To avoid crowds, visit areas that are less crowded.

GPT-4 perk: even though it still hallucinates academic papers, these papers tend to be much closer to ones that actually exist. Often the titles are right, authors are mostly right, and venues are wrong (easily searchable).

Sometimes we are not sure if our security research is ethical. I am maintaining learning materials for "ethics in computer security research", including ethics principles (both general and domain-specific), guidelines, review committees, research papers, and more. 1/2 #EthiCS

Doe v. GitHub out (copilot generative AI exploitive?): Copyright, DMCA claims based on removal of copyright management information (CMI), and license breach claims were not dismissed; everything else dismissed with leave to amend (save civil conspiracy). storage.courtlistener.com/recap/gov.usco…

#PurdueCS paper at IEEE S&P One Key to Rule Them All: Secure Group Pairing for Heterogeneous IoT Devices Habiba Farrukh, Muslum Ozgur Ozmen, Faik Kerem Ors, Berkay Celik

ANNOUNCEMENT: The U.S. National Science Foundation Secure and Trustworthy Cyberspace (SaTC) 2.0 Vision Statement is publicly available on arXiv! Check it out here: arxiv.org/abs/2308.00623

Had an AMAZING time today presenting DScope at the USENIX Security Symposium! For those who missed it: never fear! You can find an extended version of the talk RIGHT NOW at dscope.org/sec23.

ChatGPT OpenAI roon Ethan Mollick thebes Wild result. gpt-4-turbo over the API produces (statistically significant) shorter completions when it "thinks" its December vs. when it thinks its May (as determined by the date in the system prompt). I took the same exact prompt

Thanks, Patrick McDaniel! I couldn't think of a better conclusion to the academic year than celebrating this award with you and Paul. Onwards and upwards!