The Shape of Inference #2

​We’re hosting a dinner for people building AI products and infrastructure, working with models, investing in the space, or simply curious about what happens between a prompt and a useful answer.

​There’s a lot behind that answer. A team has to decide which model to use, where to run it, how fast it needs to respond, and which tradeoffs actually matter to the person on the other end. The choices that look small in a demo can feel very different once real people start using a product.

​Over dinner, we’ll swap notes on questions like:

• ​where inference is already working well, and where it still falls short,

• ​how teams balance speed, quality, and cost,

• ​which hardware and deployment choices have made a difference in practice,

• ​and what people would rethink if they were building the stack again today.

​The goal is simple: good food, candid stories, and a few unexpected connections across models, infrastructure, products, and investing.

​Come for dinner, stay for what happens after the prompt.

​Hosted by MoE Capital and ETNA Labs.