We Can Do Hard Things: Preserving Productive Friction in the Classroom

“Comet powers a shift from browsing to thinking.”1 With this and other sweeping promises, Perplexity AI announced the launch of its new AI-powered browser, Comet. Students were among the first to gain free access through a one-year promotional offer. As both a parent and an educator, I was struck by the bold claims made in the announcement. Promising to reduce “friction at every thought” and even to “become a second brain,” the browser seems to position itself as absorbing our thinking for us, turning our online searches into “seamless interactions” and “fluid conversations.”

While the idea of not having a million tabs open at once is undeniably appealing, my alarm bells went off when I read that this technology aims to remove friction from every thought. What does it mean for a generative-AI–driven browser, one that directly targets students, to vow its users will “understand everything, instantly”? If this tool becomes students’ “study buddy” or tutor, as its advertisements propose, what does that mean for how students actually learn? Beneath the promise of frictionless ease, and the implicit equation of browsing with thinking, lie more fundamental questions: what do learning and thinking actually involve? And do we really want to remove friction from the process? Taking this a step further, we can rightfully ask ourselves: what if the friction-full and repetitive tasks we so eagerly want to outsource are not enemies of our growth and self-realization but their necessary pre-conditions?

As a parent, I witness on a daily basis that a certain amount of friction is necessary for children’s development. I see my children struggle as they acquire new skills and observe the frustration that accompanies reaching new milestones. From learning to sit, crawl, and walk when they are little, to seemingly smaller, but often equally frustrating feats, such as zipping up a jacket or tying shoelaces, growing up means making your way through a string of “friction-full” experiences.

Despite a parent’s instinct to want to step in and help out, we know we need to let our kids struggle so they can learn, even if it means it takes twice as long to leave the house because those shoelaces really are a pain. As parents we watch over these friction-full moments and offer a helping hand when necessary, without doing the thing for them. And while I’m more patient some days than others, helping my kids achieve age-appropriate independence, even though it’s hard initially, is high on my list. In this way, we can see friction as an investment in the future: by allowing kids to struggle in the moment, we’re helping them achieve ease and mastery later on.

As educators, we aim to do the same. We aim to operate in this “construction zone” of learning, between what students can independently do and what they can potentially achieve. This is what Lev Vygotsky so famously called the Zone of Proximal Development, or ZPD. Within this zone, educators offer guidance and scaffold learning experiences to work toward a student’s “level of potential development.”2 Friction is a critical component of this learning zone. In fact, absence of friction indicates that we’re not in the right zone. Students are either already independently capable of understanding or performing a task or too far off—not able to get there yet, even with guidance and scaffolding. In Vygotsky’s theory, collaboration and guidance from educators or more advanced peers is crucial. Learning is relational. We learn from and with others. It doesn’t happen in isolation, and in this apprentice-like phase, educator or peer support and constructive criticism anchor the learning process.

It is important to distinguish between productive and unproductive friction, and recognize that whether or not friction-full experiences are welcome or necessary entirely depends on context. Critically, systemic obstacles to learning and uneven access to education are not what’s considered here as friction. Overcoming such inequities takes priority. Yet, within fair and supportive environments, friction can be productive, creating the conditions through which growth, reflection, and understanding emerge.

It appears I am not the only one unsettled by the promise of a frictionless life and our growing embrace of technological instantaneity. In a widely shared article for New York Magazine, journalist Kathryn Jezer-Morton dubbed 2026 the year of “friction-maxxing,” pointing out that tech companies increasingly frame everyday life itself as an inconvenience to escape from: “Reading is boring; talking is awkward; moving is tiring; leaving the house is daunting. Thinking is hard. Interacting with strangers is scary…. These are all frictions that we can now eliminate, easily, and we do.”3 As a playful countergesture, she suggests intentionally reintroducing inconveniences into daily life: ordering in less often, avoiding constant location sharing, resisting defaulting to tools like ChatGPT, and, especially unnervingly to me, not tidying the house before inviting people over.

This year (2026) has seen other, similar, Gen Z–fueled countertrends to our increasingly online life, from “rawdogging” long flights to being “chronically offline.” Still, there’s a widespread confidence that if only we can use generative AI in a “good way,” and teach ourselves and our kids AI “literacy,” that we can become more productive, even more creative. In the same vein as Comet’s advertisement, we routinely attribute some liberatory value to generative AI. If we can outsource the inconveniences of life to large language models, we can regain the more meaningful aspects of it. To me, this way of thinking harkens back to early twentieth-century Taylorism and fantasies of optimization and efficiency. Yet, as we’ve seen before and are witnessing in the extreme right now, optimalization of business processes does not necessarily benefit workers. Rather, more money tends to land in capitalists’ (read: tech giants’) pockets and workers only experience increased alienation, reducing production from a fulfilling expression of creativity to one of meaningless drudgery.4 Increased efficiency has in the past resulted in increased demands. Work more, not less.

We have to acknowledge that generative AI has its inherent problematic characteristics; that it’s not a mere matter of using it in a “good” or “bad” way. We should grapple with these built-in features, especially in our teaching. Besides the commonly known “hallucinations,” a key characteristic of all current large language models is sycophancy, or a tendency to echo and affirm users’ views. Due to human feedback training models, LLM chatbots favor agreement over truthfulness, which makes them particularly unsuitable as “study buddies.” Seeking validation from these tools, we become fearful to think for ourselves or trust our intuition. What’s more, as generative AI cannot produce anything outside of existing patterns and continually seeks out the statistical average, we risk homogenization of content and loss of creativity. It’s telling that we’ve seen a dramatic increase of certain words in student writing, and even in published research, since AI chatbots became widely adopted.

It’s crucial to find out where friction matters. Which parts of our lives can we safely smooth out and where does difficulty add meaning, growth, or depth of experience? Ironically, by embracing extreme countertrends such as friction-maxxing, raw-dogging flights, or being chronically offline, we create space to sit with our thoughts. Is this where original thinking happens? Regardless of the answer, we must acknowledge that this is the space that friction-removing technologies threaten to collapse. Tools that promise instant answers or seamless assistance can short-circuit the very processes through which understanding develops. Experts with proficiency and critical capacity can choose to outsource mundane or repetitive tasks. But when it comes to children, students, learning, and development, we’re walking across a tightrope. We need to understand how tools like Comet impact our students’ ability to do something hard, persist, and come out the other side.

Children are most defenseless when it comes to generative AI. As Jezer-Morton argues, they “don’t yet know the difference between suffering and friction.” But older students, too, are vulnerable in this age of techno-solutionism. Stuck in a performance-driven school system, students are given little space to try, fail, and tinker. As a teacher, I’ve often had to help students unlearn their focus on and anxieties around outcomes and performance before we could re-shift our focus from product to process and exploration. As students have access to more and more tools of escape, the challenge for educators is to preserve and create spaces where productive friction can still occur. I therefore suggest the following reorientation in the classroom:

Slow Down

Young children force us to slow down. When I go on a walk with my young kids, I remind myself to slow down and let them experience the world at their pace. This is when they see, wonder, and learn what’s around them.

In a world saturated with information, we can give students similar moments of pause. Assign less material but examine it more closely and critically together. Ask students to write down what they already know and don’t know before starting research. Discuss what it means to know something. Find out where your students are. Ask them to keep a journal of their learning journey. Slowing down reveals the messy space between what students currently understand and where they want to go. Let them be at ease in that space of friction. In Zygotsky’s terms, expose the ZPD and ask students to document and reflect on the steps they take within it.

Ask Questions, Relentlessly

No one is as curious, and relentless, as a toddler asking a million questions a minute.

One of our most important skills is the ability to ask questions. Surrounded by AI tools, our ability to prompt effectively and understand how prompts are processed is even more important. Critical thinking skills can only develop when we teach ourselves to always question what’s around us. Together with your students, ask questions, not only about your subject but about the tools they use alongside it.

Embrace Failure

When my son is struggling to tie his shoelaces or button his shirt and ready to give up, I remind him that he fell many times before learning to walk. All it takes is practice.

As AI promises us perfection and easy solutions, and our performance-driven culture forces us to fear failure, we need to do the opposite. Embrace failure as a path to growth. Reduce the risks of failure in your teaching by shifting focus from product to process. Assess not the end result, but the way students got there. Scaffold assignments from low to higher stakes and encourage students to grapple with difficult subject material. Documenting how they made their way through a thorny task is more constructive than just showing where they landed.

Make it Messy

No childhood is complete without a healthy mess—mud kitchens, water play, finger paint, and tornadoes of toys across the house.

A staple of any writing class is the “shitty first draft.” It’s that unruly first attempt at getting your thoughts on paper that’s likely embarrassing but essential to the process. In an age of AI-powered instantaneity, try to bring back various forms of these shitty first drafts that expose all the tensions within a topic. It’s these documents of friction and struggle that launch effective learning.

Break it Down

Knock down the block tower before building it back up.

Break down the learning process into manageable chunks. This not only helps students better grasp the material but also increases their sense of accomplishment. We derive a real sense of satisfaction from the idea that we’re making progress. Help students reach this feeling of achievement by rewarding the progress they’re making toward an assignment. Break it down into small pieces and chart a trajectory through those steps, building in assessments along the way. Focus on drafting, various rounds of feedback, and revision.

Make it Matter

Our little helpers thrive when we involve them in our daily tasks, from doing laundry to baking together. And chores and pretend play prepare them for life.

This is less directly related to opening up the ZPD and reclaiming friction, but I believe that our tolerance for difficulty is higher when we’re working toward something that has meaning to us or has real world application. In the classroom, we can offer students choices so they can work on what interests them, and we can give them ownership of their slice of the subject matter by letting them assist their peers. Additionally, we remain outcome-focused and demonstrate real-world applicability and transferable skills.

Conclusion

We need a certain amount of productive friction as we grow and learn. Friction, I believe, lies at the heart of the learning process. Without it, we risk failing ourselves and our students. At my children’s school, third- to fifth-graders wear custom shirts during standardized test-taking that read: “We can do hard things.” While I don’t agree with the tests—that’s another story—the message captures something essential. We don’t need to embrace arbitrary hardship, idolize suffering, or work for work’s sake. But in an age in which technologies increasingly promise to smooth away every inconvenience, we should preserve those spaces where thinking takes effort and growth is rocky. And remember that we can do hard things.

About the Author

Anke Geertsma (PhD Candidate, Comparative Literature; former Teaching and Learning Center Fellow)

Anke Geertsma (she/her) is Communications and Events Manager in the Office of Academic Initiatives and Strategic Innovation at the CUNY Graduate Center, where she works on communications, outreach, and programming for non-degree initiatives. She previously worked at the Graduate Center’s Teaching and Learning Center and has taught language and literature courses at CUNY and at the University of Groningen in the Netherlands, where she is from. She is currently completing a PhD in Comparative Literature on the visual post-memory of 9/11, an interest partly inspired by watching her two children, ages eight and five, grow up as New Yorkers in a world shaped by an event they did not experience directly but will encounter through the images, museums, memorials, and stories that carry on its memory.

What is the best piece of parenting advice you’ve received?

Trust yourself, and your kid(s). And something that I’m finding out as my kids are getting older is that almost everything is a phase. It won’t last forever.

What is the worst piece of parenting advice you’ve received?

Sleep when the baby sleeps.

What is the best piece of teaching advice you’ve received?

Step back. Provide structure, facilitate, and create a space conducive to learning, but let your students do the work. (Swap out “work” for “playing” and this applies to raising kids, too).

What advice do you want to offer Graduate Center student-teachers who are thinking about parenthood?

There’s never a right time to have kids. Don’t overthink. But also: don’t think you can write a dissertation with a newborn on your lap.

What advice do you want to offer all Graduate Center student-teachers?

Have fun with it! Let yourself be surprised by your students. They’re pretty amazing.

This entry is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International license.