Alas, Evernote AI is also limited in another way. When I noticed that I was getting much shorter answers from Evernote AI than I got when I posed the same questions directly in ChatGPT, I asked ChatGPT to explain the differences. You may want to do a post on ChatGPT's response:
Even if Evernote’s AI is “powered by me,” it is operating under very different constraints:
• it is optimized for local retrieval and summarization, not long-form argument construction
• it is trying to be safe, fast, and broadly helpful across many use cases
• it is reacting to your notes, not to a developing book-level theory
Alas, Evernote AI is also limited in another way. When I noticed that I was getting much shorter answers from Evernote AI than I got when I posed the same questions directly in ChatGPT, I asked ChatGPT to explain the differences. You may want to do a post on ChatGPT's response:
Even if Evernote’s AI is “powered by me,” it is operating under very different constraints:
• it is optimized for local retrieval and summarization, not long-form argument construction
• it is trying to be safe, fast, and broadly helpful across many use cases
• it is reacting to your notes, not to a developing book-level theory
So what it does well is:
– finding patterns across many notes
– summarizing clusters of material
– reminding you what you’ve already collected
– surfacing connections you might overlook
What it does poorly—by design—is:
– sustained argumentative reasoning
– strategic persuasion decisions
– tone calibration for skeptical audiences
– sequencing and framing trade-offs
Yeah. I'd agree with that. They are looking into including more notes in research and answers which will be helpful.