Sheep hate the sheepdogs until the wolves are at the door.
SAFETY
-
AKOrN: Kuramoto Oscillators Replace Threshold Units in Neural Networks
By
–
AKOrN is here: https://
arxiv.org/pdf/2410.13821.
It replaces threshold units with generalized Kuramoto oscillators. It dynamically binds neurons, generates waves of activations, is adversarially robust, calibrated for uncertainty, and can reason. If you ask me: this is the next big thing! -
AI paradoxes: Nobel warnings, Google antitrust, robotaxi advances
By
–
𝗜𝗔 𝗽𝗮𝗿𝗮𝗱𝗼𝘅𝗮𝗹𝗲
C’était la semaine des paradoxes ! #Google : deux Nobels et un démantèlement Deux Nobels alertent sur les dangers de l’IA Robotaxi et Optimus : à quoi joue Elon Musk ? Un label IA d’intérêt public Retour sur le #DigitalTourIMA de -
Political Irresponsibility Threatens Future Generations
By
–
C’est triste à pleurer La classe politique est irresponsable Nos enfants nous haïront !
-
AI Technical Debt Webinar: Financial Services Insights
By
–
Don't miss out on our webinar coming up in just 4 days! Join us for an exclusive webinar on #AITechnicalDebt with insights from Vinay Kumar, CEO of http://
Arya.ai, & Alban Bellenger, Head of Strategy at Aurionpro. Tuesday, October 22nd https://
xai.arya.ai/events/ai-tech
nical-debt-td-in-financial-services-industries-fsis-research-insights-risks-and-best-practices
… -
Aligning Large Language Models with Human Values and Intentions
By
–
align any modality large models (any-to-any models), including LLMs, VLMs, and others, with human intentions and values. More details about the definition and milestones of alignment for Large Models can be found in AI Alignmen
-
AI Safety and Regulation jobs likely filled by HR class
By
–
Sad reality that most who will end up working in “AI Safety and Regulation” will be the HR class.
-
Glasses unmask AI: Scammer realizes she’s talking to ChatGPT.
By
–
Il suffit de parler de lunettes pour démasquer une IA ?
— Defend Intelligence (Anis Ayari) (@DFintelligence) 17 octobre 2024
Dans cette deuxième partie, la démarcheuse téléphonique se rend compte qu’elle parle à ChatGPT à la suite d’un message de modération et essaye de lui faire avouer…
OpenAI a mis tout un tas de garde-fous (parfois ça en… pic.twitter.com/68iFtEc4oTJust talking about glasses is enough to unmask an AI? In this second part, the phone scammer realizes she's talking to ChatGPT following a moderation message and tries to get it to admit it… OpenAI has put in place a whole bunch of safeguards (sometimes it even becomes
-
Advanced Driver Assistance Systems Save Lives on Roads
By
–
Los sistemas avanzados de asistencia a la conducción pueden salvar tu vida en la carretera.
— Juan Merodio (@juanmerodio) 17 octobre 2024
Quedarse dormido o sufrir un desvanecimiento al volante supone un alto riesgo de accidente, pero la tecnología que implementan los vehículos más modernos pueden evitar una tragedia. Son… pic.twitter.com/dyqLjqYFKJLos sistemas avanzados de asistencia a la conducción pueden salvar tu vida en la carretera. Quedarse dormido o sufrir un desvanecimiento al volante supone un alto riesgo de accidente, pero la tecnología que implementan los vehículos más modernos pueden evitar una tragedia. Son
-
Bad if condition for water filter alert
By
–
It's like giving a bad condition to an if condition to alert on water filter control. It's not the AI's fault. It's bad settings and testing. The danger is releasing AI systems without knowing how to design and test them. But