Please scroll past if you're busy. This is maxx navel-gazing
I wrote a longish article which is purposely allusive and likely elusive in meaning. I asked two American AIs and two Chinese AIs to summarize it for me:
- Gemini 3 Pro Preview
- GPT 5.1
- GLM-4.6 with DeepThink[^1] + Web Search
- DeepSeek 128k (trained up to July 2024) with DeepThink + Web Search
I perceive a difference in tone as well as perceptiveness, based on the models.
Sharing the outputs below, without identification[^2]:
The summaries generally missed the point; were too eager to assume a clear, direct communication tone. They ignored references to specific current events which were chosen for specific meaning. Like about as much as you'd expect a loving, attentive Boomer grandparent to pick up, from listening to a terminally online grandchild talking in native jargon. Or, a college literature professor listening to students talking about sub-genres of rap music. But we can all imagine individual college professors bec...
