AI Tools Are Commodities. Taste and Framing Are Not

Last week I reviewed two presentations. Both were made with AI assistance. Both had professional graphics, clear structure, and polished language. They were nearly indistinguishable in technical quality.

But one was forgettable. The other was remarkable.

The difference wasn't the AI tool. They used the same one. It was taste. One creator knew what to keep, what to cut, how to sequence for impact, where to add human touches that AI couldn't conceive. The other accepted whatever AI generated.

This is the future of creative work: AI tools will be commodities. Taste will be the differentiator.

The Commoditization of Tools

Remember when Photoshop skills were a competitive advantage? Or when being good with spreadsheets made you valuable? Tools get democratized. Skills that depend purely on tool proficiency get commoditized.

AI is following the same pattern, just faster:

• Text generation is becoming free and ubiquitous
• Image creation is accessible to anyone with a prompt
• Code generation is eliminating boilerplate work
• Analysis tools are becoming point-and-click simple

When everyone has access to the same powerful tools, having the tool isn't an advantage anymore. The advantage shifts to how you use it, and that's where taste comes in.

What Is Taste, Actually?

Taste is hard to define but easy to recognize. It's the quality that makes some things feel right and others feel off. In creative work, it shows up as:

Selection. From infinite options, choosing what belongs and what doesn't. AI can generate a hundred variations; taste knows which one sings.

Restraint. Knowing when to stop, what to leave out, where less is more. AI has no filter for "too much." Taste provides that filter.

Context sensitivity. Understanding what works for this audience, this moment, this purpose. AI optimizes for generic appeal; taste optimizes for specific resonance.

Coherence. Making many elements feel like they belong together, creating a unified experience rather than a collection of parts.

As I discussed in using ChatGPT as a thinking partner, the quality of your judgment determines the quality of AI-assisted output.

Why AI Can't Replicate Taste

AI is trained on averages: what worked across millions of examples. This makes it excellent at producing "acceptable" output that fits established patterns.

But taste is about exceptions, not averages. It's about knowing when to break rules, when convention is wrong, when the unexpected choice is exactly right. AI predicts the most likely next token; taste sometimes demands the least likely one.

Taste is also deeply contextual. What's tasteful for one audience is gauche for another. What works in one medium fails in another. What's fresh today is cliché tomorrow. AI can learn patterns but struggles with the situated judgment that taste requires.

The Taste Gap in Practice

Look at any AI-generated content and you'll see the taste gap:

AI-generated writing is technically correct but often feels flat. It has no voice, no perspective, no edge. It's the literary equivalent of beige.

AI-generated images can be beautiful but often feel sterile, missing the "wrongness" that makes human art interesting. Everything is too perfect, too smooth.

AI-generated code works but often lacks elegance: it's verbose where it could be concise, generic where it could be specific, conventional where a different pattern would be better.

In each case, AI provides the raw material. Taste transforms that material into something worth caring about.

Developing Taste

Taste isn't innate; it's developed through exposure, practice, and reflection. Here's how to cultivate it:

Consume intentionally. Taste develops from exposure to quality. Read great writing. Look at great design. Experience great products. And don't just consume: analyze. Why does this work? What choices did the creator make? What would have been worse?

Practice selection. When AI gives you options, don't just pick the first acceptable one. Generate many, then ruthlessly curate. The muscle of choosing what belongs is the core of taste.

Study the masters. Find people with taste you admire and study their choices. What do they include? What do they omit? How do they sequence and structure? Reverse-engineer their decisions.

Get feedback. Show your work to people whose taste you trust. Where do they wince? Where do they light up? Taste calibrates through external feedback.

Develop strong opinions. Taste requires having a point of view. "It's fine" is not taste: it's the absence of taste. Practice having strong opinions about what works and what doesn't, even if those opinions evolve.

Taste as Competitive Advantage

When tools are commoditized, taste becomes the moat:

In content: Everyone can generate articles with AI. Those with taste create content people actually want to read.

In design: Everyone can generate images with AI. Those with taste create visuals that genuinely resonate.

In products: Everyone can build features with AI assistance. Those with taste create experiences people love.

In strategy: Everyone can analyze data with AI. Those with taste see what the data doesn't show and make judgment calls that create differentiation.

This connects to what I wrote about in the new digital divide being judgment, taste is judgment applied to creative decisions.

The Human Premium

Here's the opportunity: in a world flooded with AI-generated mediocrity, human taste becomes more valuable, not less.

People can tell the difference between AI-generated generic content and work that has a human sensibility behind it. They might not articulate it as "taste," but they feel it. And they'll pay for it: with attention, money, and loyalty.

The future belongs to people who can collaborate with AI while adding something AI cannot: the distinctly human judgment about what's good, what matters, and what resonates.

Problem framing: taste applied before the prompt

Taste shows up in judging output. It shows up earlier too, in how you state the problem in the first place. That upstream half was previously its own article.

What Is Problem Framing?

Problem framing is the skill of understanding what you're actually trying to solve. It includes:

Defining the real problem. Not the surface symptom, but the underlying issue. "Sales are down" isn't a problem statement. It's a symptom. The problem might be product-market fit, messaging, targeting, or a dozen other things.

Identifying constraints. What's fixed and what's flexible? What resources exist? What must be true for any solution to work?

Understanding success criteria. How will you know if you've solved it? What does "good enough" look like?

Exploring the solution space. What categories of solutions exist? What's been tried? What assumptions limit thinking?

Problem framing happens before you ever talk to AI. It determines what conversations are worth having.

Why Framing Beats Prompting

Several dynamics make framing more valuable:

AI can optimize but can't question. AI is excellent at optimization within frames. Give it a clear problem, and it generates solutions. But it can't step back and ask whether the problem itself is right. That's your job.

Prompts will commoditize. Problems won't. Prompt libraries are everywhere. Best practices get documented and shared. The "secrets" of prompting become common knowledge quickly.

But problem framing is contextual and judgment-intensive. It doesn't reduce to templates. As I discussed in AI tools becoming commodities, the repeatable parts get automated. The judgment parts retain value.

Garbage in, garbage out. The oldest principle in computing still applies: output quality depends on input quality. The most sophisticated prompt engineering can't save you from a poorly framed problem. Frame it wrong, and you're just getting wrong answers faster.

Problem Framing in Practice

Here's how I approach problems before touching AI:

The Five Whys is the first tool. Ask "why" five times to drill from symptoms to root causes. Then ask "why this problem, now?" to understand context and urgency.

Example:

• "Our content isn't performing." Why?
• "Not enough traffic." Why?
• "Low search rankings." Why?
• "Content doesn't match search intent." Why?
• "We're writing what we want to say, not what they want to know." Why now?
• "Competition has improved, and we've stood still."

The real problem isn't "content isn't performing." It's misalignment between content strategy and audience needs, made urgent by competitive pressure. Very different AI conversations follow from this reframe.

Invert the problem. Instead of "how do we increase sales?", ask "what would guarantee we never make another sale?" Then prevent those things.

Inversion often reveals obvious blockers that direct approaches miss. It's also easier to identify what causes failure than what causes success.

Question the goal itself. Is this the right goal? "We need to increase website traffic" might be wrong if the real objective is revenue. Sometimes solving the stated problem doesn't solve the actual need.

Define what "solved" means before seeking solutions. Specify what success looks like. Quantify where possible. "Better content" is vague; "content that ranks page-one for target keywords" is actionable.

This connects to what I track in my decision journal. Clear success criteria make better decisions possible.

The Framing → Prompting Workflow

Here's how framing and prompting work together:

Step 1: Frame the problem (no AI)

• What's the real issue?
• What constraints exist?
• What does success look like?
• What categories of solutions are possible?

Step 2: Explore with AI

• Share your framing
• Ask for challenges to your assumptions
• Request alternative frames
• Generate options within your frame

Step 3: Evaluate and refine

• Does AI's output address the real problem?
• What's missing or wrong?
• Does the frame need adjustment?
• Loop back to Step 1 if needed

Most people jump straight to Step 2, optimizing prompts for a problem they haven't properly understood. The work in Step 1 makes everything else more effective.

Developing Problem Framing Skills

How do you get better at framing? Unlike prompting, there's no quick tutorial. But there are practices:

Study frameworks. Mental models like first-principles thinking, systems thinking, and constraint analysis provide scaffolding for framing. Build a toolkit of frameworks to apply.

Practice on real problems. Take a current challenge and spend 30 minutes framing it before seeking solutions. Write out the problem statement, constraints, success criteria. Force yourself to articulate what you're actually solving.

Review past decisions. Look back at problems you "solved" that didn't stay solved. Often the issue was framing: you solved the wrong problem. These failures teach framing better than any course.

Seek diverse perspectives. Different people frame problems differently. Expose yourself to how others think about problems, especially those from different fields or backgrounds.

Invest in What Will Not Commoditize

Stop chasing the latest AI tools. Everyone will have them soon enough.

Instead, invest in taste. Develop your sensibility. Cultivate your judgment about quality. Practice the art of curation and selection.

The tools will keep changing. The need for human taste will not. That's where the lasting competitive advantage lives. Not in what AI you use, but in the taste you bring to using it.

Want to develop better judgment and taste? Start by understanding how to learn effectively, taste develops through deliberate study and practice.