Apple Takes Practical Approach to AI, Unveiling New Features and OpenAI Partnership at WWDC
Apple is set to unveil its latest AI advancements at its Worldwide Developers Conference next month, focusing on practical applications for everyday users. The company will integrate AI tools into core apps like Safari and Photos, powered by a combination of on-device processing and cloud computing via its M2 Ultra chips. New features include AI-powered voice memo transcription, photo retouching, enhanced search functionality, and even the ability to generate custom emojis on the fly. Apple will also introduce “smart recaps” to summarize missed notifications and content. While playing catch-up to competitors like Google, Apple aims to leverage its vast user base to drive adoption of its AI features. The company will also address privacy concerns, emphasizing its commitment to on-device processing and secure cloud infrastructure. In a significant move, Apple will announce a partnership with OpenAI to integrate ChatGPT into iOS 18. This deal grants Apple access to cutting-edge chatbot technology but also raises concerns about OpenAI’s controversies and corporate stability. Apple is reportedly still pursuing a similar agreement with Google for its Gemini chatbot. Beyond AI, Apple is expected to unveil a redesigned iPhone home screen with customizable app icon colors and placement, offering users greater personalization options.
DeepMind’s AlphaFold3 Predicts Protein Interactions, No Longer Open Source
DeepMind has released AlphaFold3, a new version of its AI software that goes beyond predicting protein shapes to modeling how proteins interact with other biomolecules like DNA and RNA. This breakthrough has significant implications for drug discovery, biomaterial development, and understanding disease mechanisms. AlphaFold3 boasts 50% greater accuracy than existing methods, according to DeepMind. The AI program analyzes molecular sequences to predict how proteins, DNA, RNA, and even smaller ligands fit together, including chemical modifications. However, unlike previous versions, AlphaFold3 is not open source. Scientists can access it for free through the AlphaFold Server for non-commercial research, but usage is limited to 20 jobs per day. This restriction on code and data access has sparked debate within the scientific community.
Averlon Raises $8 Million to Combat Cybersecurity Threats with AI
Seattle-based startup Averlon has secured $8 million in seed funding to tackle the growing challenge of cybersecurity alert fatigue. The company, founded by former Salesforce executives Sunil Gottumukkala and Vishal Agarwal, uses AI to prioritize critical security vulnerabilities and guide mitigation efforts. Averlon’s platform combines predictive and generative AI to analyze a company’s security posture and simulate potential attack paths. This approach helps security teams focus on the most pressing threats and streamline remediation efforts. The seed round was led by Voyager Capital, with participation from Salesforce Ventures and Outpost Ventures. Averlon, founded two years ago, is already working with paying customers and plans to expand its 12-person team. This investment highlights the increasing demand for AI-powered cybersecurity solutions as companies grapple with the evolving threat landscape and the rise of generative AI technologies.
