I’m with you on this, though I feel a conflict within myself: on one hand, the nostalgic attachment to that raw, "in-the-trenches" coding culture; on the other, the realization that building a proprietary engine is often completely unnecessary.
Diminishing marginal utility.
Why did the PC ultimately kill the demoscene? A lack of restrictions; hardware was no longer the bottleneck that — through brilliant programming exploiting specific hardware quirks — could be coaxed into conjuring up magical visual effects (or failing to do so).
On the C64 and Amiga 500, individual ingenuity correlated directly with visual output. The PC era—ushered in by the i386, refined by the i486DX, and popularized by the 586 (Pentium)—increasingly abstracted visual effects and audio illusions away from the hardware itself.
What previously had to be created in assembly language (!) — indeed, practically forced into existence through sheer effort — was now reduced to just "Yet Another Feature."
The partnership between Carmack — whose genius was brilliantly complemented (and even surpassed, a fact often forgotten) by Michael Abrash — represented the most important development duo in the history of game engines. Ken Silverman was a sensation who, unfortunately, never quite reaped the accolades he deserved. Interestingly, this highlights exactly why it was so right and important for Carmack and his peers to be their own bosses: success doesn't end up in someone else's pockets.
Nowadays, making games is essentially akin to video editing — dependent on NVIDIA, and nothing more.
The crucial factor is the ability to deploy a game across countless systems with minimal adaptation effort; while the base version suffers little visual degradation, the architecture must still allow for high-end PC systems to push the boundaries.
In other words: back then, we hand - coded animations—graphics, code, and music were a unified whole. Today, a kid on an iPhone creating a TikTok video achieves — with a thousand times higher quality — what used to take teams weeks or months to accomplish. Development costs are astronomical; code no longer needs manual optimization because compilers are inevitably better at it (multi-core, etc.). Nothing is one-dimensional or linear anymore.
For this reason, content is all that remains.
As someone who is nostalgic — a C64 demoscener and occasional Amiga 500 assembly coder — I do feel a twinge of wistfulness; yet, as a former senior manager and platform product lead, I cannot fathom clinging to the wrong approach for so long.
That misguided romanticism made me shake my head. Visually, Doom was clearly inferior to engines like Unity. I couldn't care less whether or not that commented-out line containing the `0x5F3759DF` hack was in there somewhere.
Don't hate the player, hate the game.