Pixel 8 Pro with a privacy boost - running a local LLM

📰 Pixel 8 Pro runs Google's generative AI models on-device

Google's Pixel 8 Pro is set to be the first device to operate Google’s generative AI models directly on the hardware.

• The Pixel 8 Pro's Tensor G3 chip accelerates AI tasks and can run Google's advanced models for various apps, like image editing.

• Enhancements to Magic Eraser enable better photo touch-ups, filling in spaces with new pixels for high-quality outcomes.

• Zoom improvements provide sharper photo details.

• Audio recording on Pixel 8 Pro can summarize meetings.

• A large language model on the Pixel 8 Pro will upgrade Gboard for smarter, conversation-aware replies.

Why it Matters: Running AI models directly on the Pixel 8 Pro enhances user privacy, as data processing happens locally. This limits the potential exposure of personal information, ensuring safer and more private user experiences.

📰 Canva’s new AI tools automate boring, labor-intensive design tasks

Magic Studio, Canva's latest suite, leverages AI to transform how everyone, from novices to experts, approaches content creation.

• Magic Switch: Converts designs into different formats or translates them into over 100 languages without manual adjustments.

• Magic Media: Features new text-to-video capability, turning text prompts or images into short videos. Also includes text-to-image enhancements.

• Photo Editing Tools: Includes Magic Grab to select and edit subjects in images, and Magic Expand to increase image frame size.

• Canva Shield: A set of controls ensuring AI isn't used for creating inappropriate content, with features for workplace control.

• Creator Compensation Program: Canva will pay designers who allow their designs to train its AI, allocating $200 million over three years for this.

Why it Matters: As one of the leading design platforms, Canva's advancements signal a broader shift towards automated design. Their commitment of $200 million to compensate creators emphasizes the importance of AI training and fair compensation in this era.

📰 Meta debuts generative AI features for advertisers

Meta introduces generative AI tools for advertisers, offering ways to design backgrounds, modify images, and create varied ad text.

• Generative AI will let advertisers create multiple backgrounds for their product images.

• Image expansion tool helps advertisers adjust to different aspect ratios for platforms like Feed or Reels.

• The text variation tool can produce up to six different text versions based on original ad copy, allowing better targeting.

Why it Matters: These AI tools underscore Meta's belief in the potential of generative AI to revolutionize the advertising industry, making processes more streamlined and efficient.

💽 OneDrive 3.0 from Microsoft brings a new design, the 'Copilot' feature, and AI-enhanced access and search capabilities. Copilot, set to roll out in December, will provide summaries and answer open-ended questions related to files.

🗃️ DeepMind, Google's AI division, debuts the largest-ever robotics dataset called the Open X-Embodiment, offering an information-rich learning tool for the robotics field. It also introduced two improved general-purpose models that benefit from this diverse database.

Thank you for reading today’s edition!

Feel free to reply to our emails with questions, suggestions, or topics you'd like to see covered, or drop us a message on Twitter or Facebook.

Until tomorrow,
- Ts Yotov (Bits and Neurons)