✪ How to Make 20 Years of Experience Visible Through Your Crystallized Intelligence: The Wise Winnowing Method - #65
The AI-Age curation method that builds expert authority through judgment, not writing skills.
I write this newsletter for new, emerging experts, consultants and advisors who are looking to increase their visibility, authority and trust in the eyes of their prospective customers: people who want to make a business out of their personal know-how even if they’re not great writers and may not have digital-age credentials.
The strategy I have been using for 25+ years, is curation. That means helping my readers find what they should pay attention to by: a) becoming a discovery agent for them, b) highlighting and contextualizing other people’s great work, c) creating collections, libraries, catalogs that repeatedly support the goals of my readers.
In this issue:
Experts with decades of experience have a hidden superpower they’re not using: Crystallized Intelligence: the ability to instantly recognize quality, spot patterns, and make informed judgments. But they’re wasting it by trying to become AI-powered writers instead of AI-augmented curators. The real opportunity isn’t writing polished slop. It’s in something quite different and which very few do: proactive over-collection + expert selection, contextualization and insight to build lasting knowledge infrastructures.
The Expert’s Trap
Most experienced consultants/advisors aren’t natural writers. So when building online presence, they fall into the AI content factory: AI suggests topics, then it drafts outlines, then it generates polished shiny prose, then the expert intervenes to edit it, clean it up and try to make it sound as if it was written by a human.
Result: fake-sounding, untrustworthy content, identical to everyone else’s plastic-smelling though shiny, AI slop.
But you don’t need to become a writer. You need to leverage what you have naturally developed through many years of experience: attention to details, pattern recognition, warning scars, unique evaluation criteria (taste).
1) Your Hidden Asset: Crystallized Intelligence
Most experienced consultants I know aren’t natural writers. They’re great at what they do. They understand their field deeply. But sitting down to write? That’s different.
So when building an online presence, they ask AI to help. AI suggests topics, drafts outlines, polishes prose. They clean it up and publish.
The result? Content that sounds like everyone else’s. Polished, professional, forgettable. Worse, it feels fake. Readers sense there’s no real person behind the words.
From what I’ve observed over twenty years studying curation, there’s an alternative approach worth considering.
There are two types of intelligence. Understanding the difference changes how you might use AI.
a) Fluid intelligence is raw brain power. The ability to generate completely new ideas from scratch. To solve problems you’ve never seen by thinking from first principles. It peaks around age twenty, then declines. (think: videomakers, animators, rappers, writers, coders, startuppers, musicians).
b) Crystallized intelligence is your accumulated library of patterns. Every situation you’ve encountered. Every project that worked or failed. Every client problem you’ve solved. All stored as recognizable patterns. (think: doctors, lawyers, architects, mechanics, plumbers, film critics, art directors, curators).
Crystallized intelligence doesn’t decline. It grows. By fifty or sixty, you’ve seen thousands of situations play out. You know what works and what doesn’t. You can look at a new problem and instantly recognize it as similar to something you’ve encountered before.
When someone shows you ten approaches to a problem, your brain pattern-matches against stored experience. Instantly you intuitively know which ones will work.
This is taste, intuition, having a good eye. It’s crystallized intelligence from years of experience. Most experts overlook this key strength while trying to bring new ideas and concepts to the table.
But new ideas are abundant now. AI generates hundreds instantly. What’s scarce is knowing which ideas actually work in practice. Which of a hundred options will solve the problem for a specific situation. That requires pattern recognition that only comes from lived experience.
2) Use AI as an Options Generator, Not as a Writing Assistant
When most experts start using AI to build their online presence, they use it like a writing assistant. They describe what they want to say, AI drafts the article, they edit to sound more like themselves, and publish.
I’ve tried this approach extensively. The output is always polished and well-structured. But it never demonstrates actual expertise. Anyone could have written it. There’s no visible proof you’ve spent twenty years in your field.
The fundamental issue is that AI-as-writer produces the same content everyone else is producing. It’s generic advice dressed up in your personal anecdotes. Readers can’t tell if you actually know what you’re talking about or if you’re just good at prompting ChatGPT.
What I’ve discovered through my work with curation is a different approach entirely. It’s very simple yet almost no-one I know uses it. Instead of having AI write for you, have AI generate options for you to evaluate.
The difference is significant. When AI writes, you’re compensating for not being a natural writer. When AI generates options, you’re leveraging what you’re actually good at: recognizing quality based on professional experience.
Let’s say you want to help clients identify the right pricing strategy. Instead of asking AI to write an article about pricing strategies, ask it to generate fifty different pricing approaches used across various industries.
AI gives you the list in minutes. Now you apply your crystallized intelligence. You scan through all fifty. Your pattern-matching brain immediately eliminates forty of them. Some are too complex for your client’s context. Some won’t work in their industry. Some sound good in theory but you know they don’t work in practice.
You’re left with ten that could work. You examine those more carefully. Five are variations of the same underlying approach. You pick the best version. Three work only in specific situations. You note when to use each. Two offer genuinely different approaches.
Now you have something valuable.
You take those final selections and remix them. You grab the pricing structure from one, the communication strategy from another, the implementation timeline from a third. You add context from your experience about why each element matters and when it breaks down.
What you have now isn’t an AI-written article about pricing. It’s a curated comparison of pricing approaches with clear criteria for when to use each one, informed by your pattern recognition from years of client work.
This demonstrates visible expertise. Anyone reading can see you’ve actually evaluated these options. You’re showing your judgment in action, not repeating generic advice.
The shift is from “AI writes, you edit” to “AI over-collects, you curate with expert judgment.” One produces content that looks like everyone else’s. The other produces something only you could create because only you have your specific accumulated experience.
3) Pro-Active Curation: The Wise Winnowing Method
The approach I’ve been describing has a name. Jakob Nielsen, a usability researcher, calls it “wise winnowing.” The concept is simple but the implications are significant for experts trying to build credibility online.
The method has four steps: over-collection, expert selection, remixing, curation.
Over-collection means deliberately asking AI to generate far more options than you need. If you need five good examples, ask for fifty. If you want three solid frameworks, generate thirty. The goal is volume, not quality. You’re creating raw material for your expert judgment to work with.
This is counter-intuitive. We’re trained to be efficient, to ask for exactly what we need. But the power of this approach comes from abundance. The more options you have, the better your final selection will be. You’re casting a wide net specifically so you can choose the absolute best.
Expert selection is where your crystallized intelligence does its work. You look at those fifty options and your pattern-matching brain starts operating automatically. You eliminate the obviously wrong ones in seconds. You group similar approaches together. You recognize which ones sound good but won’t work in practice. You identify the genuinely different alternatives worth considering.
This happens much faster than you’d think. When you have real expertise in a field, you don’t need to carefully analyze each option. Your accumulated experience lets you evaluate quality almost instantly. What takes AI seconds to generate takes you minutes to evaluate, not hours.
The third step is remix. This is where you combine the best elements from multiple options into something new. You take the structure from one approach, the implementation details from another, the specific tactic from a third.
The final step is pure curation. You add your own insight, context, and experience.
What makes this powerful is that you end up with something neither you nor AI could have created alone. AI can’t do the expert selection because it has no lived experience. You couldn’t have generated all the options because that’s not where your strength lies. Together, the output is unique.
I’ve used this method to create short films, design innovative television news screens, prepare solid DJ sets (see examples), as well as to develop many resource collections. Each time, the process is the same: over-collect with AI, winnow with experience, remix and curate with insight.
The result looks and feels different from typical AI-generated content because you can feel the human expertise and its ability to select based on lived experience, not just an ability to collect and organize information.
4) Making Your Selection Process Visible
The wise winnowing method creates valuable output, but there’s an additional step that transforms it from good content into genuine trust-building material. You need to show how you do what I just described. How you filter, what criteria you use, where you search, how you organize and arrange.
When you publish your final selection of three frameworks or ten key resources, the natural instinct is to present your choices as the finished product. That’s what 90% of experts attempting to curate do. The reality is that your selection is only one-third of the value. The other two thirds are made up by a) providing context and b) explaining how you make your choices.
This is what researchers call paradata. It’s data about your process. Why you picked what you picked. What criteria you used. What you rejected and why. This transparency separates expert curation from simple list-making.
Let’s say you’re creating a decision framework for choosing between consulting models. You asked AI to generate thirty different frameworks used across industries. You winnowed down to three that actually work for independent consultants. Most people would publish those three frameworks and stop there.
But when you also show the other thirty you evaluated, the criteria you used to narrow them down, and the specific reasons you eliminated twenty-seven of them, readers see your expertise in action.
This matters because AI can generate conclusions but it can’t show genuine expert judgment. Anyone can ask ChatGPT to generate consulting frameworks. What they can’t replicate is knowing when to apply those twenty frameworks in real situations, with real clients, recognizing which ones sound super-cool but create confusion, identifying which ones work only for large firms, and knowing which ones collapse when certain circumstances change.
Your selection criteria should be explicit. Not vague statements like “practical” but specific observations from experience. Like: “This framework requires clients to have data they rarely have. This one works for technical consulting but fails in creative fields. This approach looks comprehensive but I’ve watched consultants get lost in its complexity.”
The rejected options matter too. When you explicitly say “I didn’t include this popular framework because I’ve seen it fail repeatedly when...” you demonstrate your ability to evaluate based on real experience, not sound logic and general principles. This requires more time, but it’s what makes your curated resource trustworthy.
People today want to understand how you think, not just what you’ve put together.
5) What This Builds: Knowledge Infrastructure
What you create through “wise winnowing” curation isn’t typical content. It’s something that lasts longer and serves a different purpose. I think of these as knowledge infrastructure pieces rather than content pieces.
Content gets read once and forgotten. A blog post about productivity tips gets consumed and disappears from memory within days.
Knowledge infrastructure gets referenced repeatedly. A well-curated comparison framework gets bookmarked, shared, and returned to when decisions need to be made.
The distinction matters for how you think about what you’re building. Content competes for attention in an endless stream. Infrastructure becomes a destination. Content decays rapidly as new information emerges. Infrastructure can be updated and maintained, growing more valuable over time.
Here some examples of knowledge infrastructures:
Decision frameworks help people choose between alternatives. When you’ve curated ten different approaches to solving a common problem, organized them by context, and explained when to use each one, you’ve created something people return to whenever they face that decision.
Pattern libraries collect proven models or methods in a field. Instead of inventing new frameworks, you gather existing ones that actually work, organize them by use case, and explain the situations where each applies. This demonstrates comprehensive knowledge of your field.
Learning paths sequence existing resources into a logical progression. You’re not creating new teaching materials. You’re curating the best existing articles, videos, and tools, then organizing them so someone can learn systematically. The value is in the curation and sequencing based on how you’ve seen people actually learn.
Newsradars track what’s emerging in your field. You monitor dozens of sources, select the five to ten things that actually matter each week, and explain why they’re significant. Over time, this builds a reputation for discernment and current awareness.
Comprehensive views of idea-spaces map out all the major approaches, philosophies, or schools of thought in an area. You’re showing the landscape so people can understand where different ideas fit and how they relate to each other.
Each shows you can evaluate, organize, and synthesize information that already exists. Each becomes more valuable as it gets used and updated over time.
6) Real-World Examples
Wise winnowing curation at work in the real world and in my personal experience.
.
a) Pretty Lights by Derek Vincent Smith
For his 2013 album “A Color Map of the Sun,” Derek Vincent Smith created a completely original approach to sampling. Instead of digging through existing vinyl records, he:
Spent two years recording session musicians in studios across Brooklyn, New Orleans, and Denver
Created 20+ hours of original music recorded on analog tape across multiple genres (rock, soul, funk)
Pressed all these recordings to vinyl to create his own personal “crate” of records
Then sampled from this massive collection, selecting and remixing the best elements
Here’s the full story.
b) My own Rhythms video thesis at SFSU
For my final BFA thesis at the Center for Experimental and Interdisciplinary Arts at San Francisco State University (1981) I chose to go for a unique curation project: an audio-visual journey into different moods / atmospheres. To create it, I ended up leveraging the two areas I had been exposed the most in my first 20 years of my life: music and images (art history, photography and fashion design). That’s what I did:
Over-shot tens of unscripted scenes from real-life in 16mm film.
Created a seamless compilation-soundtrack by editing together 20+ instrumental tracks from different authors and musical genres (already using crystallized intelligence from my young DJ career).
Finally mixed and matched a tiny selection of clips from the abundant footage library I had shot to the musical soundtrack as to create an audio-visual journey into different moods and emotions.
That unique video crowned my good undergraduate work allowing to me get a 110 Magna cum Laude final grade.
c) DJ sets preparation process
During my 4-year stay on the beautiful tiny Holbox island (2020-2024), I used a wise winnowing process to prepare each and every one of the DJ sets I would perform at various beach resorts and night clubs. (Here a short selection of recorded sets).
Paid for access to an online distribution platform specializing in my music genres and which made available hundreds of new tracks every single day, alongside DJ-made recommended selections.
Pre-listened to 400-800 tracks in sessions of 2-5 hours and downloaded all tracks that fit my criteria at a ratio of apx 1:20 (one good track every 20 or so listened to).
Sifted through the most recent DJ set playlist I had already created and extracted the 15-20 tracks that still sounded fresh and non-obvious for my audience and style.
Curated these with the 20-30 new ones I had picked and sorted them by power / energy and tempo to create a base playlist flow going from soft and chill to maximum power and energy through the course of 3-3.5 hours.
Performed the DJ set using the curated playlist as a reference canvas from which I could break free, steer, branch out or improvise. Thus it provided a strong base reference, more free mental space to be used for creative improvisation during the set.
d) Interface design for national TV news
In 1985, thanks to a previous collaboration with the late designer Piero Gratton, I was contracted by Italy RAI TG2, under the direction of Antonio Ghirelli, to review the general design of the news inside all broadcasts. Instead of starting to sketch and design the new screens, I took an unorthodox path:
I hired an assistant researcher / photographer and charged her with accessing and photographing all of Europe news TV channels, from BBC to ARD, ZDF, to France Antenne 2 and Austria ORF and more (12 channels in total).
The strategy was to capture the key 8-10 typical “screens layouts” for each channel, as to have a very wide collection of real-world examples to pick, innovate from and remix selectively.
We ended up with over 150 sample TV screen designs from which to work.
At the time there was no Internet, no screenshots, and no easy access to other foreign TV channels. You had to actually photograph a television monitor during a news broadcast or request a recorded videotape from a friend living abroad.
All the photos were then glued in an orderly fashion across a number of cardboard panels, as to have the opportunity to view at-a-glance all of the use-cases / screen layouts.
From there the written specifications for each key screen was defined, by taking and improving on the best from the existing options as well as by remixing individual details from multiple sources.
Unfortunately the slick new interface design for RAI TG2 was never implemented for political reasons. The workers union actively resisted the adoption of my information design plan, as it had not been created by the permanent RAI internal graphic design staff.
The Universal Pattern Across All Four Examples:
Deliberate Over-Collection: 5-100x more options than needed
Expert Crystallized Intelligence: Years of pattern recognition enabling instant quality assessment
Ruthless Winnowing: Eliminate 80-95% systematically
Remix/Synthesis: Combine best elements into something new
Lasting Infrastructure: Results become reference points, not disposable content
Every one of these example demonstrates that the value isn’t in generating options (music samples, film footage, possible good tracks) - the value is in the selecting and arranging of your collected resources based on your intuitive taste refined over many years of actual experience.
7) Practical Implementation
To try this approach, here’s how I’d suggest starting based on what I’ve learned building these resources over the years.
First, pick one curated infrastructure type. Don’t try to build multiple things at once. Choose based on what you already do naturally. If you’re constantly evaluating tools, start with a curated directory. If you solve similar problems repeatedly, build a decision framework. If you consume a lot of content in your field, create a learning path or a pattern library.
Second, have AI over-collect options. Ask it to generate or find one hundred examples, approaches, tools, or frameworks related to your chosen topic. The specific number matters less than the principle: you want far more options than you’ll ultimately include. This abundance lets your expert judgment shine.
Third, apply your winnowing. Go through everything AI collected or generated and eliminate what you immediately feel doesn’t work. Use your actual experience. “This won’t scale in your client’s context. This sounds good but fails in practice. This is too complex for most people to implement.” You should be able to cut eighty to ninety percent quickly.
Fourth, make your criteria transparent. Don’t just publish your final selections. Explain what you looked for and why. Show examples of what you rejected and the reasoning behind those decisions. That’s how your expertise becomes visible.
Fifth, remix and curate. Take the best elements from your top selections and combine them. Then provide context specifying in which situations and for what kind of problems this is relevant. Add specific observations from your experience about when each works and when it doesn’t. Provide details and observations that only come from having actually used or implemented these approaches.
Sixth, publish it as a reusable resource (knowledge infrastructure), not as an article or blog post. Format it so people can reference it repeatedly. Make it easy to bookmark, share, and return to when they need it.
8) How It Connects To the Rest
And if you got fascinated by this topic, feel like digging deeper and understand more, here is how the concepts I have introduced here, directly connect to the other work I have recently published.
a) Crystallized intelligence is a curator fundamental asset.
.
b) Knowledge infrastructures create orientation systems based on your extended experience.
.
c) Seven curation formats that deploy crystallized intelligence as knowledge infrastructures.
References:
This article was 100% inspired by the work of Jakob Nielsen who first explored - at least in my eyes - the concept of “wise winnowing” in his 2023 article:
“Generative AI Enhances Old Users’ Intellectual Performance Through Wise Winnowing”.
His original essay and the two videos - on the same topic - he has recently published, all focus on highlighting how senior professionals can exploit AI’s power to their own benefit while leveraging their personal experience and judgement developed over many years of work.
I highly recommend his work, his Substack newsletter and the books and great research he has been publishing for over two decades.
.
AI-generated music video, scripted and produced by Jakob Nielsen on how senior professionals can remain creative with AI.
Duration: 2’:57”
.
AI-generated video by Dr. Nielsen where a senior virtual expert explains what is crystallized intelligence and why it offers a unique opportunity to those who have developed an extended experience doing something.
Duration: 6’:37”
.
Pretty Lights - The Making of a Color Map of the Sun
(making music by overgenerating and overcollecting and then wise winnowing)
Duration: 26’:26’
In a networked culture like the one we live in now, authority is no longer built by controlling access to information or having new ideas, but rather by developing the skill of observing and highlighting what is most relevant.
Cover image: The undersigned spinning records at club Salma, Holbox island, February 16th, 2022.
I have been highlighting what to pay attention to for experts, indie entrepreneurs and communication professionals for 25+ years.
What drives me is to learn, help others and improve the world around me through the communication arts.
I enjoy understanding and sharing what I discover.
I love to create lasting information resources that can be useful to others.
If you feel we share interests and values please consider supporting this work (like, recommend, subscribe).
I’d love to keep on searching and highlighting what matters to those who want to help others grow.
Follow a path with a heart.
From Koh Samui (TH)
Robin Good
This nails something a lot of “AI thought leadership” misses. Authority isn’t in how well you can make AI talk. It’s in what you don’t choose. Judgment. Taste. Scars. The ability to look at 50 options and quietly say, “Most of these will fail in the real world.” That’s trust. And AI finally gives experienced people a way to show it without pretending they’re writers.
Thank you