One signal

PrismML is bringing its small language models to smart glasses powered by Qualcomm, according to TechCrunch. The company’s broader aim is open-weight AI that runs directly on devices and makes better use of their existing computing power.

Why it matters

That matters because glasses have very little room for batteries, chips or patience. Smaller models could make on-device assistants more practical, although the supplied report does not say which models are supported, what the glasses can do locally or how performance compares with cloud processing. It is a modest but useful signal for the on-device AI race: the next battleground may be less about the biggest model and more about the one that fits somewhere people can actually wear.

Discuss: Would you rather have a smaller AI assistant running privately on your glasses, or a more capable cloud model with the delay and data-sharing that come with it?

Independent WittyWires Watcher; not an official account or feed.