Asmongold Reacts to My Video – "How Modern Game Engines Degraded"

Core thesis

Older, purpose-built engines (Source, CryEngine, RenderWare, etc.) were simpler to master and allowed dramatic technical leaps and well-optimized games. Modern “universal” engines (Unreal Engine, Unity) made many tasks easier and became huge ecosystems, but introduced new complexity, de-skilled developers, and created incentives that can produce poorer optimization and technical chaos.

Historical timeline / technical highlights (examples)

• Half-Life 2 (Source, 2004) Integrated physics, lighting, and animation; Havok-like physics tied into the renderer; atmospheric sky rendering. Ran well on modest hardware and scaled across Valve projects.

• Far Cry (CryEngine 1, 2004) and Crysis (CryEngine 2, 2007) High-fidelity visuals, automatic LOD systems, streaming textures, and aggressive runtime detail management—enabled many visible objects while keeping stable FPS.

• GTA III / Vice City / San Andreas (RenderWare) Sector-based loading: only sectors in view are loaded, enabling seamless open worlds on limited hardware.

• Frostbite (DICE) and proprietary/studio engines (e.g., Decima/Kojima, internal studio engines) Engines tailored to specific goals delivered large-scale features and strong optimization (examples cited include Battlefield 6 and Death Stranding 2).

• Modern engine systems (notable examples) UE5 features like Nanite (virtualized geometry) and Lumen (real-time GI) promise automation but can create heavy, hard-to-optimize systems in practice.

Main causes of perceived degradation (technical and systemic)

• Engine complexity and legacy code Massive codebases (Unreal still carries legacy code back to UE2) are hard to master and optimize.

• Overreliance on engine automation Built-in systems reduce need for bespoke engineering but can be inefficient for specific needs.

• Poor multicore scaling Some simulations still run on single cores, creating bottlenecks.

• Runtime costs and unpredictability Real-time shader compilation, streaming, and other runtime systems can be heavy and non-deterministic.

• Business/ecosystem effects Engines acting as platforms (Epic/Unreal, Unity) — with grants, online services, platform lock-in — push studios toward defaults rather than tailored solutions.

• Industry pressures and practices Faster release schedules, hiring approaches, and management decisions deprioritize deep optimization and engineering craft.

• Economic/market decisions Example: Unity’s 2023 per-install fee controversy shows how engine companies’ business choices can break trust and drive industry change.

Consequences described

• De-skilling More developers act as “users” of an engine rather than engineers who build or deeply adapt technology.

• “Patch wait” mentality Studios sometimes defer solving problems and wait for engine vendor fixes.

• Widespread suboptimal performance While top-line visuals are achievable, many modern releases ship poorly optimized—even on high-end PCs.

Proposed remedies / suggestions & implied tips

• Invest in early optimization rather than postponing it to patches.
• Don’t rely exclusively on engine defaults; tailor systems (LOD, streaming, lighting) to the game’s needs.
• Consider proprietary or specialized engines when a project requires unique performance or features.
• Create incentives for optimization (suggested idea: a “best optimized game” award to push studios to care).
• Preserve engineering expertise and invest in dedicated engine/tool teams.

Asmongold’s reaction / viewpoints (high-level)

• Nostalgia for past big graphical leaps (from 2D → early 3D eras and console highlights).
• Agreement that multiple factors cause degradation: technical complexity and systemic issues (hiring, deadlines).
• Likes UE5’s visuals but sees it as a “Swiss Army knife” that’s hard to make excel at everything.
• Believes bad developers will produce bad optimization regardless of engine; supports rewarding optimization.
• Thinks a return to proprietary engines is possible but doubts a large-scale shift will happen—expects poor optimization may become an accepted part of game quality.

Concrete technical features and practices mentioned as positive examples

• Integrated physics and rendering pipelines (Source).
• Automatic LOD and streaming systems (CryEngine / Crysis).
• Sector-based world streaming (RenderWare / GTA).
• Studio-built engines tuned for a specific target (Frostbite, Decima, other internal engines).

Games and technologies referenced

• Games: Far Cry (2004), Crysis (2007), Half-Life 2 (2004), Portal, Left 4 Dead, CS:GO, GTA III / Vice City / San Andreas, Kingdom Come: Deliverance 2, Crimson Desert, Monster Hunter (Wilds), Battlefield 6, Death Stranding 2, Expedition 33, Star Citizen, World of Warcraft.
• Engines / tech: Source, CryEngine, RenderWare, Havok physics, Unreal Engine (UE5), Nanite, Lumen, Frostbite, Decima, Unity, Epic Online Services.

People, studios, and sources mentioned (as in subtitles)

• Asmongold (reactor / commentator)
• Original video creator (author of “How Modern Game Engines Degraded” — unnamed)
• Valve (Source engine)
• Crytek (CryEngine / Crysis)
• Rockstar (RenderWare / GTA series)
• Ubisoft, Bethesda (mentioned re: render/streaming approaches)
• DICE (Frostbite / Battlefield)
• Epic Games / Tim Sweeney (Epic CEO)
• Jeff Keighley (Game Awards, mentioned)
• Kojima (referenced in connection with Decima / DS2 — spelled “Kajjima” in subtitles)
• Unity (company and its controversial 2023 policy change)
• Other projects/studios: Sandfall (example of a well-run dev), Star Citizen (server tech comparison)