Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

There's nothing quiet about it. There's a whole district near where I live dedicated to it, it's covered by my health insurance, I've contracted to an IVF company here to do ML for face matching, many of my friends have used it, it's well enough known that the acronym didn't need to be defined.


Pardon my ignorance, but they use ML to find sperm and egg donors that match similar facial features to the “parents”? Cause that is honesty pretty cool.


If you're talking about DNA, yes, but no. I've worked on that too and it's much harder. In the above instance we used photos of the donors and customers and the typical case was to get a close match for the partner who was shooting blanks in order to produce an offspring who ideally would look like both "parents".


That is really awesome, would you happen to know how effective the photo method is? As in, do the kids end up looking similar to their “parents”?


I couldn't imagine an ML project where you have to wait 10-15 years before finding out if the results were good enough!


It’s old knowledge that parents (more typically fathers) bond with newborns that look like them, so you could have results within a year and again within 5 when the child looks more like their adult selves


> I've contracted to an IVF company here to do ML for face matching

This is actually a good idea. I have a hard time believing it actually performs well though




Consider applying for YC's Summer 2026 batch! Applications are open till May 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: