Ad
Short

Alphabet's AI lab Deepmind cut employee costs by 39% in 2022, according to a filing with the UK government, with staff costs falling from nearly $1.2 billion in 2021 to nearly $731 million. This was accompanied by a more than 40% drop in profits over the same period, from nearly $126 million in 2021 to nearly $74.9 million in 2022. To further improve efficiency, Alphabet merged its Google and Deepmind AI divisions in April 2023 to better share knowledge and computing resources. Google Deepmind's answer to OpenAI's GPT-4, Gemini, is expected to be released later this year.

Short

ElevenLabs introduces its AI Dubbing tool, a voice translation feature that converts spoken content into another language while preserving the speaker's voice. The tool can translate audio and video content into more than 20 languages supported by the Eleven Multilingual v2 model, including Hindi, Portuguese, Spanish, Japanese, Ukrainian, Polish, and Arabic. This feature aims to break down language barriers and make content globally accessible, benefiting creators, educators, and media companies. You can try it out here.

Ad
Ad
Short

Adobe and the Coalition for Content Provenance and Authenticity (C2PA) have developed an "icon of transparency" that can be added to content metadata to indicate its provenance and whether it was created using AI tools. The icon will be available on Adobe platforms such as Photoshop and Premiere, and eventually on Microsoft's Bing Image Generator. Hovering over the mark will reveal information about the owner of the content and the AI tool used to create it. Andy Parsons, senior director of Adobe's Content Authenticity Initiative, likens the mark to a "nutrition label" for media provenance. C2PA members, including Arm, Intel, Microsoft, and Truepic, will implement the new mark in the coming months.

Image: Adobe
Short

Tesla is building a space in Austin, Texas, to house its Dojo supercomputer, which will run the AI software for its self-driving vehicles, The Information reports. CEO Elon Musk has hinted that the Dojo could eventually offer a cloud service similar to Amazon Web Services.

The project reflects Musk's intention to take greater control of the technology, which is essential to Tesla's AI-powered products and also Musk's overall AI plans. Musk wants to invest well over $1 billion in the supercomputer by the end of 2024. The Dojo supercomputer, based on an in-house chip designed by Tesla, could reduce the company's reliance on Nvidia. Analysts predict that the Dojo project could add as much as $500 billion to Tesla's enterprise value, and could even be used in other Musk companies such as X and SpaceX.

Ad
Ad
Ad
Ad
Google News