Launching today
MakeSense
Learn anything through what you already love
12 followers
Learn anything through what you already love
12 followers
The fastest way to learn something new is to connect it to something you already know. That's the idea behind MakeSense. Choose a topic. Choose your anchor. We generate a structured knowledge tree with parallel explanations connecting one world to another. Upload PDFs, articles, or YouTube videos to make every lesson your own. This is digitalization of my learning method that has helped me a lot as a student. The demo is free. Do check it out!
👋 Hey Product Hunt - Arshil here.
Over the last year, I won 4 national hackathons and 1 international Google hackathon as a student. Each of them required different fields of knowledge. I also learned a new language during this period while also maintaining a decent GPA. For the first time in my life, I was actually consistent in the gym too, though there is still work left to be done in that department.
I am telling you this because it was only possible thanks to a new learning style I used. Although I have heard bits and pieces, I never seen anyone use the full method.
- The first thing I do before you learn anything is to make a birdeye map of the concept. Focusing heavily on why I are learning what. How it connects to the next topic(rarely anything exists in isolation). Half of my work is already done, by the time I know why I am learning and how much do I need to learn.
- The second part is diving into each "island" and memorizing them. For this I often found myself using LLM to connect the topic to fandoms I love(Pokemon). The result was abysmal. Then I started adding tiny systems to improve the result, over the past year. By the time I was done, I had a system that specialized in analogous learning.
This is the exact system I am selling through my app.
Would love to learn about your learning styles...
How does it actually generate the knowledge tree, is it pulling from a fixed set of sources or using the uploads I give it to build the connections between the topic and my anchor?
@hasanmutl9pxx If you give it the uploads, it will limit the knowledge to that. If not, you can select between 5 depth levels, and based on the 80-20 rule it will generate the relevant scope.
For generating the trees themself, we are using an agent equipped with proper examples and tested instructions.
For fandom knowledge, we are using a mix of google search and custom data. Lesser used anchors are searched, or the agents internal knowledge base is used. Popular anchors are manually(will automate this) trained on relevant data.