This week wasn’t about shipping features it was about understanding user cognition.
We started understanding something deeper.
What changed this week
Over the past few days, we collected feedback from:
Students
Self-learners
Builders
Across different regions, including Thailand and Korea.
At first, the responses felt different.
But when we looked closer, a pattern started to form.
The pattern we discovered
Students are not struggling because subjects are hard.
They struggle because:
They memorize without understanding
They misunderstand questions
They don’t see step-by-step thinking
They get overwhelmed by too much information
And most importantly:
Different students think in different ways.
The real problem
Most learning systems assume:
“One explanation works for everyone.”
But in reality:
Some students need step-by-step breakdowns
Some prefer simple explanations
Some think logically
Some need real-life context
Some learn through emotional understanding
When the explanation doesn’t match the learner,
confusion happens.
What we improved in BAINT AI
Based on this, we updated our classroom assistant demo:
Added multiple explanation modes
→ step-by-step
→ simple
→ logic
→ context
→ human
Improved how answers are structured
Focused more on clarity, not just output
What we are learning
We are starting to see that:
Learning is not just about information
It is a system that combines:
Memory
Understanding
Thinking
Context
Emotion
And if one part is missing, the system breaks.
Shift in thinking
Before:
“How do we build better AI answers?”
Now:
“How do we help people think and understand?”
This shift is changing how we build everything.
What’s next
We are still in the early stage.
Next, we are focusing on:
Making explanations more adaptive
Reducing confusion in first-time users
Improving how users interact with the system
Closing thought
We are not building:
“An AI that gives answers”
We are building:
“An AI that helps people
understand and apply knowledge in real life”
Top comments (0)