A working prototype. A benchmark we’ve never seen. A weird research direction. We answer every pitch within 48 hours — yes or no, with a reason.
We don’t need a deck. Two sentences and a benchmark are enough to get a real answer in 48 hours.
We don’t need polish. We need to see the thing run. A loom video and a benchmark beats a deck every time.
What does inference actually cost you? At what scale do unit economics turn? If you can answer in numbers, we’ll move fast.
We’ve seen the obvious wedges. Show us the one that sounds slightly insane on the first read.