Resolves YES if an LLM is transported from earth to the moon, physically, as part of an equipment on a mission to the Moon (earth's moon). Otherwise NO.
The LLMs in the market title does not mean that multiple LLMs are required to go to the moon. A single LLM would be sufficient for a YES resolution
The weights and activations themselves must go to the moon, whether printed out or as circuitry does not matter. api calls back to earth do not count.
End date: Jan 1 2030
The first 24 hrs from the creation of the market act as a buffer period to solidify market resolution criteria.
How does this resolve if a LLM (weights and biases and all) is uploaded to a lunar device via radio or laser data transmission? It isn't "transported physically as part of equipment", but it is physically on the equipment on the moon. One could even argue that it would be "physically transported" because radio and laser are physical phenomena.
maybe the astronauts want to bring their AI girlfriend/boyfriend along with them - @firstuserhere would this count?
@ArmandodiMatteo Wikipedia says GPT-1 was the first LLM at 117 million parameters:
https://en.wikipedia.org/wiki/Large_language_model
Gemini Nano-1 is 1.8 billion so it should count.
@ArmandodiMatteo from wikipedia
The three Gemini models share the same software architecture. They are decoder-only Transformers, with modifications to allow efficient training and inference on TPUs.
They have a context length of 32,768 tokens, with multi-query attention. Two versions of Gemini Nano, Nano-1 (1.8 billion parameters) and Nano-2 (3.25 billion parameters), are distilled from larger Gemini models, designed for use by edge devices such as smartphones.
It counts.