The most impressive example I've seen so far involves vision input:
llm -m gemini-2.0-flash-thinking-exp-1219 \
-a http''s://storage.googleapis.com/generativeai-downloads/images/geometry.png \
"What's the area of the overlapping region?"
Top-level
The most impressive example I've seen so far involves vision input: llm -m gemini-2.0-flash-thinking-exp-1219 \ 2 comments
@simon thanks, I ran some fun prompts, some good reasoning there but not perfect, still lots of bias llm -m gemini-2.0-flash-thinking-exp-1219 "will AI agents be the better incarnation of homo economicus than humans (homo sapiens) and what outcomes will this lead to?" llm -m gemini-2.0-flash-thinking-exp-1219 "what is the main successful cause for stopping climate change in the book 'the ministry of the future'? Evaluate and rank all contenders and present a reasoning for the final selection" |
@simon In one hand: yeah, this is pretty impressive. On the other: is it really when it only takes a single look without even thinking to solve it in your head?