Denis Dyack on Unreal Engine, Optimization, and the State of the Industry
Large Unreal teams + extreme specialization
• Denis Dyack argues that most Unreal Engine projects are built by
very large teams.
• Those teams often function like an
assembly line, where individuals specialize in very narrow areas.
• Example:
• An artist whose entire focus is just
blades of grass
• They optimize their grass assets, but may have
no visibility into how other systems interact
• With
200–1,000+ developers, everyone pushes their piece forward, making full technical oversight extremely difficult.
Experience gaps + why optimization is hard
• Many teams include talented developers, but a large portion may only have
2–5 years of experience.
• Denis contrasts this with his own background:
•
35 years in the industry
• Experience
building custom engines
• Key point:
optimization is extremely hard, even for experienced teams.
Unreal as an "all-round" engine
• UE5 is described as a general-purpose engine designed to handle everything.
• But optimization needs vary drastically depending on the game:
• Racing game
• Open-world RPG
• 2D title
• Each genre requires
deep, game-specific optimization knowledge.
• There is no universal solution that automatically makes everything run well.
It's not the engine — it's the industry
• Denis Dyack's central claim: poor performance in many UE5 games is
not the engine's fault.
• Instead, it reflects the
current state of the video game industry.
Shipping without optimization
• According to Denis, many studios are in such a difficult position that:
• If the game "works," they
ship it
• There isn't enough time or budget to properly optimize
• Result: games launch in technically rough condition.
Optimization vs. content dilemma
• Optimization takes time and resources.
• That time could instead be used to:
• Add new content
• Add features
• Market "what's new"
• In many cases, decision-makers prioritize content over performance polish.
• Optimization is often seen internally as "not worth it" compared to visible additions.
Industry turmoil: "black swans" + "extinction-level event"
• Denis references multiple
black swan events in recent years.
• He also describes an
"extinction-level event" in the industry.
• Consequences:
• Many colleagues are gone
• Investment and funding have shrunk
• He predicts that for the next
2–4 years, we'll continue seeing the after-effects of earlier mistakes.
Final takeaway from Denis Dyack
• UE5 itself is not inherently the problem.
• The real issue is:
• Reduced budgets
• Lost talent
• Tight timelines
• Industry instability
• In his view,
performance issues reflect systemic industry problems, not engine design flaws.
