Summary: In which I explore why tinkering—messy, creative, often undervalued and overlooked—is not only a valid way to approach teaching, but perhaps one of the most honest.
My friend Josh Brake recently wrote a Substack post (Don’t tinker with AI in the classroom) warning educators against “tinkering” with generative AI tools in the classroom, especially in K–12 contexts. He advocates for careful experimentation instead.
As it happens, I’m a fan of tinkering, both the act and the idea—which made me wonder why Josh resists the term. While I respect his perspective and share his concerns about careless implementation, I want to make the case for tinkering, especially in light of the complex realities of teaching.
Educational contexts are messy. Just consider the myriad variables educators have to contend with. First are the learners, each of whom brings their unique talents, passions and histories into the classroom. Moreover, classrooms bring their own physical, technological, and architectural constraints into the mix. Further these classrooms are embedded within broader administrative, policy, economic, social, and cultural frames which support or constrain what educators can and cannot do. Finally, teachers are not homogeneous. They bring diverse backgrounds, knowledge, and experiences to the process.
None of these elements are static, they are continually changing and evolving, due to factors internal and external (many of which are invisible or only partly visible) to the educator. And changes in one factor can ripple through the systems in unpredictable ways, making it difficult to identify simple cause-effect relationships.
Melissa Warr and I, drawing on design theorist Richard Buchanan, refer to these as inherently indeterminate systems— systems that are characterized by incomplete information, complex interrelationships, and often no single “right” solution. Herbert Simon reminds us that in such contexts, we aim not for perfection but for satisficing—making the best possible move given what we know, then observing how the system responds before adjusting course. In other words, “we learn our way towards a solution.”
This, to me, is tinkering.
“Tinkering” is often diminished by the qualifier “just,” implying unseriousness or low stakes. But this misses something crucial about how skilled practitioners operate in ill-structured environments. As Mitchel Resnick and Eric Rosenbaum note, tinkering is “a valid and valuable style of working,” defined by “a playful, exploratory, iterative style of engaging with a problem or project,” constantly trying, adjusting, refining, and trying again.
Josh, by contrast, describes tinkering as “an almost haphazard approach,” marked by “the lack of a fully formed and researched plan.” He describes experimentation as something more formal: a step-by-step plan, a defined outcome, and a clear method for assessment. This distinction is conceptually useful—but it starts to fray when we start to recognize the complex interdependencies and unpredictability inherent in these systems.
This process often appears messier than formal planning—but in many ways, it’s more responsive. While planners may survey a situation, identify problems, make a plan, then execute, tinkerers react to the specific details of the particular situation and through that process engage in what Donald Schön called a “conversation with the material”—or in an educational context, a conversation with the evolving classroom.
Resnick and Rosenbaum point out that this tinkering mindset has deep roots across cultures and disciplines. They reference Claude Lévi-Strauss’s concept of bricolage: making do with materials at hand. The bricoleur improvises continuously, in contrast to the engineer who first plans, then gathers precisely what’s needed.
This ethos resonates strongly with how teachers actually work. Traditional education research often tries to sanitize this complexity—offering “best practices” to be implemented “with fidelity.” But that ignores the multi-dimensional, interdependent nature of teaching and learning. Teachers are natural bricoleurs, adapting creatively to shifting student needs, limited resources, and evolving constraints. What they do is not haphazard—it’s responsive, professional, and often deeply reflective.
A typographical design exploring the invisible presence of the self (“I”) across acts of bricolage, improvisation, tinkering, iteration, and design. The absent “I” haunts each word—subsumed, yet central—while the red text, “The ‘I’ can see” plays on both eye and ego, hinting at perspective, agency, and the tension between disappearance and authorship.
Of course, indiscriminate tinkering can go wrong—as Josh’s example of fixing his bike makes clear. That’s why I advocate for what I call “thoughtful tinkering at the edges.” It’s about making small, intentional adjustments grounded in professional knowledge and intuition. For instance, every time I teach a class, I try and experiment with one new technique or technology—not across the entire semester, but for a unit or two. Based on how it goes, I either refine it or discard it. Over five years, this approach has transformed my teaching.
This method builds on existing expertise, respects the complexity of classrooms, and allows for course correction without catastrophic failure. It also honors teachers’ capacity for judgment and responsiveness in ever-changing contexts.
And over time I have come to embrace some key principles: Start small by making incremental changes that can be easily reversed or modified. Build on expertise by using your professional knowledge and experience to guide decisions. Stay connected to purpose by ensuring changes align with your pedagogical goals and student needs. Remain responsive by paying attention to how students, colleagues, and the system respond to changes. Iterate thoughtfully by using what you learn to inform the next small step. And above all, accept indeterminacy by recognizing that you can’t control all variables or predict all outcomes.
To be clear, this is not an anti-data stance. Data informs our decisions, but doesn’t drive them. There’s humility in acknowledging the limits of what we can know. As I wrote in my tribute to Lee Shulman earlier this year:
Education is a design profession, and like all design professions, works in the space of indeterminacy. Data is often messy and incomplete, and sometimes contradictory. But waiting for all the data to arrive was not an option, or maybe even impossible. The only way forward was thoughtful, grounded exploration, seeking insights not just from experiments, data and theory but also from history, philosophy and experience.
That to me is the foundation of the tinkering mindset.
Josh’s concerns about thoughtless adoption of AI tools are entirely valid. At some level, our disagreement may be more semantic than substantive—we both advocate for thoughtful, purposeful change. What matters isn’t whether we call it tinkering or experimentation, but whether we approach teaching with care, curiosity, and respect for complexity.
In the end, educators have always been bricoleurs—crafting solutions with what’s at hand, responding to the moment, and slowly, patiently, shaping the future.
Thoughtful tinkering has stood the test of time, and I am confident it will serve us well in our AI-suffused future.
Reference: Resnick, M., & Rosenbaum, E. (2013). Designing for tinkerability. In Design, make, play (pp. 163-181). Routledge.
0 Comments