Yeah, look, this is not a huge advance in terms of AI functionality and development. But as a PR exercise, it is a somewhat significant leap.
Late last week, Meta announced that it’s working with Booz Allen to “deploy a fine-tuned version of Llama 3.2 aboard the International Space Station (ISS) National Laboratory,” which will enable astronauts to use its AI models for problem-solving and content generation in space.
As explained by Meta:
“Using a compact, energy-efficient system similar to those used in satellites, the team sped up AI tasks from minutes to just over a second using models fine-tuned with NVIDIA CUDA software and its acceleration libraries, including cuDNN and cuBLAS. As a result of this tech stack, which is believed to be the first of its kind used beyond Earth, astronauts will be able to harness the advantages of digital technologies for research, fast decision-making and agile data access without having to manage the risk of losing connectivity.”
Meta says that “Space Llama” will include both generative and multimodal AI, so it can process multiple inputs to provide AI-based recommendations and assessment:
“For example, Space Llama can be used to help researchers easily retrieve information from crucial technical reference documents and instructions, no internet needed.”
Meta says that this is “a standout case of American innovation,” and a key example of how its open source AI approach will help to sharpen the U.S.’s competitive edge.
“This moment is just the beginning of Space Llama. We envision a future where open source models like Llama play a vital role in space exploration and research, working “hand in hand” with astronauts as they solve complex scientific problems and help advance innovation for all.”
Which is all hyperbole, with the announcement being much more bombastic, and nationalistic, than Meta’s usual, more technical explainers.
Yet, in terms of development, this doesn’t really matter to the average person, to the average Facebook or IG user, this is not some major innovation with widespread application or implication as a result.
So why is Meta so keen to highlight this as a major milestone in AI achievement?
Well, for one, the capacity use its AI tools in space does feel noteworthy, given the computational and connective requirements. Though astronauts have been able to access the internet from the ISS since 2010, so really, it doesn’t seem like that big of a technical leap, in general context.
What Meta seems more interested in highlighting, however, is the need for the U.S. government to clear the guardrails on AI development, in order to keep America in the lead over international rivals.
China, in particular, is spending big on AI development, and if the U.S. wants to keep a hold on its leadership in the AI race, it needs U.S. regulators to stay out of the way, and enable companies like Meta to keep pushing their progress.
Which also comes with risks, and various groups are already warning against allowing AI development to go too far, unchecked.
But big announcements like this, which merge the opportunities of space exploration with the latest AI developments, could help to build Meta’s case for broader investment and open development of AI.
And at the same time, it also feels like a bit of a slap in the face for Elon Musk, who would consider himself to be the leader in space-related innovation. The fact that Meta’s AI model is being used in space first, over Musk’s xAI, will no doubt stick in his craw, and that may well have been another motivation for Meta to highlight this.
Either way, it’s interesting to see how Meta’s open source approach is not only expanding its AI reach and innovation, but also how it’s looking to talk about such, and promote those achievements to the public.
