The Visionary vs. The Machine: Why Josef Fares Isn't Buying the AI Hype
The Bottom Line: Hazelight founder Josef Fares argues that while AI is a functional tool for streamlining development pipelines, generative AI has already hit a quality ceiling. We believe his stance highlights a growing industry realization: algorithms can assist in "crunching" assets, but they lack the creative spark required to ship a Game of the Year contender like It Takes Two.
Josef Fares has never been one to filter his thoughts. From his legendary "F*** the Oscars" moment to his unapologetic defense of his publishing partners, he represents a "vision-first" breed of developer that is becoming increasingly rare in an era of corporate consolidation. When Fares speaks on the industry's current obsession—AI—we should listen, not because he’s a tech evangelist, but because he’s a creator who prioritizes the "feel" of a game over the efficiency of the machine.
The State of the Industry: AI by the Numbers
Before we analyze Fares’ skepticism, we need to look at the ground truth. The industry isn't just flirting with AI; it's already in a committed relationship with it. Our analysis of recent data shows a significant shift in how games are being built behind the scenes.
| Source | Key Statistic | Primary Use Case |
|---|---|---|
| Unity 2024 Report | 62% of studios utilize AI tools | Animation & Prototyping |
| GDC 2024 Survey | 33% of developers use AI personally | Coding & Speeding up workflows |
| Hazelight Reality | Long-term AI integration | Realizing creative vision |
Why Generative AI is Hitting a Wall
Fares made a stinging observation regarding generative tools like Midjourney: the "wow factor" has plateaued. We’ve seen this before in gaming history. Early procedural generation promised infinite universes but often delivered "mile-wide, inch-deep" experiences (think back to the initial launch of No Man’s Sky). Fares suggests that generative AI is following the same trajectory—initial shock followed by a lack of meaningful evolution.
Our take: We agree. Generative AI is excellent at creating "sludge"—filler content that looks impressive in a vacuum but lacks the intentionality required for high-level level design. You can’t prompt an AI to create the emotional "clutch" moments that Fares’ Split Fiction or A Way Out are known for. Those require a human to understand pacing, empathy, and player psychology.
The "Sony Ghost" and the Threat of Excessive Hand-holding
While Fares focuses on development, the tech giants are looking at the player experience. Sony’s recently resurfaced "AI Ghost" patent—designed to play the game for you if you get stuck—is exactly the kind of QoL (Quality of Life) feature that risks stripping the soul out of the medium. Much like the controversial "yellow paint" on ledges or excessive UI clutter, an AI "auto-pilot" threatens to nerf the sense of achievement that comes from overcoming a challenge.
- The Problem: Outsourcing player agency to an algorithm.
- The Consequence: A reduction in the "Mastery" loop that keeps players engaged long-term.
- Fares' Counter-Point: You still need a human vision to decide what the game is, not just how to finish it.
The Verdict: Vision Over Volume
Fares also took a moment to defend Electronic Arts, claiming the publisher gets "more shit than they deserve." While that's a hot take that will likely trigger some gamers, it aligns with his pragmatism. He cares about the partners that let him execute his vision.
We believe the industry is at a crossroads. We can use AI to eliminate the "grunt work"—the tedious animation cleanup and bug testing that leads to developer burnout—or we can use it to replace the creative core. If we choose the latter, we lose the "Fares factor." You can’t automate heart, and you certainly can’t automate the kind of ballsy, expletive-laden passion that makes Hazelight’s games stand out in a sea of generic AAA titles.