open ai is able to eat solve multiple problems, pass the bar exam do well in the sat and do coding. these things are remarkable but the potential for misuse is there. we need to think through that and they are still problems with the systems. they are not foolproof, they hallucinate facts and make up false information sometimes and it is hard to tell the real from the falls. they are going to be ruled out in a hull load of different contexts in our world and is going to be a challenge for governments, scholars and regulators to keep up with it. the open letter didn t cause for pause and all ai research, it introduced the idea of a clause on training the very biggest models, so on the models even bigger than the
technology and businesses. was still advance their ai - technology and businesses. so, a pause technology and businesses. so, a pause doesn t technology and businesses. so, a pause doesn t make technology and businesses. so, a pause doesn t make sense - technology and businesses. so, a pause doesn t make sense and i technology and businesses. so, al pause doesn t make sense and i m surprised pause doesn t make sense and i m surprised it pause doesn t make sense and i m surprised it on pause doesn t make sense and i m surprised it on musk pause doesn t make sense and i m surprised it on musk is pause doesn t make sense and i m surprised it on musk is involved i pause doesn t make sense and i ml surprised it on musk is involved but you have surprised it on musk is involved but you have this surprised it on musk is involved but you have this next surprised it on musk is involved but you have this next several- surprised it on musk is involved but you have this next sev
the idea of a pause now and having the idea of a pause now and having the public conversation around it is a good idea because if we want in a year or two, the idea would have been explored and debated. i would say that we have a wealth of opportunities to explore the applications of the latest cutting edge research without slowing down those applications. we need some time to think through how we apply this to coding, to checkbox that people use, two masts and to many different things to stop if we want to have public interest in the technologies that were deploying, we need to know that there safe and reliable and that their predictable. that they don t go off the rails. we ve already had several large modules basically be recalled in the last year, meta recalled their calexico model because it was making up calexico model because it was making up scientific papers and microsoft had some embarrassing situations and