It feels like a lot of AI discussion is still cloud-first, but some of the most interesting shifts seem to be happening at the edge.
A few areas that seem especially important:
- autonomy and robotics
- low-power always-on vision systems
- private local LLMs and on-device inference
- bandwidth-constrained industrial deployments
Curious how people here see it:
Over the next few years, where do you think edge AI matters most, and which hardware/software stacks actually win in practice?
[link] [comments]