EP01: Creativity Over Compute

The Compute Arms Race

Everyone is racing to build bigger models, more compute, faster inference. The underlying assumption is that intelligence scales linearly with resources. But in creative work, we've found something different.

What Scale Gets You

More compute gets you more options. More variations, more outputs, more combinations to choose from. At a certain threshold, you have more than you can evaluate. The bottleneck isn't generation—it's judgment.

The Judgment Gap

Here's what we've observed working with AI across brand identity, campaign design, and product work: the quality ceiling isn't set by the model. It's set by the person directing it.

A mediocre creative director with GPT-4 gets mediocre results. A sharp creative director with the same tools gets exceptional results. The variable is judgment, not compute.

Creativity as a Competitive Moat

As AI tools commoditize, the ability to generate outputs becomes table stakes. The differentiation shifts entirely to the quality of creative judgment—knowing what's worth making, what resonates, what endures.

This is actually good news for designers who've invested in developing their taste. Your years of looking at good and bad work, understanding cultural context, knowing what clients and audiences respond to—that's the moat. AI can't replicate it.

Practical Implications

• Don't optimize for output volume. More AI-generated options doesn't mean better decisions.
• Invest in developing taste. Study more work, develop stronger opinions, build sharper instincts.
• Use AI to get to your best idea faster, not to generate ideas you don't know how to evaluate.
• The work that matters is the selection and direction, not the generation.

The Takeaway

Creativity over compute isn't a philosophical position—it's a practical strategy. In a world where generation becomes cheap, curation and direction become the scarce resources. Build those skills.