of reminds me about like what my dad, like how many of you guys, your dad knows way more about cars than you do? And I think that's pretty normal. And how many of you guys do you know way more about computers than your dad? And so, and now, I think computers are becoming like that. Have you noticed that, I mean, like back in the day, like if you're around my age, you

remember like in order to use a gaming computer you had to have a certain level of technical knowledge you had to understand how computers kind of worked especially because like there weren't pre-built factories that were just printing these fuckers out all the time so you probably had to figure out how to install a graphics card how to put in more RAM which means that you had to get a new power supply which meant that the power supply won't fit inside the motherboard. And so you learn these things,

and I think that in the same way that, you know, cars have become a lot more streamlined, and our generation is a victim of the optimization of cars, that the future generation of young people now are a victim of the optimization of electronics and computers. And I think probably the same thing's happening with game design, is that the same people, you know, the people that wrote The Ancient Magic and Cobol, a lot of those guys are on

Social Security now, so they're outside of the industry, and they built such good mechanisms that the people now that are making computers and, you know, know making games they don't need to worry about this anymore. They don't need to even think about it. So like what is the understanding like okay well your car is making this sound okay well maybe that might be the carburetor well you don't need to know that anymore because you can just plug in a machine and the machine will give you an error code because the engine can self-diagnose

itself in most cases. So like, think about the amount of, you know, information that you just simply do not need anymore. And I think that's basically what's happened. I think this has happened with a lot of things. And I think that it's happened by the way with everything.

Like if, you know, if you asked a boomer, I'm sure a boomer would tell you that their grandparents or their dad probably said the same thing about another piece of technology. And so I think this is just how things evolve. What about Expedition 33? They don't even know how to make games? Well, I mean, I think that there's a lot. Remember what I said about how there are structural reasons? One of the structural reasons is bad management,

unreasonable deadlines, and a lack of focus. And I think Sandfall didn't have those problems. And so they were able to deliver a good game because of that.

In my opinion, it's not just Unreal Engine that has degraded. I would also like to mention the former king of indie games, Unity. I don't know, maybe 10 years ago, every other person launched this engine with the desire to create their own indie game. Unity has always been the opposite of Unreal. Simple, open, with a low barrier to entry. Dozens of cult indie games were born on it. It was an engine for

those who didn't have money, but had ideas. This was its strength. But over the years, everything changed.

After going public, the company began to think not about technology, but about profit.

And that's normal.

I don't see anything wrong with that, but the way they did it was fatal. In 2023, when Unity announced a per-install fee, developers had to pay for each installation, even for free games. For indie developers, this was a disaster. The reaction was immediate. Hundreds of studios openly declared that they were switching to other engines.

It was the moment when, for the first time in many years, the industry realized that an engine is not just software, but a tool of control. Unity has long since canceled its rule that you have to pay for each installation, but everyone is already indifferent to this engine anyway. After all, that would placate people that are like players of like an MMO but the issue is that whenever you're talking about like making a change like that and

announcing it if you have a person who's building a company and building an entire brand through this engine well you can't like the level of trust that you need to have has to be way higher and unity broke that trust they did they broke the trust. And now, you've got what's happened, is that Unity's basically dead. And Unity being dead, or not, it's not basically dead, but like, I think the amount of usage of it

probably declined massively, and I don't think it's recovered. At least, that's my guess. And it's a bad thing that Unity's dead. Not because Unity was great, but because there's an ecosystem of competing products. The moment that you lose the ecosystem of competing products is the moment that you start having a bad time. The moment that, you know, one company starts having quality degrade and everything else. Look at Unity's stock, I don't know. Unity and Unreal are not competitors. They are two different manifestations of the same problem. We can see how this universality and monopoly negatively affect the entire gaming industry.

But now we can see that there is another way, one that developers used 10 to 15 years ago. DICE stayed with Frostbite, and Battlefield 6 showed that a proprietary engine tailored to a specific goal can produce beautiful graphics with large-scale destruction of the entire map while delivering incredible optimization. And I think that another component to this too is that a lot of these engines are not from the ground up. A lot of them are forks from CryEngine or from other engines that do exist that were then used to, you know, basically be a more fine-tuned version of what the developer needed.

The game literally runs at 60 FPS on a 10 year old graphics card.

That's insane. I don't like seeing this video. How many times did I die in there? One of the most beautiful games in terms of graphics this year was also created on the studio's internal engine. What Kojima did in DS2 with his Dekuma engine is literally impossible to do on Unreal. Super fast loading, stability, and graphics that are truly next-gen. Decema is an example of good old-fashioned, highly specialized engineering without attempts

to adapt the engine to all possible scenarios. And in fact, I want to believe that in the future, developers and studios will return to their own engines more often, creating them for their own needs. I understand that this may be more difficult, but this complexity and painstaking- I actually find that to be unlikely. I find that it will be unli- I think that bad optimization will just be seen

as a larger component of game quality, and you're not going to see people go back to using different engines. I could be wrong, but that's not what I would predict.

Work always translates into success.

Well, either way, I think this is actually a really good video. I'm going to link it to you guys. How are modern how modern game engines degraded and who's to blame? I think there's a lot of reasons why. And I'm a big fan of Unreal five, but I think that his point about how because Unreal five is a Swiss army knife

That it's basically impossible for there To be anything that it's like really extremely good at 1.1 million views. I mean he deserves it This is a really great video. I gave it a I gave him a sub I gave him a like too. I think this is very Very very true and bad devs will always do bad optimizations, even with the best engine. That's true too.

Is that if you have somebody that's a dumbass, they're going to be a dumbass no matter absolutely what. It doesn't matter what happens or what's said or anything else or what they change. or anything else or what they change. They will just be dog shit no matter what.