Dan Hendryks thinks Aschenbrenner is overestimating the rate of improvement in algorithmic efficiency (https://twitter.com/DanHendrycks/status/1798177460028346722) and that at the actual rate we shouldn’t expect an intelligence “explosion” (https://twitter.com/DanHendrycks/status/1798390576922178004). (I think “explosion” is a bit vague here, and it would be more interesting to get a sketch from Hendrycks of the actual rate of improvement he predicts. But I guess that wouldn’t fit in one tweet.)
Dan Hendryks thinks Aschenbrenner is overestimating the rate of improvement in algorithmic efficiency (https://twitter.com/DanHendrycks/status/1798177460028346722) and that at the actual rate we shouldn’t expect an intelligence “explosion” (https://twitter.com/DanHendrycks/status/1798390576922178004). (I think “explosion” is a bit vague here, and it would be more interesting to get a sketch from Hendrycks of the actual rate of improvement he predicts. But I guess that wouldn’t fit in one tweet.)