We are living in an era of ideas. Everyone has one. AI can now generate a hundred more ideas before lunch. Consultants, coaches, and well-meaning contacts add still more ideas, often without ever testing whether any of it works in practice. For a business owner trying to actually move forward, the result is often overload rather than inspiration.
I want to talk about what that overload actually does to people, psychologically, and about a specific memory that has stayed with me for years.
Being Called Don Quixote
Early on, I worked with business coaches through SCORE. Some of their ideas were useful. Others were decades old, dated from well before I was born, and offered with the confidence that they would still work unchanged in a completely different market.
When I didn't adopt every idea they gave me, I was labeled resistant.
At one point, a coach told me directly that I would fail in trying to change the U.S. healthcare system and compared me to Don Quixote.
For anyone unfamiliar with the reference, Don Quixote is the title character of Miguel de Cervantes's seventeenth-century novel, a man who reads so many chivalric romances that he loses touch with reality and sets out on quixotic, doomed adventures, most famously charging at windmills he believes to be giants. "Tilting at windmills" has since become shorthand for pursuing an idealistic goal that the speaker considers delusional or unwinnable. It is rarely a compliment. It is a way of saying: you are not seeing reality clearly, and your effort will be wasted.
I want to be precise about what was actually happening in that exchange, because I think it illustrates a pattern that shows up constantly in business advice, and increasingly in AI-generated advice as well.
I was not rejecting ideas because I didn't understand them. I was declining to implement advice that didn't fit my situation, my market, or my evidence, while continuing to test the ideas that did. That is a different thing from resistance. It is filtering. But when someone has a large volume of ideas to offer and expects all of them to be adopted, filtering can look, to that person, like defiance.
What Actually Happened Next
I did not become a cautionary tale.
I went on to win the Comcast RISE 2024 grant. I presented my own findings at WIDIC. I worked with Cuyuna Regional Medical Center, Mayo Clinic Jacksonville, and several U.S. physician-owned private practices, and they changed their approach to spinal CSF leak care based on the feedback I brought them. And during the same stretch of years I was being told I was tilting at windmills, I was also working through my own hyperthyroidism, tracking it, testing it, and bringing it back down over roughly six agonizing months, largely by treating my own body the way I treat a business problem: as something to be measured, hypothesized about, and tested rather than simply labeled.
None of that makes the coach who invoked Don Quixote a bad person. I think it makes a useful point about a very common failure mode in business advice: mistaking "did not adopt my idea" for "will not succeed."
Why More Ideas Can Make People Worse Off, Not Better
This is where cognitive load theory becomes useful. Cognitive load theory, developed by the educational psychologist John Sweller in the 1980s, describes how working memory has a limited capacity for processing new information at any given time. When the amount of incoming information exceeds that capacity, comprehension and decision-making both suffer, not because the person is incapable, but because the system is overloaded.
A business owner facing a wave of advice, from coaches, from consultants, and now from AI tools that can produce dozens of suggestions instantly, is often experiencing exactly this. The problem is rarely a shortage of ideas. It is an oversupply, arriving faster than any one person can evaluate, prioritize, or test.
This connects to what the psychologist Barry Schwartz called the paradox of choice: beyond a certain point, more options do not improve decision-making. They increase anxiety, second-guessing, and the likelihood that no decision gets made at all. Research on decision fatigue, associated with Roy Baumeister's work on ego depletion, describes something related: the quality of decisions degrades as the number of decisions a person has to make in a given stretch of time increases. Every additional idea someone hands you is not a gift. It is a withdrawal from a limited account.
There is also a cultural pattern worth naming honestly. In some professional cultures I have worked across, a great deal of time is spent talking through ideas hypothetically with contacts and colleagues, crowdsourcing opinions and possibilities, without ever moving into the stage of actually testing anything in the market. Irving Janis described a related phenomenon in his research on groupthink: groups can generate agreement and even enthusiasm around an idea while actively avoiding the step that would expose whether the idea is any good. Talking about an idea, especially in a comfortable, agreeable group, can start to substitute for testing it.
Why Breaking Things Down Is the Actual Skill
I spent years as an instructor before I became a consultant, and the instinct that experience built in me is not "add more information." It is "break this down into a piece someone can actually use right now."
This draws on some well-established ideas in cognitive and developmental psychology. George Miller's classic 1956 paper on the limits of working memory, often summarized as "the magical number seven, plus or minus two," described how much information a person can hold and process at once, and it remains a foundational argument for chunking, breaking large amounts of information into smaller, manageable units rather than delivering it all at once. Lev Vygotsky's concept of the zone of proximal development, and the related idea of scaffolding developed by Wood, Bruner, and Ross in the 1970s, describes how a learner progresses fastest not by being given everything at once, but by being given the next appropriately sized piece, with support that can be gradually withdrawn as competence grows.
When I work with a business owner who is overwhelmed by ideas, and increasingly overwhelmed by an AI tool generating even more of them, my job is rarely to add another idea to the pile. It is to identify the one thing that is actually testable right now, help them test it, and use what the test shows before moving to the next piece. That is the opposite of what the coach who called me Don Quixote was doing. He was adding volume. I try to remove it.
This also borrows directly from the logic of the Lean Startup methodology that Eric Ries popularized: build, measure, learn, in small cycles, rather than debating a large plan hypothetically until it feels complete. Ideas that have not been tested are not knowledge. They are guesses, however confidently they are delivered.
What My Own Body Taught Me About Testing Under Real Stakes
I want to describe something I don't usually put into professional writing, because I think it is directly relevant here.
My body reacts, sometimes severely, to things most people tolerate without a second thought: a supplement gummy, a new food, the smallest possible amount. When it happens, it does not happen mildly. I have been on the floor, feeling like my brain was swelling inside my skull, my heart racing into tachycardia, my fingers too sweaty to operate my own phone, unable to do anything but scream for help.
And then, hours later, I am expected to function again. I track my vitals. I schedule the windows in which I test something new for times when I am not working, specifically so that if it goes wrong, it does not take my business down with it. I try to identify, from the specific sensations I experienced, what might explain the reaction. I form a new hypothesis. I test again, carefully, deliberately, and always with a plan for what happens if it goes wrong.
There is a body of research on stress inoculation, developed by the psychologist Donald Meichenbaum, that describes how controlled, repeated exposure to manageable stress, followed by recovery, can build a person's capacity to function under future stress rather than simply wearing them down. There is also research on post-traumatic growth, from psychologists Richard Tedeschi and Lawrence Calhoun, describing how people who go through severe adversity sometimes emerge with genuinely increased psychological resilience and a clearer sense of what actually matters, not as a guarantee, but as a documented outcome for a meaningful number of people. And Albert Bandura's research on self-efficacy describes how a person's confidence in their own ability to handle a future challenge is built primarily through direct, lived experience of handling past ones, what he called mastery experience, more than through being told they can do it.
I have had a great deal of practice, under circumstances with real physiological stakes, in doing exactly what I ask overwhelmed business owners to do: notice what is actually happening, form a hypothesis instead of a panic response, test it carefully, and adjust based on what the test shows rather than on how frightening the last attempt felt.
The Actual Point
None of this is a claim that ideas are bad, or that advice has no value, or that AI-generated suggestions are worthless. Some of the coaching I received through SCORE was genuinely useful, and some AI-generated ideas are genuinely good starting points.
The point is that volume of ideas is not the same thing as progress, and the ability to sit with someone in overwhelm, help them filter the noise down to the one testable next step, and walk with them through the discomfort of actually trying it, is a distinct skill from generating more ideas for them to sort through on their own. It is a skill built the same way most durable skills are built: through repeated, deliberate practice under conditions where the outcome genuinely mattered.
I was called Don Quixote for declining to accept every idea I was handed. I would rather be the person who tests one thing carefully than the person who talks about a hundred things and tests none of them.
