Odd Lots: on‑device AI still matters

- Odd Lots emphasized that even as cloud models scale, on‑device inference and local models remain strategically important for privacy, latency, and differentiated UX. - The episode cited hardware‑software integration advantages and argued that cheaper models push competitive value toward apps, data, and trusted device execution. - That framing suggests firms with secure edge hardware, unified memory, or tight OS integration may capture disproportionate long‑run value. (youtube.com)