Over the last few sessions there has been interloping on the subject
of artificial intelligence from the perspective of engineers and
computer scientists. We could approach this subject by considering the
redeeming qualities of an AI engineer. As a case in point: When
seeking to employ such a person, Bungie studios voiced preference for
a "megalomaniac genius with severe god-complex". Unfortunately,
reflection on the demographic committed to AI won't yield much on the
materiality of member's labours. I propose we approach AI by
concentrating on two opposing doctrines/frameworks in machine
learning. In this session, we could concentrate on connectionism --
that something I chose not to choose -- and one alternative. Let's
call the alternative relationism. The latter is the doctrine that
relations, and hence objects, have a necessary role in realising
AI. The former, and by far the most ubiquitous, is that artificial
neural networks will suffice. I shall not reject outright the torrent
of connectionist propaganda. Rather I seek pragmatism is something
more practicable. After all, flying a helicopter upside-down is only
one trick in-this-'ere multanimous circus.