Microsoft’s research has led to the development of AI models small enough to run on personal devices like phones and laptops without significant compromises to their capabilities. This advancement could potentially unlock a new era of computing.
The AI model, named Phi-3-mini, is part of a family of smaller AI models recently released by Microsoft. It can run on a smartphone or a laptop, and it performs comparably to GPT-3.5, the OpenAI model behind the first release of ChatGPT.
The Phi-3-mini model was tested on several standard AI benchmarks designed to measure common sense and reasoning. The results were promising, indicating that the model is just as capable as its larger counterparts.
Microsoft also announced a new multimodal Phi-3 model capable of handling audio, video, and text at its annual developer conference, Build. This model, along with the rest of the Phi family, suggests that it’s becoming possible to build handy AI apps that don’t depend on the cloud. This could lead to more responsive and private applications.
The development of these pocket-sized AI models also reveals something about the nature of modern AI. It suggests that being more selective about what an AI system is trained on could provide a way to fine-tune its abilities. This is a significant shift from the current practice of feeding large language models with vast amounts of text from various sources.
Read more: www.wired.com