Thanks for that post. Which decisions are most discriminating for us, or would pair us up most productively? Who you take money from? Who you’re trying to serve? Pragmatic incrementalism vs shooting the moon?

Some initial questions, with incomplete possible answers in parentheses: • What’s the #1 problem to attack? (open-ended) • Are you trying to use what’s already been built? (binary compatible, source compatible, muscle memory compatible, conceptually similar, nope) • What level are you operating at? (app, UI toolkit, CLI toolkit, specific programming language, software library, OS kernel, everything) • What’s the representation? (Free text, structured text, visual, kinesthetic, natural language) • … Hopefully that’s a starting point. I suspect the list of possible answers in each case is just a sample of points on a continuous spectrum, with infinite extent in either direction, and infinite detail between any two points. But then there’s the open-ended answers (and there are probably more such questions to ask). I’m hoping there we start out with all unique answers but can over time gradually notice opportunities for coalescing based purely on structural similarities in follow-up questions (similar turns in the maze). I’m a little disappointed that my questions are mostly about solution characteristics rather than problem statement characteristics. Hopefully others can help figure out that side of things. I think it’s the more important side.