This crowdfunded $109 hub wants to pool the RAM in your old PCs, Macs, and phones to run bigger AI models — but the engine underneath is software you can already get for free (www.techradar.com)

TradeMAV has launched a crowdfunding campaign for its RAMDeck hub, priced at $109 for early backers, which aims to pool the RAM from various devices—such as PCs, Macs, and Android phones—within a home network. This innovative approach comes at a time when DDR5 memory prices have skyrocketed, making the acquisition of additional RAM financially burdensome. The RAMDeck leverages llama.cpp's RPC backend, a free and open-source software, allowing larger AI models to operate across pooled memory instead of relying on the hardware of a single device. While the RAMDeck's framework prioritizes accessibility and convenience by enabling a diverse range of devices to collaborate, it’s essential to note that it does not merge memory at the hardware level; instead, it splits model layers among connected devices, passing only intermediate results over the network. Though potential users are encouraged to explore DIY options using llama.cpp and its competitors like Exo Labs’ exo, TradeMAV's offering distinguishes itself by simplifying setup, with automatic discovery and a user-friendly dashboard. However, performance trade-offs exist; pooled configurations may deliver slower inference speeds, which could temper interest from enthusiasts seeking high efficiency.