yaroslavvb

They spent a lot of effort on consistent naming. It's basically equivalent of hundred of open source math/visualization packages merged into one, and a single person has made sure for 30 years that functionality integrates in a consistent way. It's much less overhead to remember that using a large collection of OSS libraries. So the naming structure in itself is valuable

You can use eigenvalue solver to find singular values of any matrix but not vica versa. Implementing eigenvalues needs complex numbers

I had similar experience, Claude made report of MRI for achilles tear, it measured the gap, but it was completely hallucinated. Achilles tendon is black on the MRI, it instead measured 13mm distance between two completely different things (looked white), the radiologist looked at and saw no gap at all

This project gave me motivation to build the deep learning next token prediction integration for JetBrains because I was using PyCharm at the time. (Eventually, it wasn't continued because it was kind of expensive to host)

I don't hold grudge, GPT-2 wasn't that great of a model, so releasing it would be more of a publicity value. But the blog post already had that purpose.

Thanks for the share! Didn't realize eleutherai launched around same time

Modern AGI discourse defies the voice of reason

Stricter (but not looser) standards can be imposed on state level. Canada has no binding national drinking water law, they leave it to territories/provinces to decide how to implement guidelines.