At the end of Thanksgiving holidays, I finally finished the piece on reward hacking. Not an easy one to write, phew. Reward hacking occurs when an RL agent exploits flaws in the reward function or env to maximize rewards without learning the intended behavior. This is imo a
ETHICS
-
Samsung See-Through Truck: Safety Innovation or Hazard?
By
–
🚛 Samsung's "See-Through" Truck: Brilliant or Dangerous?
— Pascal Bornet (@pascal_bornet) 2 décembre 2024
A truck with front cameras streaming road view to rear LED screens, helping drivers decide when it's safe to overtake. While it aims to prevent accidents, the question remains: does this tech improve safety or create more… pic.twitter.com/N87qPoKN5FSamsung's "See-Through" Truck: Brilliant or Dangerous? A truck with front cameras streaming road view to rear LED screens, helping drivers decide when it's safe to overtake. While it aims to prevent accidents, the question remains: does this tech improve safety or create more
-
AI thinks differently: alien intelligence reshapes human capability
By
–
AI is an alien intelligence. Much of what is easy for us is hard for AI and vice-versa. And that's (mostly) a good thing:
-
LLM-Based Generative Agents Simulate 1000 Real Individuals
By
–
9). Generative Agent Simulations of 1,000 People – introduces a new agent architecture that uses LLMs to create behavioral simulations of real individuals, achieving 85% accuracy in replicating human responses on the General Social Survey and reducing demographic biases compared
-
ChatGPT Language Games: Measuring Bullshit in LLM Outputs
By
–
10). Measuring Bullshit in Language Games Played by ChatGPT – proposes that LLM-based chatbots play the ‘language game of bullshit’; by asking ChatGPT to generate scientific articles on topics where it has no knowledge or competence, the authors were able to provide a reference
-
Survey on LLM-as-a-Judge: Building Reliable Evaluation Systems
By
–
7). Survey on LLM-as-a-Judge – provides a comprehensive survey of LLM-as-a-Judge, including a deeper discussion on how to build reliable LLM-as-a-Judge systems.
-
Decentralized Open Source AGI Development Path Announced
By
–
This is the first step on the path to fully decentralized and open source AGI at scale, ensuring humanity won't be dependent on big corporations https://t.co/ZCUav9LvUm
— Marek Rosa | European🇪🇺 | South African🇿🇦 (@marek_rosa) 30 novembre 2024This is the first step on the path to fully decentralized and open source AGI at scale, ensuring humanity won't be dependent on big corporations
-
Next Decade’s Big Tech: DAOs, Biotech, AI, and Quantum
By
–
Next Decade's, Next #BigTech ? DAOs for trustless governance Synthetic Bio for industry Carbon Markets for credits Ethical AI deployment Space infrastructure Quantum computing services Personalized medicine Metaverse ecosystems #MegaTrends
-
Primordial Wisdom Principles for Effective Leadership
By
–
How to Apply Primordial Wisdom to #Leadership Vision: Chart a course with clarity and purpose Compassion: Lead and grow with understanding Presence: Be fully present, every moment Equanimity: Stay calm and decide with wisdom Authenticity: Lead by living
-
Enterprise Leadership Priorities: ESG, Tech, Supply Chain, Workforce
By
–
Large Enterprise #Leadership Agenda Focus Areas from Recent 10-Q Filings: ESG & sustainability focus continues Tech investments intended to drive growth & efficiency Supply chain strategies evolve in an unsettled world Employee welfare & safety prioritized