I Launched the AI Safety Clock. Here’s What It Tells Us About Existential Risks