Tuesday, July 16, 2024
- Advertisement -
More

    Latest Posts

    Here are all the products launched at Google’s I/o 2024

    The conversation at Google’s annual developer conference, I/O 2024, on May 14 was dominated by AI. While Google shared updates on its AI models, particularly Gemini, it also announced new products and features. These span from generative AI models with creative output to integral hardware required to train AI models. Here are some of the new products launched at the event:

    GENERATIVE AI PRODUCTS

    VEO

    Google announced the launch of Veo, a text-to-video AI model capable of generating high-definition video, developed by Google Deepmind. Veo is capable of creating videos with 1080p resolution based on textual prompts. According to Google, Veo is capable ofaccurately capturing a prompt’s tone and rendering details in longer prompts.” VEO can also perceive cinematic terms like “timelapse” or “aerial shots of a landscape” and incorporate them into its output. It is available for select creators in private preview on VideoFX , Google’s new experimental tool for video generation. Google also plans to expand Veo’s capabilities to YouTube Shorts and other products.

    Imagen 3

    Google also introduced Imagen 3, a text-to-image model capable of creating photorealistic images. Google added that it had “far fewer distracting visual artifacts than our prior models.” It is also capable of rendering text, which Google acknowledged as “a challenge for image generation models.” Imagen 3 is available for select creators in private preview in ImageFX, Google’s AI-generated tool for text-to-image output. Based on customer requests, Google has added image editing controls and higher-quality images to ImageFX allowing users to add, remove or change specific elements in images by simply brushing over them.

    MUSICFX

    Google shared that through its partnership with YouTube, it was collaborating with artists to explore generative AI in music. Google also shared that it was developing a suite of music AI tools called Music AI Sandbox. Further, it added a DJ Mode to its generative AI tool MUSIC FX which would allow users to mix beats by combining genres, instruments, etc.

    WATERMARKING TECHNOLOGY

    Google also plans to introduce watermarking capabilities for AI-generated text and video, through its watermarking tool, SynthID. Launched last year to watermark AI-generated images, the tool was created to curtail the spread of misinformation through AI. The SynthID watermark, allows users to identify if the content is AI-generated. Google said that the watermarking technology for synthetic videos embeds a watermark directly into the pixels of every video frame, building on the technique used for watermarking images.

    However, Google noted watermarking’s limitation saying that it was not a “silver bullet” but “an important building block for developing more reliable AI identification tools and can help millions of people make informed decisions about how they interact with AI-generated content.”

    They also plan on making open-source SynthID for text watermarking and publishing a paper on text watermarking so developers can incorporate it into their models.

    AI HARDWARE

    Google announced the sixth-generation version of its AI-specific hardware, Tensor Processing Units (TPU) called Trillium. The new TPU aims to reduce latency and costs. Google said that it had 4.7 times increase in peak compute performance per chip, double the High Bandwidth Memory (HBM) capacity and bandwidth, and Interchip Interconnect (ICI) bandwidth than previous versions of the TPU. It is also 67% more energy-efficient than TPU version 5.

    Also Read:

     

    The post Here are all the products launched at Google’s I/o 2024 appeared first on MediaNama.

    Latest Posts

    - Advertisement -

    Don't Miss

    Stay in touch

    To be updated with all the latest news, offers and special announcements.