SF tech lab the "SL5 Task Force" just launched an ambitious new AI datacenter security standard to defend frontier development against nation state attacks. Targeted at securing multi-agent workflows doing large scale automated AI R&D in 2028/2029.
calling on hacker teams to help figure out how to avoid weights theft, loss of control risks & get a defense dominant future as fast as possible
calling on hacker teams to help figure out how to avoid weights theft, loss of control risks & get a defense dominant future as fast as possible
main site: sl5.org
Also available on x: https://x.com/SL5TaskForce/status/2031419965572628624?s=20
reply