I wonder how much of what you know would be applicable to the lowest levels of computational biology, protein folding, that sort of problem. The problems are ob

I wonder how much of what you know would be applicable to the lowest levels of computational biology, protein folding, that sort of problem. The problems are obviously different but have a lot in common with the kind of problems you'd run in to in astrophysics. I've thought about this, but if you set out to sell software to researchers it's very hard to find customers. Anyone working in this field already has some workable software, and your software would have to be an awful lot better in order to persuade them to switch... and the best-existing codes probably represent many man-years of work. I've thought about starting a company in my own particular subfield (first-principles simulations on the atomic level) but have never quite got to the point where I come up with enough imaginary customers to make the effort worthwhile. And that's in my own field, which I already understand... catching up with computational biology from an astrophysics background would take a lot longer.