I'm starting my fifth year in math PhD. For two years I've been thinking on a well known problem (well known enough to have a Wikipedia page) in my field with no success and all of a sudden I think I did solve it with a very short solution. I haven't written it down yet but I can do so in 1 week. The bottom line is that I'm in a situation that there is nothing positive about my CV except solving the conjecture. I do not have any other publications and I'm very unknown,no professional mathematician in my field is aware of my existence. I'm planning to apply for postdoc positions this coming fall but what I'm struggling with is the letters of recommendation. Let's say I'll write my arxiv version of the paper in 1 week and about 1 week it will take it for me to explain it to my adviser and proofread it. My adviser hopefully will help me to give some talks at some places (I haven't given any talks outside of my university during my PhD). I think the standard procedure is that after giving the talk I should start asking for the LOR from the people who attended my talk until I can get 2 LORs. Problem is that this procedure will potentially take a lot of time and as people require at least a month to write LORs by that time the deadlines will be passed. I wanted to ask for some ideas on how to get LORs in my situation?
I was thinking maybe I can send emails directly to the people in my field with my preprint and directly ask for the LOR but someone told me this is very unusual and might not be a good idea. What are your opinions on this?