Java to Shed Legacy Applet API in Upcoming JDK Release
OpenJDK proposal aims to remove deprecated Java Applet API, citing lack of support from JDK and web browsers, and security concerns.
Riley King
Google has announced that its AI video generator, Veo, is now available in private preview for customers using Vertex AI, its AI development platform. This move marks a significant expansion of Veo's capabilities, which were first unveiled in April, and enables Google Cloud customers to generate high-definition videos from images and prompts.
Veo is a flagship generator that can produce 1080p clips of up to six seconds in length, capturing different visual and cinematic styles, including shots of landscapes and time lapses. The model can also make edits to already-generated footage and supports masked editing for changes to specific regions of a video. According to Google, Veo is competitive with today's leading video-generating models, including OpenAI's Sora and models from Adobe, Runway, Luma, Meta, and others.
The private preview of Veo is already being utilized by two customers, Quora and Mondelez International, the owner of Oreo. Quora is using Veo to bring its Poe chatbot platform to life, while Mondelez International is leveraging the technology to create marketing content with its agency partners. "We created Poe to democratize access to the world's best generative AI models," said Spencer Chan, product lead at Poe. "Through partnerships with leaders like Google, we're expanding creative possibilities across all AI modalities."
Google attributes the delay in releasing Veo's API to "enterprise readiness." Warren Barkley, senior director of product management at Google Cloud, explained that the company has augmented, hardened, and improved the model for enterprise customers on Vertex AI. "As of today, you can create high-definition videos in 720p, in 16:9 landscape or 9:16 portrait aspect ratios," he said.
Veo has been trained on a vast amount of footage, which enables it to generate videos that are remarkably realistic. However, the model is not without its limitations. Objects in Veo's videos can disappear and reappear without explanation, and the model often gets its physics wrong. For instance, cars may inexplicably reverse on a dime. Despite these limitations, Veo is a significant advancement in AI video generation.
Google's approach to training Veo raises important questions about data sourcing and ownership. The company refuses to disclose the exact sources of its training data, citing "agreements" with YouTube creators. This has sparked concerns about the potential misuse of copyrighted content and the need for clearer guidelines on data usage.
In response to these concerns, Google has implemented prompt-level filters for Veo, including for violent and explicit content. The company also maintains an indemnity policy that provides a defense for eligible Veo users against allegations of copyright infringement. "We plan to indemnify Veo outputs on Vertex AI when it becomes generally available," said Barkley.
Veo's potential applications are vast, ranging from marketing and entertainment to education and beyond. However, the technology also raises important questions about the future of work in the creative industries. A 2024 study commissioned by the Animation Guild estimates that more than 100,000 U.S.-based film, television, and animation jobs will be disrupted by AI by 2026.
Google's cautious approach to Veo's rollout may be a response to these concerns. The company has not provided a timeline for Veo's general availability in Vertex AI, nor has it disclosed when Veo might come to additional Google platforms and services. "We typically release products in preview first, as it allows us to get real-world feedback from a select group of our enterprise customers before it becomes generally available for wider use," said Barkley.
In a related announcement, Google revealed that its flagship image generator, Imagen 3, is now available for all Vertex AI customers without a waitlist. Imagen 3 has gained new customization and image editing features, although these are currently gated behind a separate waitlist.
As Veo continues to evolve and expand its capabilities, it will be important to monitor its impact on the creative industries and the broader implications of AI-generated content. With its private preview now available to Google Cloud customers, Veo is poised to revolutionize the way we create and consume video content.
OpenJDK proposal aims to remove deprecated Java Applet API, citing lack of support from JDK and web browsers, and security concerns.
Japa no longer top savings goal in Nigeria, economic pressures squeeze disposable incomes, and savings rates decline.
Microsoft's Copilot Vision feature, now in limited testing, allows AI to browse the web with users, providing assistance and answering questions.
Copyright © 2024 Starfolk. All rights reserved.