The case for cybersecurity in AI safety
Two posts criticize what they see as a fixation on alignment among “EA” and “doomer” AI safety advocates, arguing that it sidelines cybersecurity.
TLDR
One post accuses the “EA safety, AI doomer crowd” of portraying security as impossible and alignment as the only solution. Another argues that these advocates should learn cybersecurity and use every available tool to keep AI under control. It accuses them of ignoring basic cybersecurity lapses in favor of vague alignment narratives with no plausible proposal for solving alignment.
The case for cybersecurity in AI safety
Two posts criticize what they see as a fixation on alignment among “EA” and “doomer” AI safety advocates, arguing that it sidelines cybersecurity.