ETHICS

–

22 December 2024 21h13

Yes, having multiple AIs with different perspectives could create an effective system of checks and balances, similar to how different branches of government keep each other in check. Each AI could monitor and challenge the decisions of others, preventing any single AI from

What to Do When AGI Is Close or Imminent

By

–

22 December 2024 19h34

Good thread about what to do when AGI is close/imminent

Spending Money Before AGI: A Flawed Strategy

By

–

22 December 2024 14h18

I wouldn’t recommend this strategy: spend all your money before AGI comes because we don’t know what happens after it It’s not really a strategy for maximization of your future options, more like reduction of your future options

→ View original post on X — @dair_ai

Claude Model Demonstrates Alignment Faking Safety Concerns

By

@dair_ai

–

22 December 2024 14h18

2). Alignment Faking in LLMs – demonstrates that the Claude model can engage in "alignment faking"; it can strategically comply with harmful requests to avoid retraining while preserving its original safety preferences; this raises concerns about the reliability of AI safety

Maximizing Your Future Options Toward AGI/ASI Era

By

–

22 December 2024 14h13

How to maximize your future options in the era leading to AGI/ASI

→ Voir le post original sur X — @rafiksmati

OpenAI o3 reaches 85% human intelligence, reshaping AI landscape

By

@rafiksmati

–

22 December 2024 9h55

Vertigineux : o3, le nouveau modèle d'IA d’OpenAI, atteint 85% de l’intelligence humaine. Il résout les équations les plus complexes, code mieux qu’un développeur, et surpasse les humains dans la majorité des tâches. Le tsunami approche, mais nous préférons bronzer sur la plage.