Learning In Public As A Product
I’m building a peer-to-peer app that uses AI prompts to guide Internal Family Systems sessions. Here I share each tiny experiment and lesson in plain view, inviting honest feedback as I go. The aim is to learn in public while staying clear-eyed and free from the doom-scroll.
Experiment: "You don't have to be right, just see for one month or maybe 2 weeks"
Priorities
What it is:
- Integrated: learning in public
- Part of my workflow
- Community to practice IFS is the product, not me. What it is not:
- Choosing between visibility and productivity, but both simultaneously Prioritiy above this:
- Launching with two people using it found through 1:1 connection or finding existing IFS communities and sharing.
- This is just to keep fostering human connection so it can flourish and maybe help more people.
Hypotheses:
Am I doing the above priorities or chasing likes? Am I still productive? Am I building with others and gathering feedback?
PACT
What is a PACT*
Post a video three times a week to youtube** for three weeks and review if you want to continue.
PACT* = In Tiny Experiments, Anne-Laure Le Cunff says a PACT is a self-designed promise to carry out one Purposeful, Actionable, Continuous and Trackable action for a fixed time—so you learn through tiny experiments rather than chase distant goals. So Brief (opens in a new tab)
** Youtube:
- Evergreen reach: videos keep surfacing through search and recommendations for months, so each upload compounds instead of expiring after two days.
- Deeper feedback: threaded, timestamped comments invite thoughtful discussion—ideal for nuanced IFS topics.
- Low-distraction setup: schedule uploads in YouTube Studio, answer comments in a weekly block, and skip the home-feed entirely.
- Room to grow: Community posts, playlists and live streams sit on the same channel, giving you future options without adding new apps.
Process Documentation and Scratchpad
Format for videos:
- What I did
- Why
- Positive
- Negative
- Next
Script
- Write a basic script
- See video-scripts
Content and Visuals
- Load up pages or visuals
- Visual slides: https://www.figma.com/buzz/qR3LmqFRJVTV7pamoyrEEh/Social?node-id=0-1&p=f&t=yeFmktWUH6uBFyi0-0 (opens in a new tab)
- Load up Discord to record and find the right viewport
Record
- On Descript
- Use good mic
- Make sure at high resolution
- Make sure you have the correct framing for youtube
Edit
- Descript basic, removes ummms ahhhs, etc.
- Split screen: https://chatgpt.com/c/68415730-1d0c-800a-a8b0-44daa751eb5d (opens in a new tab)
- Sound quality
- Eye contact ? Looks weird
Title and Description
- Gernerate in ChatGPT or Descript
Thumbnail
- https://fal.ai/models/fal-ai/flux-lora?fromTraining=df2b7e6e-b583-4d8e-8334-5c9b4db05806- (opens in a new tab) Create the image
- Edits to image: https://fal.ai/models/fal-ai/flux-pro/kontext (opens in a new tab)
- Buzz template: https://www.figma.com/buzz/qR3LmqFRJVTV7pamoyrEEh/Social?node-id=0-1&p=f&t=kgZUbSFCihi5cMRJ-0 (opens in a new tab)
- Add icons: https://www.tech-stack-icons.com/ (opens in a new tab)
- Output thumbnail from Buzz
Upload
- To Youtube via Descript
- Add thumbnail in youtube
- Publish
Optional short form:
Not sure this is worth it!
Setup
- Take transcript from descript
Package for short-form
- Take long form video
- Copy it in descript and then edit the layout to mobile
b-roll creation
- Take transcripts of short-forms
- Transcript to o3 and create b-roll prompts
- Prompt b-roll in Soras
short form edit
- Add captions
- Add b-roll
- Upload to instagram (stealth for now)
Review
See above to do Positive, Negative and Next and answer your hypotheses.
Later
- Put long-form video or shorts in substack and share to email subscribers
Future idea
- If you keep doing the above, automate:
- Human following the docs
- Create it as a tool - the outcome would be that you upload your video you recorded, the tool does the rest right through to publishing.
- In both human and tool could be something you sell