Anthropic's new AI Constitution profoundly challenges how we think about, develop, and use artificial intelligence, while also opening up potentially transformative possibilities
The langauge struggle here is real, especially around this idea of AIs having moral character that's neither human nor machine in any traditonal sense. I've been working with these models for a few months now, and the way they shift between seeming totally predictable and then completely surprising me still catches me off guard. The constitutional approach seems like a practical way to navigate this uncertainty without pretending we have it all figured out.
Great piece of writing! Very insightful. Couldn’t help but notice the uncanny resemblance of developing these new frontier AI model to raising a human kid!
The langauge struggle here is real, especially around this idea of AIs having moral character that's neither human nor machine in any traditonal sense. I've been working with these models for a few months now, and the way they shift between seeming totally predictable and then completely surprising me still catches me off guard. The constitutional approach seems like a practical way to navigate this uncertainty without pretending we have it all figured out.
Great piece of writing! Very insightful. Couldn’t help but notice the uncanny resemblance of developing these new frontier AI model to raising a human kid!
Anthropic going full cult mode.