Homework Is Broken — and Professors Know It
Homework is supposed to reinforce learning.
In practice, it often reinforces frustration — for both students and professors.
Recently, we spoke at length with a chemistry professor about how homework actually gets created, assigned, reviewed, and adjusted in a real college course. Not in theory. Not in marketing demos. In the messy, day-to-day reality of teaching.
What we heard wasn’t a complaint about students or technology.
It was something deeper: homework systems have drifted away from how teaching actually works.
How homework actually gets made
Most people assume homework assignments are thoughtfully designed, step by step, to match lectures.
The reality is more constrained.
Homework usually comes from publisher-controlled test banks bundled with textbooks. Professors don’t start with a blank slate — they start by scrolling through prewritten questions, hoping something fits what they actually taught.
Sometimes they get lucky.
Sometimes there are only two usable questions for an entire topic.
Sometimes there are fifty, and forty-eight of them are wrong for that class.
If the platform allows it, professors write their own questions. If it doesn’t, they’re stuck.
“If I’m writing all my questions from scratch, it takes days.”
Even with experience — even after teaching the same course for years — building a single homework assignment can take hours spread across multiple days.
That time isn’t spent thinking about pedagogy.
It’s spent fighting tools.
The real goal of homework (it’s not the score)
When professors check how students are doing, the first thing they look at isn’t grades.
It’s time.
- Did students finish too quickly? → The assignment was too easy.
- Did they spend forever and still struggle? → It was too hard.
- Is the time on task reasonable for the learning goal?
Scores come second.
Homework isn’t about correctness alone — it’s about productive struggle. The right amount of friction. Enough repetition to learn, but not so much that students disengage.
Most platforms don’t surface this insight well. They optimize for points, not understanding.
Where students actually get stuck
When students ask for help, the issue is rarely the final step.
It’s the beginning.
“I don’t know where to start.”
Professors care deeply about how students approach a problem:
- What value did they start with?
- Did they track units?
- Did they choose a sensible first step?
Units, in particular, are a constant failure point. When students drop units, numbers become meaningless — and instructors lose the ability to diagnose the mistake.
What professors want isn’t the answer.
They want to see the student’s thinking.
Academic freedom matters more than features
One of the strongest themes we heard was control.
Professors want:
- The ability to import their own questions
- Freedom to modify assignments late when lectures shift
- The ability to move or reorder questions without breaking everything
Some platforms block this entirely. Others allow it, but make it painful.
A common frustration:
- Clicking a tiny edit icon next to a delete button
- No autosave
- Multiple confirmation screens
- No drag-and-drop
- No way to edit assignments once students have started
These aren’t edge cases.
They’re daily experiences.
Fragmented workflows are the norm (and everyone hates them)
Today’s teaching workflow looks like this:
- Slides in PowerPoint
- Worksheets in Word
- Homework in a third-party platform
- Announcements in the LMS
- Email for individual communication
- PDFs everywhere
Professors make this work because they have to — not because it’s good.
Many have spent over a decade building personal systems to survive this fragmentation. New instructors don’t have that luxury. Teaching a new class often means working up until the last possible minute, constantly revising materials.
The dream is simple:
“Everything in one place.”
Not for convenience — but to reduce cognitive overhead for both instructors and students.
What not to build
One thing was very clear.
Any system that solves problems for students is a non-starter.
Professors don’t want shortcuts.
They want guided reasoning.
They want tools that ask:
- “What does this number represent?”
- “Where did you start?”
- “Why did you choose that step?”
That distinction matters.
The real failure of modern homework platforms
The biggest failure isn’t technology.
It’s priorities.
Most platforms obsess over:
- Locking content down
- Preventing cheating
- Enforcing rigid flows
Meanwhile, they ignore:
- Instructor time
- Academic flexibility
- How learning actually unfolds
They make the student side polished and the professor side painful.
That tradeoff is backwards.
Why this conversation matters
This interview wasn’t about features.
It was about respecting the craft of teaching.
Good professors are already doing the hard work:
- Designing iterations of problems
- Adjusting difficulty in real time
- Watching how students think
- Constantly refining materials
The right tools shouldn’t replace that work.
They should amplify it.
That belief sits at the core of what we’re building at LabNotes.ai.
Not answers.
Not shortcuts.
But systems that align with how learning actually happens.