What I learnt from Kickstarting Persuaide
The way to Persuaide
One year of Deep Learning AI capabilities. Starting to see use cases everywhere. Doing a structured literature review to understand how persuasion and text work together. Find amazing companions that are inspired by the idea. Apply for funding. Be all excited.
Major Challenges
- Everybody is nice. Really? Peter Gloor.
- You non-tech cofounder does not believe that AI is already there? How do you help him?
- Test everything with users. Try to get as many people as possible in front of a camera and your tool. Make everybody feel the pain.
- Speak up w/ your founder family. Be honest. Kick their ass. If you feel like something is going wrong - for me it was transparency, tell them. It is not easy but there is no time to not do so.
Major Findings
From Project to Project.
- Lots of ideas, MVP the shit out of them. 80% solutions are easy. 95% solutions create quite some difficulty.
- Building datasets is a horse shit of work, and needs all kind of data engineering.
- Document everything, have frequent iterations.
From AI to Everything.
- You are not only responsible for the deep learning and machine learning.
- But: marketing strategies and growth. Being a coach and motivator.
- Also: Talk to investors, friends, interested people.
Uncertainty says Hi. Embrace it.
- Funding gets scraped. You wait forever. 2 months. Then someone says no.
- Covid prevents you from traveling whereever you want. Also from meeting your people.
- Just make a plan. Commit. 3 weeks is enough. Or even 2?