Scaling laws work capabilities keep keep jumping but our ability to have alignment and safety assurance just be built in seems not there. um maybe never going to be there if that is the case then what do we do so then you got to come outside the model.
The models have advanced to a point where there is a new use case like they're useful in a new way. Not just knowledge retrieval, not just go and write a bunch of code, they are useful, but that requires a new type of harness. Just like when the models got good at code, the harness became very important.
For the first time in human history, we have a mathematical model of all of human knowledge. It's not perfect, but that's essentially what all of the LLMs are.
There might be like sarcasm hidden in the vectors doesn't make it through to the actual tokens and you realize that the AI was joking when it was saying, "I won't turn you into a paperclip." Wink wink.
There's a lot here, including the idea that you'll have an AI model literally talking to itself to figure out how to drive your car, which I find both incredibly interesting and incredibly funny.
