Fall in Love With the Problem: Teaching Judgment in an Age of Instant Solutions
Fall 2026 Issue 02 | by DR. CARLOS BAZAN AND ANGELA AVERY

Earlier this year, entrepreneurship educator Steve Blank encountered an unexpected problem. Students in his Lean LaunchPad course arrived on the first day with products that previous cohorts might have spent much of the semester developing. Using generative AI, they had built sophisticated apps, digital twins and other working solutions before the course had really begun.
At first, Blank and his teaching team were delighted. Then they realized something was wrong. Students were building faster, but they were not necessarily learning more.
Blank summarized the problem memorably: Creating products rapidly and inexpensively had allowed students to “make bad ideas go faster” (Blank, 2026a).
From building solutions to understanding problems
The story resonated with us because it reflects something we have encountered repeatedly while teaching entrepreneurship and working with graduate student researchers interested in mobilizing their research: We naturally tend to fall in love with our solutions.
Entrepreneurs become attached to their ideas. Researchers understandably become attached to technologies, discoveries and methods they may have spent years developing. Students are accustomed to assignments in which the problem has already been defined, and their task is to produce the best possible answer.
But entrepreneurship begins somewhere else. Before asking, “How can I build this?”, we need to ask whose problem we are trying to solve, how significant it is, what people are doing about it today, what assumptions we are making about their needs and what evidence might convince us that we have misunderstood the problem.
This is why customer discovery is central to experiential entrepreneurship education. Students engage with potential users and stakeholders, test assumptions and use what they learn to modify—or sometimes abandon—their original ideas. The objective is not simply to validate a solution. It is to learn.
When a polished product looks like learning
Generative AI complicates this process.
Blank found that students could now create remarkably sophisticated products before understanding the customers they were supposedly building them for. The polished product created what he calls “evidence theater”: something that looked like evidence of progress without necessarily representing customer discovery, hypothesis testing or critical thinking (Blank, 2026a).
In the second part of his reflection, Blank went further. Traditionally, the Minimum Viable Product (MVP) evolved alongside students’ understanding of customers and served as evidence of both technical capability and learning. With AI, students can create a convincing product before accumulating that evidence. Blank thus provocatively relabels these artifacts Initial Untested Products (IUPs) (Blank, 2026b).
The distinction has implications beyond entrepreneurship.
Generative AI can help students produce polished reports, presentations, analyses, computer code and other artifacts much faster than before. Increasingly, therefore, the quality of a finished artifact may tell us less about the intellectual process that produced it.
This does not make AI an enemy of learning. A recent systematic review found substantial evidence that generative AI can support critical-thinking, problem-solving, analytical and metacognitive skills when appropriately incorporated into learning (Daniel et al., 2025). At the same time, emerging research cautions that the relationship between AI and critical thinking depends heavily on how students engage with these tools and how learning activities are designed (Nasr et al., 2026).
The educational question, then, is not simply whether students should use AI. It is what intellectual work we want students to perform while using it.
Teaching the problem, not just the solution
In this regard, entrepreneurship education offers some useful lessons. We can give students more opportunities to encounter problems that are not already specified. We can ask them to distinguish what they know from what they assume. We can require them to seek evidence from people affected by a problem rather than relying exclusively on readily available information—or AI-generated representations of those people. And, importantly, we can reward students for changing their minds when the evidence warrants it.
Perhaps even “fall in love with the problem” does not go quite far enough. Students should fall in love with understanding the problem while remaining willing to change their minds about what the problem is.
That disposition is as fundamental to research as it is to entrepreneurship: Hold assumptions provisionally, seek evidence, test interpretations and revise when the evidence demands it.
When answers become cheap, judgment becomes valuable
Blank argues that AI has shifted a critical bottleneck in entrepreneurship. When almost anyone can build something quickly, the difficult questions become what we should build, for whom and why (Blank, 2026a).
This also offers a broader lesson for higher education.
Generative AI has made plausible answers and sophisticated outputs extraordinarily inexpensive. It has not made judgment inexpensive. Students still need to determine which questions matter, interrogate assumptions, evaluate evidence, understand context and recognize when they should change their minds.
Perhaps that is one of the opportunities AI presents us as educators and mentors. When answers become abundant, we can devote more attention to teaching students how to ask better questions.
And before they fall in love with a solution, we can teach them to understand the problem.
Entrepreneurship in practice at Memorial University
Research Knowledge Mobilization Fellowship
Insight Business Consulting (IBC) Program
References
Blank, S. (2026a, September 21). The year AI came for us: Teaching entrepreneurship will never be the same. Poets&Quants. https://poetsandquants.com/2026/09/21/the-year-ai-came-for-us-teaching-entrepreneurship-will-never-be-the-same/?pq-category=ai
Blank, S. (2026b, September 25). AI killed the MVP – Long live the IUP. Poets&Quants. https://poetsandquants.com/2026/09/24/ai-killed-the-mvp-long-live-the-iup/?pq-category=ai
Daniel, K., Msambwa, M. M., & Wen, Z. (2025). Can generative AI revolutionise academic skills development in higher education? A systematic literature review. European Journal of Education, 60(1), Article e70036. https://doi.org/10.1111/ejed.70036
Nasr, N. R., Tu, C.-H., Sujo-Montes, L., Haniya, S., Yen, C.-J., & Hossain, S. (2026). Generative artificial intelligence impact on critical thinking in higher education: A comprehensive bibliometric analysis and systematic review. Journal of Computers in Education. https://doi.org/10.1007/s40692-026-00396-z
DR. CARLOS BAZAN is associate dean (Graduate Programs & Research) and associate professor of technology entrepreneurship in the Faculty of Business Administration at Memorial University, with a cross-appointment in the Faculty of Engineering and Applied Science. A former Engineering Chair in Entrepreneurship, his research sits at the intersection of academia, innovation and entrepreneurship, with a particular emphasis on research mobilization. His work explores how institutional and organizational support mechanisms can accelerate translating academic research into societal and economic impact.
ANGELA AVERY is a business development advisor with Memorial University's Research Innovation Office (RIO), where she supports initiatives that advance research partnerships, commercialization and knowledge mobilization. Drawing on experience in both private industry and higher education, her work focuses on connecting researchers, students, industry, government and community partners to create pathways for collaboration and impact. Her interests include research translation, student success, innovation ecosystems and the role of strategic partnerships in advancing societal and economic outcomes.
.png)