Good morning. It’s Monday, September 30th.

Did you know: On this day in 1996, the Nintendo 64 was officially released in North America.

In today’s email:

• OpenAI’s Future Pricing Plan

• California AI Bill Vetoed

• Apple Pulls Out of OpenAI Funding

• Robot Arm Demo

• 5 New AI Tools

• Latest AI Research Papers

You read. We listen. Let us know what you think by replying to this email.

OpenAI plans to double ChatGPT's price in five years, targeting $100 billion in revenue by 2029

OpenAI intends to raise the price of ChatGPT from its current $20 per month to approximately $44 over the next five years, aiming to more than double its pricing structure while targeting a staggering $100 billion in revenue by 2029. Internal documents reveal that OpenAI's revenue skyrocketed to $300 million in August, fueled by a user base that surged from 100 million in March to 350 million by June.

The free version of ChatGPT plays a crucial role in this growth, with projected earnings of $2.7 billion this year, a significant increase from $700 million last year, including about $1 billion from business users. Microsoft, as OpenAI's largest cost driver due to its computing power, continues to significantly invest in this collaboration. Read more.

Gov. Newsom vetoes California’s controversial AI bill, SB 1047

California Governor Gavin Newsom has vetoed SB 1047, a bill that sought to regulate AI model development. Authored by Senator Scott Wiener, the bill aimed to hold AI developers accountable for implementing safety protocols, targeting models costing over $100 million and requiring 10^26 FLOPS during training.

Despite opposition from key Silicon Valley figures like Meta’s Yann LeCun and U.S. Congressman Ro Khanna, Newsom acknowledged the bill's good intentions but critiqued its broad application, stating it imposed unnecessary burdens on lower-risk systems.

Meanwhile, Wiener voiced frustration, calling the veto a setback for AI safety oversight. Newsom's office pointed out his recent signing of 17 other AI-related bills and his commitment to crafting more targeted regulations with guidance from experts like Fei-Fei Li, who had previously argued that the bill would harm California’s AI ecosystem. Read more.

Apple pulls out of latest OpenAI funding round  

Apple has pulled out of discussions for OpenAI's upcoming $6.5 billion funding round, which could have marked a rare cross-investment by the tech giant. Meanwhile, major players like Microsoft and Nvidia continue negotiations. Microsoft, already invested in OpenAI to the tune of $13 billion, is anticipated to add another $1 billion. Thrive Capital leads the round with a commitment of approximately $1 billion, joined by participants including Tiger Global Management and UAE state-backed MGX.

OpenAI is also navigating a complex transition from a nonprofit to a for-profit structure due to investor demands. If this restructuring isn't completed within two years, current investors can reclaim their investments. Ongoing funding discussions may still see changes in both participants and amounts before the round closes.

In light of recent executive departures, including CTO Mira Murati, OpenAI CFO Sarah Friar reassured investors of the company's strong position. She emphasized a "talented leadership bench," the commitment to sustainable revenue models, and ongoing innovation, particularly with its AI tool, Sora. Read more.

Robot hand can detach from arm, crawl over to objects, and pick them up

Researchers at Switzerland's EPFL have introduced a groundbreaking robotic hand that can detach from its arm, crawl, and grasp objects. Presented in the paper “Beyond Manual Dexterity” at the IEEE International Conference on Robotics and Automation, this study advances dexterous manipulation techniques using deep learning for multi-fingered grasping.

It explores contact information-guided grasping in cluttered environments, challenging the limitations of traditional, biologically inspired designs. The hand's ability to bend backward enables it to handle a wider variety of objects, pushing the boundaries of conventional robotics.

While its human-like appearance might creep you out, it's a sign that robotics is getting more creative. As researcher Xiao Gao cheekily told IEEE, "Although you see it in scary movies, I think we're the first to introduce this idea to robotics.” Read more.

• China makes AI breakthrough, reportedly trains generative AI model across multiple data centers and GPU architectures

• Amazon's $4 billion investment in Anthropic gets green light from UK regulator

• Microsoft Prepares to Relaunch AI-Powered Recall Offering

• Google and Meta update their AI models amid the rise of “AlphaChip”

• Meta tests personalized AI content in Facebook and Instagram feeds

• OpenAI's o1 probably does more than just elaborate step-by-step prompting

• Microsoft Prepares to Relaunch AI-Powered Recall Offering

• Raspberry Pi launches camera module for vision-based AI applications

• Huawei takes on Nvidia in China's AI chip market with new Ascend 910C processor

📄 Emu3: Next-Token Prediction is All You Need

📄 MIO: A Foundation Model on Multimodal Tokens ​​

📄 MaskLLM: Learnable Semi-Structured Sparsity for Large Language Models  

📄 EMOVA: Empowering Language Models to See, Hear and Speak with Vivid Emotions

📄 LLaVA-3D: A Simple yet Effective Pathway to Empowering LMMs with 3D-awareness

Thank you for reading today’s edition.

Your feedback is valuable. Respond to this email and tell us how you think we could add more value to this newsletter.

Interested in reaching smart readers like you? To become an AI Breakfast sponsor, reply to this email!