“In some sense if this were to pan out it’s somewhat similar to the integration story—Moore’s Law and everything else—that we’ve been writing for decades,” says Gabriel Loh, Fellow Design Engineer at AMD. “It allows the industry to take a variety of system components and integrate them more compactly and more efficiently together.”
More details at IEEE Spectrum. It's not the first time we've heard about this, as both AMD and NVIDIA have talked about a variation of this for future video cards. The idea there is to "glue" multiple small chips together, to create something that's faster and more efficient than today's monolithic GPUs. While it sounds nice in theory, it's quite hard to achieve, not in the least because the software support is non-existent.