"bounded regret algorithm" in @yalereview 👾💜
this poem came out of my sense that we can learn a lot about ourselves from how we teach machines to learn
new paper draft: 𝗰𝗼𝗻𝘁𝗲𝘅𝘁𝘂𝗮𝗹𝗹𝘆 𝗽𝗿𝗶𝘃𝗮𝘁𝗲 𝗺𝗲𝗰𝗵𝗮𝗻𝗶𝘀𝗺𝘀!
in this paper, andy haupt (@andreas_h0wpt) and I offer a framework for comparing the privacy of different mechanisms.
arxiv.org/abs/2112.10812
Advanced AI may make it easier for bad actors to deceive others online. In “Abstracts,” we discuss one solution being explored by a multidisciplinary team: a credential that allows people to show they’re not bots w/o sharing identifying information. msft.it/6014lsOim
a short podcast on a recent research collaboration with @Microsoft and @OpenAI!
generative AI poses serious threats to the privacy and authenticity of our digital communications
we offer a new framework for thinking about and responding to these threats
In the “Abstracts” podcast series, project manager Shrey Jain & Harvard fellow Zoë Hitzig talk contextual confidence, a framework for understanding & addressing the challenges generative AI poses to identifying & protecting the context of communications: bit.ly/3R9BnSt