Vlad

This is a good move to try take back entry market share from the Capcuts and Canvas.

Gotta know your way around Figma. Basically table stakes if you want to iterate quickly.

I find it's better to ship fast vs delay and polish. Early user feedback is priceless, and if you're working closely with them they'll be quick to forgive any rough edges.

Depends how young you are. In your 20s you can loose it all without it mattering too much big picture. High risk ideas often pay off so long as you aren't playing lottery ticket odds.

Crazy turnaround, I remember when it first launched it was considered total slop.

Maybe. The output would need to be grounded on human generated information/analysis (historical books, encyclopedias) to have any chance of exceeding quality of info on Wikipedia. The good news is that all the articles already cite their sources, so it'd be pretty straightforward to implement a "deep research" with those sources as the root.

I'll answer the easier question first: Yes, too much context is bad for model output quality This is for a couple reasons, but as a rule of thumb if you can make the context smaller, do it. Deciding which information to cut is tougher. Without knowing your exact situation it's tough to give specific advice, but here are a few generic pointers: Long prompts tend to duplicate information, that is...

I like it! This solves a really big pain point for us, getting the word out there without polluting the internet with AI slop. Giving it a shot!