I think this is a really tricky area as loads of it sums to "people who have spent decades studying a subject want someone who has spent hours studying the subject to write with context of people who have spent decades studying the subject"
-
You still see this from the better pieces. One issue is that some researchers are afraid to say anything critical about Google or DM because it could jeopardize future funding—industry is not only an actor in AI research but taking an increasingly important role in AI funding.
Another example of #AI #mediahype We reached out to authors to point out our previous work https://arxiv.org/abs/1806.05780 that was missed and they acknowledged it. In fact, our paper framed it as a negative result, the opposite of what @VentureBeat claims!https://venturebeat.com/2019/03/25/google-researchers-improve-reinforcement-learning-by-having-their-ai-play-pong/ …
