I am hugely excited and energised by discussions of how we can make artificial intelligence safer and more reliable; how we can ground our symbolic requirements both in real-world phenomena and also in the sub-symbolic representations learned by the massively over-parameterised machine learning tools that are currently in vogue; and I am fascinated by how we can use synthetic data and domain randomisation to bridge those divides and achieve that symbol grounding.

Even in technical communities, it's hard to find people who are as engaged, excited and energised by those topics as I am. In 'everyday' life, people either disengage and switch off immediately, or react in a (presumably defensive) mocking and derisive manner.

It's only by spreading our net further, by reaching out to a global community of academics, scientists, engineers and founders that we can build ourselves a more sympathetic and engaged social network, and that we can fully satiate both our social and our intellectual needs.