3 ms·
"could be inferred from the new founder bios." Could be, but that's a lot more work to do when reading applications. Inferring these facts from a heterogeneous
by jonnathanson 12y ago
"could be inferred from the new founder bios."
Could be, but that's a lot more work to do when reading applications. Inferring these facts from a heterogeneous and unstructured data source, like hundreds or thousands of bios, requires a lot of effort that is better spent elsewhere. Setting these up as simple red flags, through a series of boxes to check or not check, solves the problem more efficiently. Right off the bat, you can sort all applications by any of the criteria in these boxes. You can do that before even starting to read them. If you're inferring from bios, on the other hand, you have to read every bio before you can sort.
- nash11 12y agoThere would be many applications this year i would believe with the new look. Paul's version shooed away people who weren't serious about themselves applying for YC, That was a good tactic anyways.
- TheMakeA 12y agoI think we're on the same page. Each founder already has to answer "Will you live in the Bay Area January through March 2015, if we fund you?" and "Will you commit to working exclusively on this project for the next year?" in a drop down in the new founder bio page. if numWontLiveInBayArea >= total / 2 then flag and if numNonExclusive > 0 then flag Could be totally automated.
- jonnathanson 12y agoMakes sense. Perhaps the two questions are there as a redundancy. Some people might answer "Yes" to the first question, but then "No" when asked in a slightly different way. In the first question, they're just answering for themselves. In the second question, they're taking a deeper mental look at their team and assessing if anyone on it isn't up for the move or the 100% commitment. To whatever extent there's conflict between answers to A and B, that's a sign that some people aren't being totally honest in answering A. (This is purely speculation, of course).