Optoelectronic Intelligence - Jeffrey Shainline
This paper sets out an architecture that appears to be both scalable to brain scale neuron counts and connectivity. To achieve these scales they use a combination of superconducting loops and optical interconnects, which can mirror the behaviour of collections of human neurons. The proposal still faces many challenges in scaling in areas such as photon loss, however on paper and early results their approach seems plausible. A human scale and beyond brain capable of processing 4-5 orders of magnitude faster, at equivalent energy costs, is truly sci-fi.
Why I am optimistic about the silicon-photonic route to quantum computing - Terry Rudolph
I've found a lot of Terry Rudolph's papers enjoyable to read, with touches of humour often mixed in. In this paper he is especially light hearted, making a case for the photonic measurement based approach to quantum computing. This was one of my favorite introductions to this approach and definitely convinced me.
Neuroscience-Inspired Artificial Intelligence - Demis Hassabis
With the current rate of progress in AI it can be easy to forget the root of a lot of the ideas. In this paper the authors give a detailed history of the co-evolution of the two fields. The paper makes a good case for continued cross-pollination, not just for the sake of advancing AI but also because it may "yield insights into some of the deepest and the most enduring mysteries of the mind".
