OpenAI plans to release 'Strawberry' for ChatGPT in two weeks

OpenAI is preparing to release a new AI model named "Strawberry" as part of its ChatGPT service within the next two weeks, according to a report by The Information. The model focuses on improving reasoning capabilities, setting it apart from other conversational AI models by allowing it to "think" before responding rather than providing instant answers. This new approach could enhance the accuracy and depth of responses, appealing to users seeking more thoughtful interactions.

OpenAI, led by CEO Sam Altman, has garnered significant attention and investment in the AI sector, particularly with backing from Microsoft. The company has reported strong adoption of its AI products, with over 1 million paying users across its business services. The success of ChatGPT, powered by advanced large language models, has played a key role in driving this growth.

While Strawberry will initially focus solely on text input and output, it is not yet multimodal, meaning it won’t be able to process or generate images. The model is described as a standalone offering within the ChatGPT ecosystem, though details on how it will be presented to users remain unclear. Both OpenAI and Microsoft have yet to provide further comments regarding the upcoming release.

Finally, a developer tool that makes digging around for the answer a thing of the past. 

Other developer tools can’t tell you how your codebase works and why. Unblocked can. We augment your source code with context from Slack, Confluence, Jira (and more), so your team gets helpful answers instantly.

Unblocked gives you:

• Instant answers about every aspect of your codebase 

• Relevant, historical context for any file open in an IDE

• A powerful Slack integration that answers questions in channels, removing interruptions to you

⌛Engineers who use Unblocked save an hour or more a day.

“This is the best application of AI’s potential in enhancing developer proficiency, making developers themselves more effective — rather than typing for them.” Mathew Spolin - SVP Engineering at AppDirect

Roblox announces AI tool for generating 3D game worlds from text

Roblox has announced plans to introduce a new generative AI tool aimed at simplifying the creation of 3D game environments on its platform. This tool, still under development, allows users to generate 3D objects and entire game worlds using text prompts. In a demo shared by the company, a user types commands like "create a race track" or "make the scenery a desert," and the AI instantly creates corresponding 3D models. The tool is expected to appeal to users without extensive 3D design skills, streamlining game creation for a broader audience.

The AI system operates through what Roblox calls a "3D foundational model," which predicts the next structural element in a 3D sequence, similar to how large language models (LLMs) process text data. By tokenizing 3D blocks, the model can predict and generate entire environments. However, challenges in sourcing high-quality 3D data have caused occasional glitches, like inaccurate object formations. To address this, Roblox employs a second AI model that acts as a visual moderator, refining and correcting errors in the generated 3D structures.

Roblox's decision to open-source its 3D foundation model could allow developers, and even competitors, to contribute to and improve the technology. This aligns with the company's broader goal of democratizing game creation, making it accessible to more users and developers. The announcement was made at Roblox's annual Developers Conference, where the company also set an ambitious goal to capture 10 percent of global gaming revenue and introduced new features aimed at enhancing the user experience.

Overall, the new AI tool is part of Roblox’s continued efforts to integrate artificial intelligence into its platform, building on earlier AI-powered features such as the Code Assist beta and Material Generator. Roblox envisions a future where generative AI plays a central role in real-time game creation, making it more accessible and efficient for users across the platform.

iPhone 16 gets AI superpowers

Apple has unveiled the highly anticipated iPhone 16, which comes equipped with the powerful A18 chip and a suite of advanced Apple Intelligence (AI) features. This new device promises to revolutionize how users engage with their smartphones, leveraging AI to enhance everything from writing and communication to photography and productivity.

One of the standout features is the AI-driven writing tools that allow users to rewrite emails, edit notes, and even create custom emojis tailored to conversations. The enhanced search capabilities make it easier to find specific photos or videos through natural language commands. AI also helps manage daily tasks by summarizing emails, highlighting important notifications, and prioritizing what needs attention, keeping users focused on their most essential activities.

The iPhone 16’s camera is also equipped with powerful visual intelligence, allowing on-device AI to analyze photos without compromising user privacy. Additionally, it seamlessly connects with third-party tools like ChatGPT. Siri, Apple's virtual assistant, has also received a significant upgrade, enabling it to handle more complex natural language queries, incorporate contextual information, and execute commands with greater precision.

This release signals a major shift in Apple's approach to AI, with the iPhone 16 integrating generative AI capabilities into everyday tasks. By developing its own models and integrating tools like ChatGPT, Apple is positioning itself at the forefront of AI innovation, ready to challenge other tech giants in the space.

Related:
- The real power of Apple Intelligence will show up in third-party apps
- AI-powered visual search comes to the iPhone

Other stuff

• Sergey Brin says he’s working on AI at Google ‘pretty much every day’

• LLMs are a dead end to AGI, says François Chollet

• Elon Musk says Tesla has ‘no need’ to license xAI models

• Harrison.rad.1, the latest frontier in radiology-specific foundational models

• ‘Right to Repair for Your Body’: The Rise of DIY, Pirated Medicine

• MI6 and CIA using generative AI to combat tech-driven threat actors

• Cops lure pedophiles with AI pics of teen girl. Ethical triumph or new disaster?

Hope you enjoyed today's newsletter

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

OR