exrhizo

TypeSafe employee here, This could change the narrative around AI and safety. The Jev model isn't a reasoning model, it doesn't have hidden thinking tokens. This is how to bring engineering back into the conversation, design your reasoning systems as code, that way they are explainable. Explicit reasoning is the interface for humans to constrain and control, this opens completely different approaches to alignment.

Richard Watson's (Author) YouTube videos on this are good: https://youtu.be/zdmY6q2ZWm8?si=ilcIwg9ghTH7rMBO

Raw milk doesn't seem to me the most anti science thing to me

But I believe the premise that financial interests aren't being challenged

A good reason to have LLM provider swapping built into these things

Rather than banning AI in technical interviews, better to see how the candidates use it and if they can comprehend what the LLM is saying, the quality of their prompts and own thinking.

Thank you sir.

Awesome, made this worth it from my pov

Perhaps it's of use to others, The possibilities for healing were not as diverse then as they are now. Besides the clinical dbt or cbt, there are many coaches and bodyworkers with diverse offerings on how to transform trauma. Somatic practices like Hakomi, breathwork like holotropic. Many many options, all of which can have a contribution to healing.