*shrug*, I don't really care who does the research, I care if I learn anything from reading it. FWIW, their interp papers are pretty separate from AGI

As a former interpretability researcher who has skimmed their work but not read it closely, I just don't find it terribly interesting or novel. Also, frankly, I find the writing style for the papers pretty hard to parse (as they don't follow standard paper formats) and a tad grandiose, as they tend to avoid standard things like comparing against other methods or citing other work. Relatedly, I think their choice to avoid peer review has impacted how people perceive their work, and limited its distribution.

I think that's always been the case, foundation models or not