Natural emergent misalignment from reward hacking in production rl [pdf]assets.anthropic.com3 points·by neapolisbeach·8 maanden geleden·0 comments
Whisper leak: a side-channel attack on large language modelsarxiv.org3 points·by neapolisbeach·8 maanden geleden·0 comments
Google Private AI Compute [pdf]services.google.com1 points·by neapolisbeach·8 maanden geleden·0 comments
Boeing whistleblower found dead in USbbc.com1,335 points·by neapolisbeach·2 jaar geleden·556 comments
Finding a command injection vulnerability in my smart thermostat (CVE-2023-4212)hmcguinn.com2 points·by neapolisbeach·3 jaar geleden·0 comments
Vm2 discontinued due to unfixable security issuesgithub.com2 points·by neapolisbeach·3 jaar geleden·0 comments
neapolisbeach·vorig jaar·discussNsjail, firecracker, gVisor, or v8 isolates are all good options with different tradeoffs
neapolisbeach·3 jaar geleden·discussIf you're interested in time-lock puzzles, drand [1] achieves this but through randomness instead of computational effort.[1] https://drand.love/