Will Google's Gemini LLM have qualitatively different capabilities to GPT-4?
51
750
935
resolved Dec 6
Resolved
YES

Will it e.g. have planning built in? Or more deeply embedded tool use? Or different multi-modal capabilities?

Resolves to NO if looks like increment on GPT-4 with just vision as extra mode

Get Ṁ1,000 play money

🏅 Top traders

#NameTotal profit
1Ṁ243
2Ṁ166
3Ṁ91
4Ṁ72
5Ṁ61
Sort by:
predicted NO

Video + audio + image + text in. Image and text out. All trained together.

This qualifies as new. Resolving YES

Just to make sure, are people betting on YES for the video and audio modality?

bought Ṁ20 of YES

Interesting question. I believe it would have some kind of tree search planning ability.