Person doing PhD in AI here (Ive seen all of OpenAI's research, been to their office couple times, know some people there) - tbh the piece was a pretty good summary of a lot of quite common somewhat negative takes on OpenAI within research community (such as that they largely do research based on scaling up known ideas, have at times hyped up their work beyond merit, changed their tune to be for profit which is weird given they want to work in the public interest, and that despite calling themselves OpenAI they publish and open source code much less frequently than most labs -- and with profit incentive they will likely publish and open source even less). The original article also presented the positive side (OpenAI is a pretty daring endeavor to try to get AGI by scaling up known techniques as they are, and people there do seem to have their heart in the right place) .
I feel OpenAI has stayed fairly close to its mission and principles. I'm not sure about why so much secrecy, they could have just NDAed and let the reporter take look at things which is common practice even at Apple. However, apart from that, they are the only ones explicitly sensitive to safety and economic issues surrounding AGI. Specifically, AGI is certain to generate the first trillionaire or even deca-trillionaire and throw the wealth gap between countries completely out of control. Unlike guns and nukes, it would be a far more difficult tech to clone for less fortunate groups. I think OpenAI is the only group actively working against that outcome (i.e. max 100X return provision) while most others wish to race to AGI first, patent the hell out of it and become the most powerful richest entity world would have ever seen. Also, until recently with only 200 folks and limited pay (relatively speaking), they have amassed extraordinary talent that has given run for its money to teams 10X larger. When people say that they just try to copy/scale others innovation, they grossly underestimate very impressive contributions that has touched virtually everything from supervised, unsupervised, RL, NLP and robotics.
