Overall, I’m glad that the idea of emergent abilities is being discussed more. I’m excited about work that would enable us to predict emergent behavior, since emergence includes risks as well as abilities. I’d love to discuss more with you on twitter or at the next conference!
SAFETY
-
Digital Intelligence Overtaking Humans: 5-20 Year Timeline Prediction
By
–
I now predict 5 to 20 years but without much confidence. We live in very uncertain times. It's possible that I am totally wrong about digital intelligence overtaking us. Nobody really knows which is why we should worry now.
-
Improving AI Sharpness and Prompt Interpretation Accuracy
By
–
Sharpness, less rogue elements and more accurate reading of prompts.
-
Pi Chatbot Appears More Advanced Than GPT-3, Not OpenAI-Based
By
–
I don't believe it's based on OpenAI's models also it seems much more advanced than GPT-3 at defending against jailbreaks and at performing creative tasks https://
forbes.com/sites/alexkonr
ad/2023/05/02/inflection-ai-ex-deepmind-launches-pi-chatbot/?sh=6096129b3d6d
… -
Microsoft CSO Discusses AI Risks and Six-Month Pause Letter
By
–
Microsoft’s Chief Scientific Officer weighs in on the dangers of A.I. and the open letter for a 6-month pause https://
bit.ly/3VjD7Je -
AI Future Impact: Regulation Needs Clarity on Risk Scenarios
By
–
I'm not sure the future of AI can be put in the same category as the introduction of planes and cars. Plus, there seems to be a general consensus among AI authorities on the need for regulations. I'm curious about what type of incident needs to happen in order to trigger serious
-
SSL Models Privacy Risks from Unintended Image Memorization
By
–
Do SSL models have Déjà Vu? Meta AI researchers uncovered previously unknown privacy risks resulting from unintended memorization of image-specific information in SSL models + suggest practical mitigation strategies for new models. Paper https://
bit.ly/413zhoZ -
AI Pioneer Leaves Google Warning of Future Dangers
By
–
‘The Godfather of #AI’ Leaves Google and Warns of Danger Ahead https://
nytimes.com/2023/05/01/tec
hnology/ai-google-chatbot-engineer-quits-hinton.html
… @nytimes #ArtificialIntelligence #MachineLearning #DeepLearning @ahier @gvalan @avrohomg @anijov @CatherineAdenle @Shi4Tech @Damien_CABADI @enilev @FmFrancois @data_nerd @JAdP @tunguz -
Maximizing Civilization’s Net Happiness Over Time
By
–
There will always be sadness. What matters – I think – is maximizing cumulative civilizational net happiness over time.
-
Google CEO admits lack of understanding of AI system behavior
By
–
Google CEO says that they don’t fully understand their own #AI system after it did things it wasn’t programmed to do: