Anthropic researchers: AI models can be trained to deceive and the most commonly used AI safety techniques had little to no effect on the deceptive behaviors
[images] Abraham Samma / @abesamma@toolsforthought.social : Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training — This is some sci-fi stuff right here (even if unsurprising)...
Source: a leaked April 2023 internal document by Luke Sernau, a senior software engineer at Google, argues that open-source AI will outcompete Google and OpenAI
warning the company doesn't have one in A.I. John Battelle / John Battelle's Search Blog : Google and Commoditization: Anyone Need a BackRub? Agne Cimermanaite / Metaverse Post : Leaked Google Documen...
2020 political campaigns secured emails with physical security keys from a nonprofit working with Google and Microsoft; source says Biden's campaign used keys
Jordan Novet / CNBC : Tweets: @netik , @codinghorror , and @eladgil Tweets: John Adams / @netik : U2F and FIDO worked to stop email attacks this year. Stop using SMS 2FA. Let's get this in the hands ...
2020 political campaigns secured emails with physical security keys from a nonprofit working with Google and Microsoft; source says Biden's campaign used keys
- Keys from the likes of Google and Yubico became a hit among campaign workers this year, including in Joe Biden's campaign. Tweets: @eladgil , @martin_casado , @elizabeth_joh , @boblord , and @jordan...