We just turned on a 24-hour autonomous AI news channel this morning, live now: https://x.com/i/broadcasts/1nJOLQEqZZrxR. We are pulling from a mainstream newsfeed so will cover news as usual, we also are including AI-generated interviews with real guests.
For context, we’re running the whole thing directly on Claude on a single machine, and the experiment has cost us about $20K so far.
We’ve been experimenting with more real-time use cases and wanted to see what happens when avatar video runs continuously. The model takes audio (and optionally a photo/video reference) and returns back your avatar video.
We’ll have it up for at least the next 24 hours, let us know what you think (haters welcomed)!
Hi HN. I'm a PM at Mirage (we make Captions).
We just turned on a 24-hour autonomous AI news channel this morning, live now: https://x.com/i/broadcasts/1nJOLQEqZZrxR. We are pulling from a mainstream newsfeed so will cover news as usual, we also are including AI-generated interviews with real guests.
We built it to launch our Avatar X API: https://fal.ai/models/mirage-api/avatar-x/reference-to-video
For context, we’re running the whole thing directly on Claude on a single machine, and the experiment has cost us about $20K so far.
We’ve been experimenting with more real-time use cases and wanted to see what happens when avatar video runs continuously. The model takes audio (and optionally a photo/video reference) and returns back your avatar video.
We’ll have it up for at least the next 24 hours, let us know what you think (haters welcomed)!