BioML Nonsense

A place for niche jokes

---

Scaling laws for BioML and wet lab data will eventually work out in the right setting! After all, language data for LLMs was acquired by the largest wet lab experiment ever conducted: Human civilisation 🤯

---

I find it ironic that CuEquivariance houses kernels for the triangular attention and triangular update operations that basically killed equivariance.

---

Following Demis’s bold claim that ‘we can cure all disease with AI’, Isomorphic Labs will eventually pivot into wellness and yoga, promoting healthy lifestyle choices powered by AI.

---

You join Latent Labs, only to find out they are still using SE3 equivariant diffusion — how upset would you be?

---

AF3 actually got rid of equivariance and frames to be sufficiently ‘novel’ and different from RF-AA in the eyes of ML people. There are equivariant and frame-based versions of AF3 within GDM/Iso that work better for certain tasks!

---