I'm running through the learn-validate-build loop and want to share what I have learned this far. this will improve my learning, generate feedback and I can lookup my notes when I need a reminder. And I need that reminder probably very soon.

Constraints

only onchain opportunities → I want to focus on crypto ideas
no scams, casinos or pfp projects → we don't run scams and we don't pull rugs

How to treat your ideas

most ideas I get are probably bad

but the good ideas might look bad in the beginning. it's okay to just add to a long list of seemingly mediocre ideas. what worked well is to separate idea generation and judging them. I took a few hours every other day to go through the idea list to 1) analyze how excited I am about them 2) how feasible they are to build for me and 3) if it's a good time to work on it, from my personal perspective and a market / big picture point of view.

enough planning: I didn't want to spend more time in analysis mode and wanted to GO, but the default action for me is writing code. That might be a bad idea!

How I validate random ideas

how much to build vs. doing anything else

writing code is fun for me. the problem with it? I can't validate a startup idea by hiding in my code editor. What I started to ask myself: What will I answer by building this? Or am I just coding for fun and to avoid talking to users?

get feedback before it's too late (it's always too late)

the question "would you use this?" will only lead to problems (see the Mom Test)
the energy that you're sending out will approximately come back to you. people can smell if we suck energy and waste everyone's time. but we want to move fast and time is running out. how do we continue?
what worked well for me is a video showing my face talking about the idea and giving a sneak peek into the app. wait! you said you didn't write code and now you showed them an app?! thank you for paying attention, let me explain: we can't build a full app and see how users behave, so we have to move in small steps. here's what I did:
I've built a user interface that expresses my initial ideas how users might get value from it. it can look ugly, if you're not asking for feedback on design or vibes. questions that I asked: "I've built this flow roughly like this and it ends here, how should it continue?" or "Technically it's x right now, have you heard of a solution that makes it y?"
Not visible in my video: I've built a backend and database model. This answered my questions around how fast I can move with the stack and frameworks I've selected. I now have a full-stack deployed app that does almost nothing. but I can build out at anywhere in the product landscape (Login flow, frontend pages, user flow, backend logic, database queries and tables, etc).

actual feedback that I got from the video

high % of replies to my outreach, did not expect this 🙏
many questions on consumer vs B2B and first use-cases. everything that I avoided or didn't express clear enough came back as questions
show the user benefit first, then how the sausage is made. learned this for next time
feedback conversations repeatedly turned into talking about focus, which initial value will be provided and how that connects to a long-term vision. many things I haven't figured out here so far.

use AI to structure feedback and learn

I started a new ChatGPT o1 session and dumped all my notes, the transcript of my demo video and all feedback. AI is good at finding patterns and will hallucinate less if it has text to work with. sidenote: I wanted it to be a great brainstorm partner, but didn't get an LLM to help with it, yet. It's been great in structuring conversations and suggesting follow-up tasks

I'm happy with the flow for this week and the tangible improvements I can make to to it. subscribe to see next week's post how it all turned into nothing...or something