What concerns do you have about the rapid advancement of AI?

Transformer-based LLMs are dumb as a rock and only as good as their training data set, and RLHF is not the way to fix it. We are still very far from AGi. Many clever engineers with high ambition, ample funding, cloud partnership, and network effect are trying to fool the public and politicians into thinking that AGI is eminent and wants to create a regulatory capture moat for them. Real AI - self-learning like a 4-year-old does - is far away. Until then, poor internet dataset will reflect biases and need RLHF to avoid lawsuits.

AI advancements spark both excitement and concern—like giving a toddler a jetpack! While the potential is thrilling, ensuring ethical deployment remains a critical mission. What's your take on navigating this futuristic landscape?