Post
Ethan Mollick
‪@emollick.bsky.social‬
Of the three models showing their "thinking," Gemini may actually be the most interesting. Less personal and cute than DeepSeek r1, not obscured like the OpenAI models, and actually shows the pathways it took to solve the problem more clearly, making it easier to diagnose issues.
February 7, 2025 at 1:43 PM
5 reposts
93 likes
Striking how all 3 land on "lumi", and 2 on "flora", especially with a question where the answer could go in so many directions. I cannot imagine asking this of 3 people and them all thinking of the same word. Makes you wonder if increased use of LLMs will lead to more monoculture.
PS: or maybe it's as simple as the models being triggered by the "light" in "delight"
Interesting. Thanks for the luminescent observation.
how did R1 miss out on "glissora" with one letter? it'd win by a large margin with that (still wins with glissara though)