Arvind Narayanan and DeepMind's Séb Krier argue shifting AGI goalposts represent a 'false summit' of inadequate benchmarks · Digg