No New Programmers for New Game Engines - Jonathan Blow

Summary

Jonathan Blow argues that the game development industry is increasingly relying on existing engines—especially Unity and Unreal—rather than building new ones. He claims this trend reduces the number of programmers who gain deep, low-level understanding of how the hardware and software stack work.

Main Points

• Legacy engine-building required deep machine knowledge: Historically, developers had to understand the machine well to maximize hardware capability and impress players. Blow suggests this led to more robust software and fewer “misunderstanding” bugs.

• Modern developers mainly assemble games with engine “snippets”: Many programmers learn by using higher-level tools (e.g., C# components and Unity integrations) rather than creating systemic or low-level software.

• Short-term benefits, long-term costs: Reusing engines speeds development and helps teams ship sooner, but it comes at the expense of broader capability—especially the knowledge required to build and maintain foundational technology.

• Maintenance risk as expertise disappears: Unity/Unreal were created when many people at game companies continuously built engines and trained successors. If fewer people ever learn engine construction, it becomes harder to hire or find experienced maintainers—potentially reducing engine quality and slowing ramp-up for new work.

• Engines may decay over time: Even without new competing engines, existing ones could become harder to maintain as knowledgeable developers become scarce.

• Comparison to “Asimov Foundation” and historical fragility: Blow likens the current situation to societies where advanced skills are preserved by a small elite. When disruption occurs, fragile, hard-to-replicate skills can’t be carried forward because too few people understand them. He suggests that now “almost nobody knows” what’s happening on the CPU, making long-term preservation of foundational expertise uncertain.

• A potential need for a “second foundation”: He implies society may need a deliberate effort to sustain—or recreate—widespread low-level understanding, but he doubts there will be enough people doing that work (or enough higher-level developers who truly understand low-level behavior).

Presenters / Contributors

• Jonathan Blow