Veteran Dev: AI Can't Make Good Games, But Eliminates Sunk Costs of Trial and Error
draginol · x · 2026-08-03
The developer elaborates on AI's practical value in game development: while it can't generate good games directly, it eliminates much of the tedious work.
He notes that over decades, for every commercial game shipped, five were scrapped because the idea sucked. AI's true value lies in allowing teams to cheaply test ideas they wouldn't have dared to try before, minimizing the human potential wasted when projects are thrown out midway.
Related event: Veteran Dev: AI Can't Make Good Games But Lowers Trial Costs(3 posts)→
More from AGI Musings
- Vinod Khosla Predicts Medial Could Become the Best World Model — tracyingram · 2026-08-03
- Researcher Disappointed if AGI is Just Transformers, GRPO, and fp4 GEMMs — francoisfleuret · 2026-08-03
- Are We in the Singularity? Researchers Say Life Will Go On — danfaggella · 2026-08-03
- Is AI Reasoning Right for the Wrong Reasons? Quanta Magazine Explores — rao2z · 2026-08-03
- AI is Fulfilling Hilbert's Dream of Automated Theorem Proving — tomjaguarpaw · 2026-08-03
- Predicting AGI with Compute and API Pricing: Public Opinion Matches Model — anpaure · 2026-08-03