I looked for an independent primary source on on these events and wasn't able to find one. However, it may be of interest that the conference program lists the

I looked for an independent primary source on on these events and wasn't able to find one. However, it may be of interest that the conference program lists the talk title as "Creating delightful libraries and books with nbdev and fastdoc" [0], whereas the delivered talk has the title of "I Like Notebooks" and is a direct response to Joel's talk. The conference organizers, or others, may have expected a different keynote and felt deceived by the change in topic. Note also that, according to Jeremy's blog post, the version of the talk posted on YouTube is not the version that was presented at the conference. As third parties we have very little information. I'm also uncertain if the investigative committee actually intended to impose any consequences. Jeremy expressed concern that the situation may prevent his talk from being available, but I don't see any talks posted on the JupyterCon site yet. The only definite action is the decision that a code of conduct violation had occurred and the notification thereof. The notification clearly caused Jeremy Howard significant distress, which should have been avoided. However this may have been unintentional on the part of the conference organizers. For all we know, the situation has also caused them similar distress. Hopefully the involved parties can still find some remediation despite the incident blowing up on social media. My main take-aways are that (1) small technical organizations are often poor at community management and investigating potential misconduct, and (2) the public internet is worse, with a tendency to devolve into public shaming. [0] https://cfp.jupytercon.com/2020/schedule/general-sessions/ https://cfp.jupytercon.com/2020/schedule/general-sessions/