👉 ChatGPT now available via phone and WhatsApp

OpenAI has launched a new initiative that allows users to connect with ChatGPT through a dedicated phone line, 1-800-CHATGPT, and via WhatsApp. This service is designed to provide easy access to the AI for real-time conversations, offering up to 15 minutes of interaction for free per phone number each month. The goal of this initiative is to make AI more accessible to a broader audience, especially those who may not be familiar with traditional text-based interfaces.

Kevin Weil, OpenAI’s Chief Product Officer, announced this feature during a recent livestream, emphasizing its potential to introduce AI to new users in a familiar format. The integration of ChatGPT with WhatsApp further expands its reach, allowing users globally to engage with the AI without needing to navigate complex platforms.

👉 Google Whisk: A visual creation tool for rapid testing

Google has launched Whisk, an experimental tool that allows users to generate AI visuals using image prompts instead of traditional text-based descriptions. This platform is built on Google’s Imagen 3 generative AI model and is currently available for free to users in the U.S. Whisk aims to simplify the creative process by enabling users to upload images such as subject, scene, and style, and generate an image without needing a text prompt.

Whisk is designed to cater to both novice and experienced users by allowing for rapid experimentation with visual concepts. Users can upload reference images, and Google’s Gemini model analyzes these inputs to create corresponding visuals. This approach not only enhances accessibility but also encourages creativity by enabling users to explore diverse artistic styles in a more intuitive manner.

→ Try it here

👉 Create moodboards with Midjourney

Midjourney released "Moodboards", which let you personalize models using collections of images. The new functionality allows users to upload various reference images (100 maximum) that guide the AI's creative output, with a new moodboard code. This feature is particularly valuable for creatives who wish to explore complex combinations and styles in their generated results. Users can add images directly, link to URLs, or select from previously generated images, facilitating a more intuitive and collaborative creative process.

In addition, they also added support for multiple personalization profiles and made image ranking 5x faster. The introduction of these features comes alongside Midjourney's plans to release its v7 image model and a new video model after the Christmas holiday.

👉 OpenAI unveils new reasoning models: o3 and o3-mini

OpenAI has confirmed the upcoming release of two advanced reasoning models, named o3 and o3-mini, which are successors to the previously launched o1 and o1-mini models. Announced during the final day of OpenAI’s "12 Days of OpenAI" livestream event, these models are expected to enhance the capabilities of AI in performing complex tasks that require sophisticated reasoning skills. CEO Sam Altman highlighted the significance of this launch as a pivotal moment in the evolution of AI.

The initial rollout of o3 and o3-mini will be limited to selected third-party researchers for safety testing, with o3-mini available by the end of January 2025. Altman stated that the models are intended to enable more intricate problem-solving abilities in various domains, including science and mathematics. OpenAI’s commitment to safety and alignment in AI development remains a core focus, with the new models designed to adhere to rigorous safety standards.