hard to decide, it would be one of these two: 1) Finishing the project (last 15-20%): this is mainly for two reasons: - with previous generation of LLM models used for vibe coding it was very easy to get stuck in a "halucination" loop once your code base became fairly complex. At certain point a bug in core functionality would appear. By fixing this bug, two other bugs would be introduced and it was usually very difficult to fix them. Once I've fixed them both, the first bug would reappear and this would happen over and over again. After getting frustrated this actually made me a better vibe coder - I've learned to reset session instantly the LLM hallucinated for the frst time. I've learned to write better prompts and make my code base more granular, split problems into smaller problems. - I am an ADHD person, an idea maker boasting with ideas all the time. I'm a thinker, not much of a doer (although that has changed a lot thanks to AI). I love to start new things but once they are 60% done they start to bore me as I don't see further problems/challenges with that given idea and I like to move to another project. 2) Although I have been in digital marketing for 24 years (and maybe because of that), I find marketing very boring in the era of AI and marketing automation. I dont have a problem to draft perfect go-to-market strategy but kind of lazy to execute it. For these two reasons I'm seeking for people to collaborate on different projects. I have 40+ startup/product ideas in my drawer, with a functioning business model / monetization. (I've built tens of digital products for various customers including Samsung, BMW or Microsoft and I've co-founded 4 startups and exited 3 of them)