It’s just sugar with a teensy bit of the natural brown color from unrefined molasses left in it. I don’t find your observation that it takes 5 or 10 times as much of it to sweeten something to be true for me whatsoever, it’s almost exactly the same, and leaves me wondering if perhaps you also find that today’s low-flow toilets need to be flushed dozens of times to work, or that you turn on modern showers and just a tiny trickle comes out :)
So they slapped some reinforcement learning on top of their LLM and are claiming that gives it “reasoning capabilities”? Or am I missing something?