Great design rarely comes from knowing what you're doing

Color-object binding accuracy, best text-to-image model

23%

Users who rated AI reproductions unsatisfactory

60%

Users who gave up before getting what they meant

27%

Less hue drift when generation is bound to tokens

62%

What people get wrong about taste

Everyone agrees that taste is subjective. What most people miss is that subjectivity has layers. There is bad taste, good taste, and great taste. They are not the same thing, and the difference matters enormously for how we think about design, creativity, and AI.

The distinction is not academic. It is the entire difference between competent output and output that moves people. And it is exactly where current AI gets stuck.

AI has been trained on the largest corpus of human creative work ever assembled. It has mastered the rules. It produces endlessly competent, technically sound, instantly forgettable work. AI is the best "good taste" machine ever built. But great taste? That requires knowing when to drift.

Lessons from making sounds

I produced and published music for over ten years before I switched to design. I can say with complete conviction that the most important lesson from a decade of making sounds applies directly to every design decision I've made since.

Good music is structurally sound. It understands the laws of rhythm, melody, and harmony. It is in perfect pitch. It satisfies a million technical constraints. You can study theory for years to produce music that is technically flawless — and no one will remember it.

Great music is fundamentally imperfect. It is something you cannot fully put into words. Paradoxically, it often comes from the moments when you do not know exactly what you are doing. Every time I stuck rigidly to the rules, the output was competent and boring. But when I drifted — when I went outside the structure and captured the fuzziness of that specific moment — that is when I had my best work.

I call this disciplined drift. It is not random experimentation. It is an extremely disciplined approach to pushing past what is known and comfortable. You keep drifting, but you are razor-sharp focused on how you are drifting and why. Eventually you end up somewhere that feels "just right" in a way that is difficult to articulate — and you stop. One step further would be noise.

The practice of not knowing

The greatest products and discoveries in history are rarely the result of intentional, perfectly planned design. They are accidents. Penicillin: a contaminated petri dish. The microwave oven: a melted chocolate bar in an engineer's pocket near a magnetron. X-rays: fluorescent glow from an experiment aimed at something else entirely.

These are not design examples — they are discovery examples. But the principle is the same. The people who made these breakthroughs were deeply skilled in their domains. They had the craft. What they also had was the willingness to follow the unexpected result instead of discarding it. They drifted from their plan. And they had the judgment to recognize when the drift produced something extraordinary.

Zero great breakthroughs started with complete certainty about the outcome.

For fifty years, the economy has organized designers around certainty. Clear briefs. Defined requirements. Measurable outcomes. Stakeholder alignment. Ship something that works and move on. That framing made sense in a world where iteration was expensive — where every design cycle cost weeks of engineering time and real production dollars.

AI fundamentally changes the economics of experimentation. Iteration that took weeks now takes minutes. The cost of trying something unexpected has collapsed to nearly zero. And yet most design teams are still operating as if experimentation is expensive. They are still optimizing for certainty in a world where the cost of uncertainty has evaporated.

This is not an argument against usability. Doors should still open the way people expect. But "design for everyday things" as a mindset — settling for the first solution that works — is a choice, not a constraint. When experimentation is nearly free, there is no excuse for stopping at competent. The discipline now is to keep drifting until you find the version that is not just correct but inevitable.

Test your taste

You have taste you can't fully explain. On the original version of this piece there are six design forks, and the exercise is to pick the one you'd ship — then notice two things: how fast you decide, and how impossible it is to say why. One pair sets a centered, symmetrical, safe layout against an asymmetric, tension-driven one. Most people choose in under a second and then spend a minute failing to justify it.

That gap between knowing and explaining is exactly where taste lives. It's also exactly what AI can't do. If you want to run the forks yourself, they're on the original piece.

Our hallucinating friend

Everyone talks about AI hallucinations as a problem to solve. In domains where correctness is life-or-death — medicine, law, engineering — hallucination is failure. But in creative domains, hallucination is something else entirely. It is drift. And drift is the entire point.

The irony is that the mechanism behind hallucination — the tendency to generate something that departs from the literal training data — is exactly the mechanism that creative work requires. The problem is not that AI hallucinates. The problem is that AI hallucinates without discipline. It drifts without knowing how far to go or when to stop.

Look at what happens when we try to "fix" AI with human feedback. RLHF (Reinforcement Learning from Human Feedback) collects preferences from thousands of annotators and optimizes toward whatever the broadest crowd accepts. The output converges. The diversity collapses. The model gets better at being average. It learns good taste — follow the rules, satisfy the majority — and moves further from great taste with every training round.

AI already knows the rules. Every design convention, every grid system, every typographic scale, every color theory principle — it has absorbed more of these than any human ever will. The capability to follow rules is already there, especially with agentic workflows. What is missing is the mechanism to break them like an artist: with awareness, with intent, with the judgment to know when the drift has landed somewhere extraordinary and it is time to stop.

The data is clear: the better the structured taste layer feeding the model, the better the output. Unconstrained models produce wild variance. Models informed by structured design context — tokens, principles, decisions — produce output that is coherent, intentional, and dramatically closer to what the designer actually wanted. The competitive advantage in AI-augmented design is not which model you use. It is the quality of the structured taste informing it.

Design's Napster moment

AI is fundamentally rewriting the laws of intellectual property in creative work. Design is about to have its Napster moment — the point where the old model of ownership breaks and a new one has to be built.

Right now, AI companies are profiting from your experience and knowledge. Every design system you published, every case study you shared, every Dribbble shot and Behance project — it was consumed by training pipelines that now generate competing work at scale. The people whose judgment and taste made those outputs possible receive nothing.

Music went through this. I lived it. Napster shattered the old distribution model. The industry responded with DRM and lawsuits — and lost. What eventually worked was Spotify: a new system that made sharing the default and built compensation on top. But Spotify got the economics wrong. It centralized value extraction. Artists stream billions of plays and earn pennies. The shape was right. The distribution of value was not.

Design needs the right shape and the right economics, eventually. A system where designers can share their taste — their principles, their decisions, their judgment about which rules to bend — and get compensated when AI systems use it. Not a walled garden. Not DRM for design tokens. An open infrastructure where the people who provide the intelligence receive a share of the value it creates.

What I'm building

So how do you encode taste? How do you capture disciplined drift in a form that AI can actually use? Not a prompt. Not a style guide PDF. Not a Figma library. Something machine-readable that preserves the judgment behind the decisions: which rules to bend, how far to push, and when to stop.

That's what I'm building with Morpius. Capture your design decisions — principles, judgment calls, the why behind every choice — into a structured knowledge graph you own. Connect any AI tool to it over MCP, and your system gets retrieved on demand instead of averaged toward the mean.

Not to replace your taste. To extend it.

Later: license your structured taste to teams who want to design like you, or borrow taste from designers you trust. That's the marketplace — Spotify showed the right shape for creative compensation, it got the economics wrong, and we can do better. It's coming. The memory comes first.

References

1. Sibley, F. — "Aesthetic Concepts" (1959). Philosophical Review 68(4), 421–450.
2. Hume, D. — "Of the Standard of Taste" (1757). Kant, I. — Critique of the Power of Judgment (1790).
3. Wallace et al. — "Diffusion Model Alignment Using Direct Preference Optimization," CVPR 2024.
4. Vendi Score — Friedman & Dieng, arXiv 2210.02410. Diversity measurement in generative models.
5. Google Research — "Introducing NIMA: Neural Image Assessment" (2017); AVA dataset.
6. GenColorBench — 44,464-prompt color binding evaluation across 5 tasks (October 2025).
7. Fleming, A. — "On the Antibacterial Action of Cultures of a Penicillium" (1929). British Journal of Experimental Pathology.
8. Spencer, P. — Raytheon microwave oven development, accidental magnetron discovery (1945).
9. Röntgen, W.C. — "On a New Kind of Rays" (1895). Discovery of X-rays.

This piece first ran on morpius.com/taste in May 2026. If you're wrestling with how to get your own taste into the tools your team uses every day, get in touch.