It is apparently very difficult to put lasers on Si. Si itself can't turn electricity into light (except a few experiments in academic research we've read about), and it is difficult to put other materials onto Si.

Recall that Taylor said at the 2012 AGM that nobody would dare subject Si to a high temperature anneal such as they do in POET. Why not? The only difference I can see is that POET uses GaAs throughout the process, while others try to deposit materials like GaAs, Ge, etc. onto Si. So with POET, anything you do to the material is naturally compatible amongst all devices. Lattice mismatches and different temperature tolerances during the process make it very difficult for GaAs, etc. to be added to an Si substrate.

IBM seems to be the most prolific in the number of different approaches to Si photonics. Not only do they appear in POET's white paper twice, but they also have the "holey optochip" which is post-process method of passing light by physically drilling holes in the chip. Even, IBM hasn't managed to add the laser to the chip.

I don't think HP and POET are working on "the machine" because the work on memristors predates POET's drive towards commercial markets by years. Plus we have no way of contributing to anything related to Si-based photonics.