Springer Nature retracts ~40 publications that trained ANNs on 'bonkers' dataset | Hacker News Reader