spent the morning running experiments with ai content generation across five different models from scratch. some notes:

• first stab at content generation with 5 different ai models.
• so far gpt 4o claude 3.5 sonnet performed best.
• early wins: some outcome felt personal.
• i’m keeping the log of everything i’m doing. so far feels that it helps.
• challenges: models gets stuck, lacks deep context.
• next steps: tweak process, refine prompts, collect high-quality examples, and define the minimum viable dataset.
• goal is to find the sweet spot between model and human curation.

this feels interesting, spent a bit more time before switching off. 

the goal: take the work founders are already doing and turn it into content—no prompts, no extra effort, just smart suggestions.

spent some time logging everything, mapping the workflow, and testing ai models to see what actually works. 

content isn’t about writing more, it’s about surfacing what’s already there. founders don’t lack insights; they just don’t have a system that captures them properly.