There’s a bit of a middle ground where the user may say something like “tech articles from new site” which would require the agent to move around reading and aggregating information to return to the user. Is this the former or the latter?
I think the dividing line is more "happens in anticipation there might be a user session which wants the data" and "happens as the result of a user session which wants the data" than it is about how it was done. E.g. prefetching content after a page load is done as a result of the user session getting to the page, it just as going around to any sites it can find to cache it on the chance the user might need them.
You mean combine sort of related information from all over the internet into new statistically generated phrases that appear to answer the question, but still contain all kinds of hallucinations, fabrications, and wrong statements?
While true Dream Theatre is prog metal, it's also great evidence that prog rock was not an embarrassing dead end when you have amazing bands like Dream Theater, Animals as Leaders, Devin Townsend, and Between the Buried and Me as their new prog legacy.
I’ve been fortunate to see all of those bands, on the recent tour and years ago on a Prog Nation tour.
Animals I need to listen some more too, not something I can pick up on in a single pass. This isn’t uncommon. I think DT is actually really approachable, but even still it takes me several listens to find the groove of the album (in contrast to LTE which takes even more time). Animals is just flat a lot to absorb in one hit.
Townsend was super entertaining and has a remarkable voice. But much like Buried when I saw them years ago, I like the music but it’s hard for me to get past the dirty singing. Same with Opeth, at the same tour as Buried.
This discussion feels very meta. In the sense that Twitter is the "bad" person, and we are the actors contributing our individual outrage induced take.