Don't fear superintelligence, fear not having it.
SAFETY
-
Technology Winter Risk: Microsoft Outage as Warning Sign
By
–
La panne mondiale de Microsoft n'est qu'un apéritif de la prochaine grande crise qui surviendra. La technologie apporte le progrès, mais elle nous rend aussi plus vulnérables. Je vous partage ces pages de mon livre "Le Nouveau Temps" où je parle du risque "d'hiver technologique".
-
Domino Cloud Achieves FIPS Certification Towards FedRAMP Authorization
By
–
Exciting news! Domino Cloud is now FIPS-certified, a major step towards FedRAMP High Authorization. This ensures top-tier AI security & data protection for our federal customers. Learn more: https://
domino.buzz/4f55yUF #AIGovernance #FIPS #FedRAMP #DataSecurity #AI -
AI Hallucinations: Logical Route Planning Failures
By
–
When you ask a generative AI engine to generate the most logical route from Iceland to Ireland: #technlogy #tech #ai #artificialintelligence #hallucinations
-
Discrete Gaussian Differential Privacy Wins Caspar Bowden Award
By
–
Excited to share that "The Discrete Gaussian for Differential Privacy," joint work with Clément Canonne (@ccanonne_) and Thomas Steinke (
@shortstein
), was selected for the 2024 Caspar Bowden PET Award! Honoured for this work to be recognized. 1/2 -
Repeating Past AI Mistakes: Critical Lessons for Better Outcomes
By
–
Why do we keep making the same AI mistakes? Our #AITodayPodcast dives deep into this critical issue. Learn from past errors and drive better AI outcomes. Tune in: https://
cognilytica.com/are-we-still-r
epeating-the-same-mistakes-with-ai-ai-today-podcast/
… #AI #AIWinters -
Fine-Tuned Models Leak Private Pretraining Data Information
By
–
Nice talk by John Abascal on "TMI! Finetuned Models Leak Private Information from their Pretraining Data," a joint work with @the_stanleywu
, @AlinaMOprea
, and @thejonullman
. They show that fine-tuning large models can nonetheless leak info about their training data. -
Defining Positive Valence and Robot Well-being Through Homeostasis
By
–
I looked at how we could define positive and negative valence for robots, linking physical distress to a "bad" gut feeling (e.g. low battery, overheating motors) and flourishing homeostasis to a "good" gut feeling.
-
Smart Bullets with AI Technology to Distinguish Targets
By
–
Should a bullet have a brain? Yes, so it hits the bad guys and not the good ones.
-
AI Technology for First Response and Crime Prevention
By
–
We develop our technology for first response, crime prevention, SAR. It is being used by a non-profit, community-based crime fighting organization, and we don't supply it to the military or spies. You can learn more about our mission here, we are transparent about our work and