Show original
Try the app
Enjoyed this article?
Use "Request tipping" to ask the author to set up tip receiving.
AI translation
I want you to know what is fundamentally different about this from other idea apps!
Try the app
Enjoyed this article?
Use "Request tipping" to ask the author to set up tip receiving.
I worked on this project to solve the problem of ideas generated through brainstorming with AI becoming similar to those of other people.
AI is structured to output statistically correct answers, so even if the starting point is unique, there's a structural weakness where ideas tend to converge toward common ones during brainstorming. I was researching ways to break through this limitation.
It's often said that in an era where AI has become widespread and creation has become easier, ideas will become increasingly important.
However, when I searched for "idea app," I only found memo apps for jotting down ideas and tools for organizing thoughts.
I felt there was no product that fundamentally addressed how ideas are actually generated, so I decided to create one myself.
I particularly focused on the internal logic.
This app is based on analogical thinking. It works by matching the user's concept with naturally occurring phenomena that have a structurally similar structure, then detecting blind spots from that match to generate ideas.
In this process, I abstractly structure things using constituent elements and the predicates that connect them. The key point I researched and refined was determining at what level of abstraction I could identify analogies between the user's concept and natural phenomena.
AIとの壁打ちで出るアイデアが他の人と似通ってくる問題を解決しようと思い制作に取り組みました。
AIは統計的な正解を出力する構造にあるので、出発点は独自でも壁打ちするうちにありがちなアイデアに寄ってしまうという構造的な弱点がありますが、それを打破するために研究していました。
AIが普及して作ることが簡単になってきた時代に、これからはアイデアが大事になってくるというのはよく言われてますよね。
ですが、「アイデア アプリ」で検索してもアイデアを書き留めておけるメモや、思考を整理するツールなどばかりでした。
どうすればアイデアが生まれるのかに本質的に向き合ったプロダクトがないと思ったので自分で作ってみようと思いました。
中身のロジックに特にこだわりました。
このアプリはアナロジー思考をベースにしており、ユーザーの構想と構造的に似ている自然現象をマッチングし、そこから盲点を検出してアイデアを生み出すという設計です。
この時、構成要素とそれを繋ぐ述語で抽象的に構造化するのですが、どの粒度で抽象化すればユーザーの構想と自然現象のアナロジーを見抜けるかは研究を重ねてこだわったポイントです。