While large language AI models like ChatGPT, Gemini, and Llama dominate the headlines, Meta is shifting focus to small language models. According to a recently published paper by Meta’s research team, the company is betting on these smaller models as the future of AI.
Large language models typically use billions, even trillions, of parameters, making them too large to run on mobile devices. This size also leads to increased cloud costs and latency issues. Meta’s researchers argue there’s a growing need for efficient large language models that can operate on mobile devices.
The scientists detailed how they created high-quality language models with fewer than a billion parameters, making them suitable for mobile deployment. This challenges the common belief that model quality is directly proportional to the quantity of data and parameters used. Their small language model showed comparable performance to Meta’s Llama LLM in some areas.
“There’s a prevailing paradigm that ‘bigger is better,’ but this is showing it’s really about how parameters are used,” said Nick DeGiacomo, CEO of Bucephalus, an AI-powered e-commerce supply chain platform. “This paves the way for more widespread adoption of on-device AI.”
A Significant Shift
Meta’s research is crucial because it challenges the current cloud-reliant AI norm, where data is processed in distant data centers, explained Darian Shimy, CEO of FutureFund, a venture capital firm. “By bringing AI processing into the device itself, Meta is potentially reducing the carbon footprint associated with data transmission and massive data centers and making device-based AI a key player in the tech ecosystem.”
“This research is the first comprehensive and publicly shared effort of this magnitude,” added Yashin Manraj, CEO of Pvotal Technologies. “It is a crucial first step in achieving an SLM-LLM harmonized approach, balancing cloud and on-device data processing. It lays the groundwork for AI-powered applications to reach the level of support, automation, and assistance that have been marketed but lacked the engineering capacity to support.”
Meta’s move to downsize a language model by an order of magnitude is significant. “They are proposing a model shrink that makes it more accessible for wearables, hearables, and mobile phones,” said Nishant Neekhra, senior director of mobile marketing at Skyworks Solutions. “This opens up a whole new set of applications for AI and provides new ways for AI to interact in the real world while addressing the deployment challenges of large language models on edge devices.”
Optimizing for Efficiency and Accessibility
Meta’s focus on smaller AI models for mobile devices mirrors a broader industry trend towards optimizing AI for efficiency and accessibility, explained Caridad Muñoz, a professor of new media technology at CUNY LaGuardia Community College. “This shift not only addresses practical challenges but also aligns with growing concerns about the environmental impact of large-scale AI operations. By championing smaller, more efficient models, Meta is setting a precedent for sustainable and inclusive AI development.”
Small language models also align with the trend towards edge computing, which aims to bring AI capabilities closer to users. “The large language models from OpenAI, Anthropic, and others are often overkill — ‘when all you have is a hammer, everything looks like a nail,’” DeGiacomo said.
Meta’s commitment to small AI models indicates a significant shift in the AI landscape, emphasizing efficiency, accessibility, and sustainability. This move could transform how AI operates on mobile devices and pave the way for broader, more practical applications of AI technology.
We have helped 20+ companies in industries like Finance, Transportation, Health, Tourism, Events, Education, Sports.
