Meta(previously known as Facebook) has released an artificial intelligence(AI) model called the Segment Anything Model(SAM). SAM can identify individual objects within an image or video, even if it has not been trained on them. Meta claims that the dataset of image annotations used to train SAM is the largest ever of its kind. This new technology will likely have applications in fields such as image recognition and computer vision.
Big tech companies have been excited about their artificial intelligence (AI) breakthroughs since Microsoft-backed OpenAI’S ChatGPT chatbox become popular last fall. This has led to a race to dominate the AI space and increased investments. Meta has hinted at several features that use generative AI, which creates new content rather than just identifying or categorizing data like other forms of AI. However, Meta has not yet released a product that uses this type of AI.
Meta, the company that owns Facebook and Instagram, is interested in using artificial intelligence (AI) to help people create things like videos and illustrations. One tool they’re exploring can turn written ideas into strange, surreal videos. Another can create illustrations for children’s books. Meta already uses similar technology to help tag photos, remove bad content, and decide which posts to suggest to users. Their CEO, Mark Zuckerberg, has said that making AI tools for creativity is a priority for Meta this year.
A company has made a new technology called SAM, which can generate images from written descriptions. They’re going to let people download SAM for free, but they can only use it for research purposes and not to make money. People who use SAM will also have to agree to only use it for research purposes if they upload their own pictures to the system.
