Sign in Subscribe

The Façade of Quality

The Façade of Quality

We are living in the age of the “perfect” first draft. With a simple prompt, an AI co-pilots can generate a business plan, a marketing campaign, a block of code, or an entire essay that is, on its surface, flawless. It is grammatically correct, logically structured, and confidently presented. It has a façade of quality. Yet this seductive illusion is one of the greatest threats to genuine innovation and deep work.

This phenomenon is the professional equivalent of the student who gets a B by flawlessly reproducing the lecture notes. Their work is correct, but it’s a performance of knowledge, not an act of understanding. They’ve built an appealing but ultimately shallow work. In the age of AI, the world is now awash in appealing but shallow work. The market for routine, well-posed competence is collapsing, because the machine can produce it infinitely nearly for free.

Deep value, the kind that transforms industries and solves mad science problems, has never been about surface-level perfection. It’s messy, effortful, and defined by constructive failure. I’ll call this deep work, and it requires deep humans.

I often tell my employees that I will happily name them a "co-founder" of one of my companies if their engagement with a project mirrors a sentiment from Carl Jung: "The meeting of two personalities is like the contact of two chemical substances: if there is any reaction, both are transformed." Deep work is a transformative reaction. When a deep human engages with an ill-posed problem, they don't just execute a task; they are changed by it, and the problem itself is changed in turn. My company is forever transformed, and they are its new cofounder.

This is the fundamental difference between the machine's autonomous output and the contribution of a human (or better still, a cyborg). An LLM doesn't wrestle with a problem; it generates a high-probability sequence of tokens based on its training data. It has no skin in the game. A deep human, a fanatic, engages with a problem with their full being. They bring their unique life history, their hard-won subjective utility, and their entire portfolio of meta-learning skills to the task. They embrace productive friction, using the problem not as a task to be completed but as a crucible in which to forge a new understanding.

This is why the most common use cases for AI—automating our "boring work"—are so dangerous. They promise to free us for deep work, but they actually train us out of the very traits required to do it. They encourage us to become passive editors of machine-generated mediocrity, smoothing the edges of a flawless, shallow façade. We risk becoming a generation of professional plasterers, skilled at hiding the cracks but having forgotten how to build a solid foundation.

How do we cultivate the capacity for deep work in ourselves and our organizations? Reject the façade of quality. Stop rewarding pretty, shallow answers and start incentivizing a messy, exploratory process. Be a leader who models the "Writer's Room" protocol, where the goal isn't to find the right idea quickly but to engage in a collaborative struggle that deepens everyone's understanding. It is facilitating this process, hybrid collective intelligence, in which AI will truly shine.

The future will not be built by those who are best at prompting a machine to generate a perfect first draft. It will be built by the deep humans who have the courage to expose their ideas to the deepest scrutiny in search of its hidden flaws and weak assumptions. They will engage in the hard, transformative labor of building something deep despite the flaws. “If you can’t build it, you don’t understand it.”

💡
Follow me on LinkedIn or join my growing Bluesky! Or even..hey whats this...Instagram? (Yes, from sales reps and HR professionals to lifestyle influences and Twitter castaways, I'll push science anywhere.)

Research Roundup

Dada Science

Our machines are brilliant Dadaists that can generate research ideas that experts judge as more novel than their own. The problem? The ideas are shallow.

Instead of just publishing a paper on “creative” LLMs, the researchers went deeper: they had experts actually execute both AI- and human-generated ideas. Once the rubber met the road, the AI's "brilliant" concepts crumbled. All scores for “novelty, excitement, effectiveness, and overall” quality dropped after facing the messy reality of execution, but the LLM’s score dropped much more.

Novelty is a quality of information. Innovation is a practice.

An LLM generates a well-formed sentence; a human wrestles with a messy problem. The LLM has been fine-tuned to produce the appealing, not to bring concepts into reality. For now, they can't navigate the thousand dead ends that constitute real discovery as well as humans or RLs.

The future isn't autonomous AI ideation. It's hybrid collective intelligence—the synergy between the AI's ability to generate infinite sparks and the deep human's capacity to discern which of them might actually catch fire. The machine can propose a thousand paths; only a human can walk one.

Second Hand News (LLM Ed.)

Why “Google” a question when you can “Claude” an answer? Surprisingly, depth.

A series of experiments compared people learning about a topic using a traditional web search versus an LLM. The “core information…was the same”, but participants who learned from the LLM “developed shallower knowledge”.

Web search, for all its limitations, demands “active engagement”. You have to sift through links, evaluate sources, and synthesize disparate pieces of information yourself. It's a process of productive friction. LLMs, in contrast, deliver a pre-digested synthesis, encouraging a “passive consumption” that creates an illusion of understanding without the deep cognitive work required to build it.

The downstream consequences are even more telling. When asked to create advice based on what they learned, the LLM group felt “less invested”, and their advice was “less original, less detailed, and ultimately, less likely to be adopted by others”. They didn't just learn less; they created less value.

This is the "GPT is the new GPS" problem I wrote about recently. Too many of us are using these powerful tools not to augment our thinking but to automate it. We're choosing the shallow path of frictionless answers over the deep path of genuine learning.

The question isn’t whether search or LLM is more intelligent. The question is which produces the greatest hybrid collective intelligence. LLMs could, but they are trained not too…and unfortunately so are we.

Rather than abandon these tools, we need to fundamentally change how we use them: a starting point, not an endpoint, a sparring partner, not a ghostwriter. Be among those doing the hard, deep work…and maybe even visit a library.

The Perfection-Performances Paradox

Why am I so against gamification and poorly designed incentives? Let’s look at the difference between excellence vs perfection.

Let’s call “excellencism” a striving for “high and realistic standards” while perfectionism pursues “high and unrealistic standards”. In a series of experiences, perfectionists held “higher creative self-efficacy and creative personal identity” than those striving for excellence and yet “scored lower on objective indicators of creative abilities”.

On creativity assessments, “participants pursuing excellence tended to generate more answers and more original ones compared with those pursuing perfection”.

Why? Despite their self-perception perfectionists showed a strong negative association with “openness to experience”. Excellence, not perfectionism, was built on openness and drove creativity, as well as performance on “divergent thinking and associative tasks”.

We train perfectionism into elite performers by enforcing bullshit incentives in school and work that punish exploration and celebrate shallow performance.

Perfectionism: caring more about keeping your job than doing it well.

💡
<<Support my work: book a keynote or briefing!>> Want to support my work but don't need a keynote from a mad scientist? Become a paid subscriber to this newsletter and recommend to friends!

SciFi, Fantasy, & Me

I know I read The Goblin Emperor some time ago, but as I kept seeing it pop up on “Best Fantasy of the Century” lists I was left wondering why I couldn’t remember anything about it. If pressed I would have said, “A kindly social outcast becomes an unexpected emperor.”

I relistened to it last week, and the reason I couldn’t remember more than that is because that is just about the entirety of the story. There are many specific details about him being an outcast and about him being kind and about him being a better emperor for those very reasons.

And yet, I also see why many cherish The Goblin Emperor. As far as cozy fantasy goes, this one has a counterpoint, revealed near the end, that that gives it real sociopolitical depth even as it celebrates the simple heroism of kindness in its (not particularly) gobliny protagonist.

Stage & Screen

💡
If your company, university, or conference just happen to be in one of the above locations and want the "best keynote I've ever heard" (shockingly spoken by multiple audiences last year)?

Vivienne L'Ecuyer Ming

Follow more of my work at
Socos Labs The Human Trust
Dionysus Health Optoceutics
RFK Human Rights GenderCool
Crisis Venture Studios Inclusion Impact Index
Neurotech Collider Hub at UC Berkeley UCL Business School of Global Health