A sad secret that most professional conference goers know: talks can sometimes be the absolute worst use of your time. The smartest people in your field from all around the world are now physically in the same room as you - and then you all sit silently.
My favourite conference experience ever was an accident. What started as a few people catching up for lunch became an impromptu language modeling / translation / knowledge graph discussion which literally became a round table (we found a round black marble table + sat at it) ^_^
There was no moderator, no plan, no audience (everyone spoke), and many didn't know who the others were or could only vaguely recall the names as an author on a random paper. We delved into successes, half done projects, failed experiments, random mysteries, and dataset dramas.
None of this is what makes it into standard academic discussion. Even on Twitter few rarely think or discuss their failures or mysteries in detail. Yet these are the exact discussions that are likely to lead to the next breakthroughs (or fundamental rethinks) of our field.
