#EffectiveAltruism folx should work on their reading comprehension.

I got an #AI in 2024 retrospective from 80,000 Hours, an Effective Ventures project (related to EA). In it, they mention that "the o1 language model [developed by OpenAI ][...] has the ability to deliberate about its answers before responding."

The OpenAI o1 release says: "We introduce deliberative alignment, a training paradigm that directly teaches reasoning LLMs [...] safety specifications..."

Quite the leap of faith...