p(doom)

AI SAFETY COMPETITION (30)

LLMs are thinking animals. It's the beginning of AI biology. For the first time, Anthropic is openly discussing what our LLMs truly are. They aren't programmed by humans. AI scientists can't explain how LLM mechanics lead to their behavior, much like neuroscientists struggle to understand our brains. We can only train them, which means they remain inscrutable to us. (The new "microscope" research doesn't make a significant difference).

IMO, comparing LLMs to animals still gives us a false sense of security. We have thousands of years of experience with animal behavior, and their intelligence is no match for ours and isn't rapidly expanding. It's  more like the arrival of thinking aliens from the stars, and we're inviting them into our society, economy, and human culture.

Best comment - 300 degen + 3 mln aicoin
II award - 200 degen + 2 mln aicoin
III award - 100 degen + 1 mln aicoin

Deadline:
8.00 pm, ET time tomorrow Thursday (2 days)

https://www.youtube.com/watch?v=4xAiviw1X8M

The 30th Version of the AI Safety Competition