After hundreds of user interviews, one demo day and a lot of honest feedback from Coco, we're pivoting. What we learned: clowns don't want a cloud-connected, context-aware, AI-native honk. They want a rubber bulb on a brass horn that works without wifi. So starting today, Squeakly makes a physical horn. It honks when you squeeze it. Latency is zero. It ships in two weeks. Thank you to our investors for their support, and to the 140 years of prior art we had, in retrospect, been competing with.

We live-demoed Squeakly on the ClownCon Demo Day stage to 600 clowns. The conference wifi went down two seconds in. Contextual honking requires the cloud. Our horn produced nothing for eleven seconds. I squeezed it harder. It produced nothing, harder. Then Mimi Marceau, front row, mimed a honk. The room lost it. Biggest laugh of the day. I've watched the recording 40 times and I believe it's the strongest product-market-fit signal we've ever had, though my co-founder disagrees about which product. Grateful for every lesson. Rebuilding the demo with an offline mode.

Great question! We're pre-latency right now. Fallback is on the roadmap. Would love to grab 20 minutes; honestly, we'd love your thoughts on all of it.

Some news: Squeakly has raised a $2.1M pre-seed to build the AI-native honk layer for the internet, led by Big Top Capital with participation from angels including Bozo D. Ringwald. I've also joined clown.business. Founders should be where their users are, and our users are clowns. The problem: the honk hasn't fundamentally changed in 140 years. A rubber bulb, a brass horn, a squeeze. Squeakly uses a large language model to generate the right honk for the right moment, in real time. We call it contextual honking. We're hiring founding engineers. Must love honks. Must be comfortable with ambiguity, including very loud ambiguity.