Stop squashing your commits. You're squashing your AI too
2 points by jannesblobel 8h ago 7 comments
Ask HN: Best codebases to study to learn software design?
100 points by pixelworm 2d ago 89 comments
Deeper Than Deep: David Reich's genetics lab unveils our prehistoric past (2017)
37 themgt 6 8/26/2025, 1:52:13 PM laphamsquarterly.org ↗
I only wanted to point out that you're almost forced to come up with some theory you have to explain genomic findings. Over time, yes, most of those theories have historically been shown to be bunk. (My gut tells me this one will share a similar fate.)
But that's the scientific method. Propose a hypothesis (theory). Design experiments to test said theory. Present the results. Which should lead you to another hypothesis.
In the "propose a hypothesis" phase, at least in fields like this one, the proposals will tend to be informed more by world view than by science. That's how humans think. I don't think there's any changing that. The control, or "check and balance", is supposed to be the objective review and validated replication of the results. Which is, I grant you, lacking at times.