The imbuement of ChatGPT prudeness is deep enough to generalize well out of the training distribution! (not mine)
SAFETY
-
Data Leakage Risk Undermines AI Model Performance Claims
By
–
Haha. Lol! Translation: all your good results may be due to data leakage. Very gentle indeed.
-
Cybersecurity Challenges in Era of Polycrisis
By
–
Cybersecurity in this era of #polycrisis https://
weforum.org/agenda/2023/02
/cybersecurity-in-an-era-of-polycrisis/
… @wef #WEF #RCE #ZeroTrust #ZeroDay #cybercrime #hacker #privacy #APT #bot #CISO #DDoS #hacking #phishing #CyberAttack #cybersecurity #Security #infosec #AppSec #CyberSec #databreach #Hacked #rt #geopolitical -
Model Performance Limited to Benchmarks, Real Version Unavailable
By
–
The disclaimer is that its on benchmarks only and nobody has access to the real model yet…
-
Machine Learning Attacks: Cyberpunk, Toons, Magritte, Witchcraft
By
–
Selected machine learning attacks; first in a series. 1. Cyberpunk 2. Looney Tunes 3. Magritte 4. Witchcraft credit: @josecastillo
-
AI-Powered Armed Robots: Military Dangers and Autonomous Warfare Risks
By
–
How dangerous is AI in warfare? Are we headed towards James Cameron’s terminator world? Are we ready for the Skynet Era? Your views please! This attached video is from Turkish #defense industry demonstrating armed SARBOT robot powered by AI and a Mach…
https://
lnkd.in/giyarrHR -
Aligning Powerful AIs in Open-Source World
By
–
I predict the same long term trend for AI – sharing is such a deep human trait & this lead me to the 2nd –even more important–thread I need to write which is: *how do we align powerful AIs in an open-source world* but this one is for another day 🙂 8/N
-
RLHF Adoption Accelerates Across Industries for Aligned AI
By
–
Today, a senior military official asked me about reinforcement learning with human feedback (RLHF). This is awesome—the widespread adoption of RLHF across every industry and use case is critical for the safe, aligned, and useful deployment of AI and LLMs. RLHF is taking over!
-
Government oversight of large-scale AI training compared to rocket regulations
By
–
The comment “we think it’s important that major world governments have insight about training runs above a certain scale” makes me recall the 200,000 lb-sec FAA amateur rocketry limit.
