Wretch in the machine (3)
Somehow also for January
This is the third in a three-part series. Prior installments:
IX. On stochastic parrots
A great deal of ink has been spilled over the Bender et al. 2021 paper describing LLMs as stochastic—meaning random—repeaters of bits of human language, without comprehension. The central point of the paper is straight out of Weizenbaum:
LMs are not performing natural l…