Post AW8V2LTRdLCMUeK3Q8 by melissadancey@mastodon.au
(DIR) More posts by melissadancey@mastodon.au
(DIR) Post #AW8JBa4rVaYCwRSfz6 by lauren@mastodon.laurenweinstein.org
2023-05-29T02:22:15Z
0 likes, 0 repeats
One day in the future:Dear AI Chatbot, how can a nuclear reactor excursion be stopped?> Pull out all the control rods is the best way for excursions.Are you sure?> Yes, rods pulled out will be good for excursions.Thanks, we'll try that. Just one more thi> Hello? Hello? Hello? Hello? Hello?
(DIR) Post #AW8RS5V5FIfsfH6mpc by pjfasano@nova.community
2023-05-29T03:54:26Z
0 likes, 0 repeats
@lauren Don’t worry, the reactor will kindly (and rapidly!) disassemble itself!
(DIR) Post #AW8V2LTRdLCMUeK3Q8 by melissadancey@mastodon.au
2023-05-29T04:34:57Z
0 likes, 0 repeats
@lauren 60 minutes Australia had a story on AI last night.
(DIR) Post #AW9UEqrCs5zEVPnR5s by _thegeoff@mastodon.social
2023-05-29T16:00:51Z
0 likes, 1 repeats
@lauren Recently had a very similar discussion with technicians (school, not nuclear!) after somebody offered ChatGPT generated protocols. If it's safety-critical you need to check ChatGPT's work to a level of detail that makes it pointless asking in the first place.
(DIR) Post #AW9WHnHCqB04B4UKFU by kentborg@social.tchncs.de
2023-05-29T16:23:34Z
0 likes, 0 repeats
@lauren I saw a Washington Post headline that seemed to be answering the question: After so many years of trying, why is AI now successful?I didn't have the stomach to click.
(DIR) Post #AW9fqQxqgi9qhU1hQG by jt7d@mastodon.social
2023-05-29T18:10:49Z
0 likes, 0 repeats
@lauren I've been wondering whether ChatGPT has killed its first human yet.