In this weekly roundup:

• From GPT-4o to AGI

• Co- founder of OpenAI’s New AI Venture

• Anthropic Introduces Claude 3.5 Sonnet

• Google AI Gets Personal with Wearables

• Second Sora Competitor in Just a Week

• Apple ‘Intelligence’ is here

• Midjourney's 'Model Personalization'

• AI Stories

Main AI Updates

😱From GPT-4 to AGI

Leopold Aschenbrenner's Vision: Leopold Aschenbrenner, a former researcher at OpenAI, has made a bold prediction: Artificial General Intelligence (AGI) will arrive by 2027.

His foresight, based on the rapid advancements in AI technology, suggests that AGI could compress decades of progress into a single year, driven by hundreds of millions of AGI instances.

This transformative leap from GPT-4 to AGI is anticipated to have significant economic and military implications, making AI security measures a national priority.

Economic and Military Impact: Aschenbrenner foresees AGI rapidly accelerating technological progress, akin to the speed of innovations during the Manhattan Project.

The national security implications are profound, with AGI's potential to redefine power dynamics globally.

Current AI labs' security measures are deemed insufficient, posing risks of sensitive information falling into adversarial hands.

The potential of AGI to reshape industries and national security underscores the urgent need for robust strategic planning and international collaboration.

Managing AGI’s rapid intelligence growth is a critical challenge, with the potential for catastrophic risks if not properly controlled.

👀Co- founder of OpenAI’s New AI Venture

Founders and Mission: Ilya Sutskever, co-founder of OpenAI, has launched a new company, Safe Superintelligence (SSI), with the mission of developing safe and powerful superintelligent AI.

He is joined by former Y Combinator partner Daniel Gross and ex-OpenAI engineer Daniel Levy. This venture comes shortly after Sutskever's departure from OpenAI.

Focus on Safety: SSI aims to create beneficial superintelligent AI within the next decade, prioritizing safety and ethical considerations.

Unlike OpenAI’s recent focus on commercial products, SSI will operate purely as a research organization, insulated from short-term business pressures.

Funding and Vision: Although specific funding details have not been disclosed, Gross confidently stated that raising capital will not be a problem.

The establishment of SSI follows months of speculation after Sutskever’s exit from OpenAI, signifying his commitment to a new path focused on safe AI development.

Sutskever’s departure from OpenAI and the founding of SSI suggest a vision for a safer approach to superintelligence.

This new venture highlights the importance of dedicated research in the AI field, ensuring that advancements in AI are both powerful and beneficial.

⚡️Anthropic Introduces Claude 3.5 Sonnet

Model Enhancements: Anthropic has released Claude 3.5 Sonnet, an upgraded AI model that outperforms competitors like GPT-4o and its predecessor Opus.

The new model is twice as fast as Opus and offers significant cost reductions, with token pricing at one-fifth of Anthropic’s previous top-tier model.

Performance and Features: Claude 3.5 Sonnet excels in benchmarks for reasoning, coding, mathematics, and general knowledge.

It introduces ‘Artifacts,’ a feature allowing users to view, edit, and build with Claude in a real-time side panel workspace.

Future Releases: Anthropic has announced that 3.5 versions of Haiku and Opus are scheduled for release later this year, along with new features like Memory, which will further enhance the capabilities of their AI models.

Latest Developments

⌚️Google AI Gets Personal with Wearables

AI-Driven Health Insights: Google has introduced two groundbreaking AI models that revolutionize personal health data analysis from wearable devices.

The PH-LLM model, a fine-tuned version of Gemini, can interpret health data and generate coaching insights for sleep and fitness.

Additionally, PHIA, an AI agent, combines language skills with code and search capabilities to analyze wearable health data.

Performance and Capabilities: PH-LLM performs comparably to human experts in providing health insights and has achieved expert performance on certification exams.

PHIA, scoring 84% on health insight questions, demonstrates strong reasoning and data analysis capabilities on complex queries.

Google's research showcases the potential of AI to elevate health wearable tracking, enabling personalized support for users.

As these AI tools integrate into popular devices, they promise to deliver expert fitness and nutrition coaching directly to users' wrists, marking a significant advancement in personal health technology.

🤯Second Sora Competitor in Just a Week

Innovative AI Video Generation: Luma Labs has launched Dream Machine, an AI model that generates high-quality 5-second video clips from text and image prompts.

Unlike OpenAI’s Sora, Dream Machine is already available to the public. This transformer model is trained on video content, producing coherent outputs that mimic real-world interactions and physics.

User Accessibility and Features: The model can generate 120 frames of video in 120 seconds, with initial wait times longer due to high demand.

Dream Machine offers a free plan with 30 video generations and a paid tier extending to 2000 outputs per month.

Dream Machine signifies a massive leap in AI video capabilities, offering users a taste of powerful new tools while raising the bar for competitors like OpenAI’s Sora.

This advancement is set to revolutionize the way users create and interact with AI-generated video content.

📲Apple ‘Intelligence’ is here

Advanced AI Integration: At WWDC 2024, Apple unveiled "Apple Intelligence," introducing advanced AI features across iPhone and Mac.

This includes an upgraded Siri with more conversational abilities, custom Genmoji, and integration with OpenAI's GPT-4o.

Key Features

• Advanced Siri: Now more conversational, Siri can handle actions across apps, manage notifications, and write and summarize text.

• Privacy: Apple emphasizes privacy with AI features operating primarily on-device and complex requests sent to Apple's "Private Cloud Compute."

• Genmoji and Image Playground: Users can generate AI-created emojis and images, with a developer API for integration.

• Enhanced Search: Detailed queries and new tools like Magic Eraser for image editing.

• OpenAI Integration: Siri utilizes GPT-4o for advanced queries while ensuring user permission and privacy.

Apple’s entry into the AI market with a strong focus on privacy highlights the importance of secure AI development.

These features promise to enhance user experience with smarter, more intuitive interactions.

🎨Midjourney's 'Model Personalization'

Tailored AI Art: Midjourney has introduced a new feature called 'model personalization,' allowing users to tune the Midjourney algorithm to their personal tastes.

This feature reduces Midjourney's training data bias, learning from user preferences to generate more personalized outputs.

User Engagement: Model personalization learns from votes in pair ranking and images liked by users, requiring approximately 200 pair rankings or likes for optimal performance.

Users can enable this feature by typing --p after their prompt, receiving a unique code for sharing their personalized model.

This innovation in AI art creation provides users with a more personalized experience, aligning the outputs with individual tastes.

Midjourney's feature represents a significant step towards more user-centric AI tools, enhancing the creative potential of AI-driven art.

Top Stories

🤖AI Stories

Google Translate is now using AI to translate a staggering 110 new languages. Imagine the doors this opens for communication across the globe.

Instagram is starting to let some creators make AI versions of themselves. This could change the way we interact on social media forever.

Open AI and Microsft are facing a lawsuit from the Center for Investigative Reporting as the news industry steps up its attack on AI. The battle over AI and journalism is heating up.

NBC is bringing AI Al Michaels to Peacock for customized Paris Olympics recaps. This is a game-changer for sports broadcasting.

Bill Gates says AI will make it easier to combat climate change but must be ‘used by people with good intent.

Thanks for reading!

If you enjoyed this, please help spread the love by forwarding this Newsletter to a friend or colleague.

I hope to see you in the next one!

AI Genesis

P.S. Whenever you’re ready, here's how I can help you:

1. Get your product in front of 4000+ solopreneurs, business owners, and professionals. Sponsor this newsletter here.