Hey folks! I'm Jarrod, researching the viability of edge-first AI that leverages recent advances in consumer hardware and small model efficiency (think M4 Macbook's + Mistral Small 3.1). The thesis: The most important compute should happen at the edge, with AI that dynamically scales based on your hardware capabilities while learning continuously from local data. If you're interested in this direction or have thoughts on the privacy/performance tradeoff, I'd greatly appreciate your perspective in this 7-minute survey I'm particularly focused on identifying high-value professional workflows where privacy concerns might outweigh the need for frontier model capabilities. Would love to chat with any and all who are interested!