An experiment with Grok

From I don't want to read a Clooney-centric account of how the Democrats allowed Joe Biden to win the nomination and then swiped it from him. by Anne Althouse.  

She is referring to a New Yorker article by Jack Tapper which excerpts material from his new book which focuses on the 2024 Democratic primary process and the later replacement of Biden with Harris.  She is highly critical of Tapper's analysis and writing and she uses Grok to develop the critique.  

She has five prompts to Grok:

1. Summarize this article: https://www.newyorker.com/news/the-political-scene/how-joe-biden-handed-the-presidency-to-donald-trump

2. I skimmed it and it seemed so Clooney-focused it put me off

3. I'd like a critical perspective on the piece that is skeptical of this interest in Clooney. I don't care to read a step by step story of what happened as Biden got excluded from the nomination he won in the primaries. I want much more skepticism about the way the Democratic Party allowed him to win the nomination in the first place and got itself into the seeming (or bullshit) jam it was in last summer

4. Did Democrats engineer a nomination for Harris that she couldn't have won in a fair primary or did Democrats accidentally wait too long to lose hope in Biden and find themselves in the position where they couldn't avoid giving it to Harris? And how did the mainstream media contribute to the problem?

5. Make a clear and concise list of why Tapper's going Clooney-centric exemplifies what's wrong with mainstream media politics and with the Democratic Party

Grok's responses are here.  

The response is, of course, astonishing in its accomplishment.  AI is still a wonder.

But it also highlights multiple issues with which people are wrestling.  

The answers are startlingly articulate and sophisticated.  However, there is a certain flatness of voice.  It feels like Grok is shaping its response based on its interpretation of Althouse's sentiment in her questions.  

Because it is a sequence of responses to a series of prompts, it answers each somewhat independently of the others with the result that there is a certain degree of repetition of stated predicate assumptions or examples of evidence.  

Grok occasionally has a relatively clear position but it seems to usually take the ambivalent stereotypical academic framing "On the one hand, . . . . But on the other hand, . . . ."  There seems to be a desire for the happy median even if its answer out on one wing or the other.

On the whole, the answers are impressive and reasonably comprehensive but they feel to me as a great averaging.  The interesting insights to human debate are the lonely voices, the iconoclastic views, those with a passionate insight.   They may be true or untrue but that is where the energy is in a conversation or debate.  

That energy feels missing from the Grok response.  It is not that its response is wrong but that its response feels like a good articulation of the average answer.  It lacks both spice or the catalyst to fresh thinking.  It seems to me.