Back in August of 2020 I issued a working paper, GPT-3: Waterloo or Rubicon? Here be Dragons. On the first page, after the cover, but before the abstract and table of contents, I said this:
GPT-3 is a significant achievement.
But I fear the community that has created it may, like other communities have done before – machine translation in the mid-1960s, symbolic computing in the mid-1980s, triumphantly walk over the edge of a cliff and find itself standing proudly in mid-air.
This is not necessary and certainly not inevitable.
But, alas, that seems to be what is going on. And it's not merely that all those resources will have been wasted through, who knows, maybe some folks will wise and fly right, there's the opportunity cost, the good work that could have been done if some of those resources had gone to other word, any other work, just to sample the space of possibilities more widely.
Are we there yet?
No comments:
Post a Comment