Apple will ‘break new ground’ in generative AI this year, Tim Cook teases

Apple CEO Tim Cook recently hinted at significant advancements in generative AI from the company, stirring excitement about what's to come later this year, possibly around the unveiling of iOS 18 at WWDC in June. Cook's comments during Apple's annual shareholder meeting underscored the company's focus on integrating AI features into its ecosystem, following a brief mention of generative AI in last fall's introduction of new auto-correct and text prediction functionalities. This move signifies Apple's commitment to innovating in the AI space, despite entering the arena behind competitors like OpenAI and Google.

The anticipation builds as stakeholders and the tech community speculate on the potential impact of Apple's foray into generative AI. While previous mentions have been relatively sparse, the promise of AI-centric enhancements in iOS 18 suggests a significant pivot towards AI, positioning it as a central component of Apple's strategy moving forward. This shift comes after the discontinuation of the Apple Car project, highlighting AI as the new focal point for innovation within the company.

Despite joining the generative AI race behind established players, Apple's unique advantages, such as system-level integration capabilities and custom silicon design, may offer a distinct edge in enhancing the user experience across its ecosystem. As the tech world awaits further announcements, Cook's teasers fuel speculation and optimism about Apple's potential to redefine its offerings through AI, marking an exciting chapter for the company and its stakeholders.

Contentful is the future of intelligent composable content

As the leading intelligent composable content platform, Contentful enables developers and marketers alike to easily deliver compliant on-brand experiences at speed and scale—all within one unified content system. With Contentful, you can create infinitely and publish instantly.

Ideogram 1.0: the most advanced text-to-image model, now available

Ideogram 1.0, the latest and most sophisticated text-to-image model was unveiled today. This new model, developed from the ground up, revolutionizes the way text is rendered in images, offering unmatched photorealism, and precise adherence to prompts, and introduces the Magic Prompt feature for crafting detailed, imaginative prompts. Ideogram 1.0 is accessible to all at Ideogram AI, inviting users to become part of a worldwide community of creative minds, share their works, and draw inspiration from others.

Ideogram 1.0 sets a new benchmark in text rendering within images, addressing historical challenges of inaccuracy in AI-generated text. This enhancement enables users to effortlessly generate personalized messages, memes, posters, and more, with a level of precision previously unattainable. Our evaluations confirm that Ideogram 1.0 significantly outperforms existing models in text accuracy, cutting error rates by nearly half.

To enrich the user experience further, Ideogram now offers both free and paid subscription plans. The free plan includes daily generation allowances, while paid subscriptions unlock additional benefits such as priority processing, private generation options, image uploads, and exclusive access to the Ideogram Editor.

Alibaba presents EMO - The most amazing audio2video I have ever seen.

EMO introduces an innovative audio-driven portrait-video generation framework that can create expressive vocal avatar videos from a single reference image and vocal audio input, such as talking or singing. The process involves two main stages: the initial Frames Encoding stage, where ReferenceNet extracts features from the reference image and motion frames, and the Diffusion Process stage, where an audio encoder processes the audio embedding. This stage incorporates facial region masks with multi-frame noise, leveraging Reference-Attention and Audio-Attention mechanisms within the Backbone Network to ensure identity preservation and motion modulation. Additionally, Temporal Modules adjust the motion's velocity, offering videos of any duration based on the input audio's length.

The framework's versatility is showcased through its ability to generate videos with expressive facial expressions, head poses, and maintain character identity over any duration. It supports inputs of singing and talking in various languages, accommodating rapid rhythms and ensuring synchronization between the fastest lyrics and expressive character animations. Furthermore, EMO can animate portraits from diverse sources, including historical paintings, 3D models, and AI-generated content, bringing a wide range of portrait styles to life with realistic motion.

EMO's potential applications extend to animating movie characters delivering performances in different languages and styles, facilitating cross-actor performances. This capability opens new avenues for character portrayal in multilingual and multicultural contexts, significantly enhancing the expressiveness and dynamism of vocal avatar videos. The method’s support for different languages and portrait styles, coupled with its ability to keep up with fast-paced rhythms, underscores its adaptability and broad applicability in creating lifelike animations from audio inputs.

Other stuff

• Google CEO Sundar Pichai Says Its Malfunctioning Gemini AI Is ‘Unacceptable’

• SEC Investigating Whether OpenAI Investors Were Misled

• StarCoder 2 is a code-generating AI that runs on most GPUs

• An AI license plate surveillance startup installed hundreds of cameras without permission

• Instagram owner Meta forms team to stop AI from tricking voters

• GPT in 60 Lines of NumPy

• Digital Media Outlets Sue OpenAI for Copyright Infringement

• AI Grifters Fill Amazon With Kara Swisher Memoir Ripoffs

Hope you enjoyed today's newsletter

⚡️ Join over 200,000 people using the Superpower ChatGPT extension on Chrome and Firefox.