Apple Seeks AI Partners as OpenAI Reports Billions in Losses

The artificial intelligence sector is currently navigating a complex landscape marked by significant executive movements, profitability challenges, and evolving ethical and legal considerations. Apple, for instance, has seen considerable executive flux, with its senior vice president of hardware technologies, Johny Srouji, deciding to remain with the company, a move crucial for its custom chip and AI future. This comes amidst other high-profile departures, including Apple's head of AI and top lawyer, and the Vision Pro software lead moving to Meta. Apple is reportedly struggling in the fast-paced AI race, especially after its Siri revamp fell short, and now plans to collaborate with external AI developers for future innovations like rumored AI glasses and a foldable iPhone by 2026. Profitability remains a major hurdle for many AI companies. OpenAI, a prominent player, reported billions in losses for 2024 and the first half of 2025, underscoring the high costs associated with creating and operating generative AI. Despite AI's potential to save businesses money, a Stanford report indicates that only 3% of US adults pay for AI services, suggesting the current subscription model may not be sustainable. Roblox CEO David Baszucki also highlighted the difficulty in keeping up with the rapid and complex pace of AI research, noting that major companies like Google are increasingly privatizing their findings. New AI products are emerging, but not without controversy. ByteDance, known for TikTok, launched an AI-powered phone, the Nubia M153, featuring its Doubao Mobile Assistant. However, rival companies like Tencent and Alibaba quickly blocked Doubao from their platforms, citing security concerns over its broad permissions and data handling, particularly with financial transactions. ByteDance has temporarily paused some AI features to address these issues. Meanwhile, Andrej Karpathy, former Tesla AI Director, advises users to interact with Large Language Models by asking what a group of experts would think, rather than what the LLM itself thinks, emphasizing that these models are simulators without personal opinions. The creative and legal implications of AI are also coming to the forefront. AI-generated music, while improving, still raises questions about lyrical quality and the changing role of artists, necessitating new quality and ethics standards. Alexis Ohanian, co-founder of Reddit, predicts that the rise of AI-generated online content, with about 20% of social media content being fake, will increase demand for authentic, real-world experiences and live entertainment. Furthermore, a German court's landmark ruling that AI training requires copyright holder permission is creating a global divide in how AI content is regulated, potentially pushing European developers to license content rather than using unlicensed data.

Key Takeaways

• Apple's hardware chief, Johny Srouji, will remain with the company, securing its custom chip and AI future, despite other senior executive departures, including its head of AI.
• Apple faces challenges in the AI race, with its Siri revamp failing, and plans to partner with outside AI developers for future products like rumored AI glasses and a foldable iPhone by 2026.
• OpenAI reported billions in losses for 2024 and the first half of 2025, highlighting the high costs of generative AI and the struggle for profitability in the industry.
• ByteDance launched an AI-powered phone, Nubia M153, with ZTE, featuring the Doubao Mobile Assistant, which was quickly blocked by rivals like Tencent and Alibaba due to security and data concerns.
• Andrej Karpathy, former Tesla AI Director, advises users to ask Large Language Models what a group of experts would think, rather than what the LLM "thinks," as LLMs are simulators without personal opinions.
• Roblox CEO David Baszucki finds the rapid pace of AI research challenging to follow, noting that major companies like Google are now keeping their findings private.
• AI-generated music, while improving, still raises questions about lyrical quality, artistic roles, and the need for new quality and ethics standards in the music industry.
• Alexis Ohanian, Reddit co-founder, believes AI's proliferation of fake online content (20% of social media) will increase demand for real-world experiences and live entertainment.
• A German court ruled that AI training requires permission from copyright holders, creating a global split in copyright enforcement and potentially pushing European AI developers to license content.
• Many AI companies struggle to profit due to high operational costs per response, with only 3% of US adults paying for AI services, indicating a need for new business models.

Alexis Ohanian sees AI boosting live entertainment

Alexis Ohanian, co-founder of Reddit, spoke at the BRIDGE Summit in Abu Dhabi. He believes artificial intelligence makes online content less human, with about 20% of social media content being fake. This shift will make people want more real-world experiences. Ohanian feels live events will become a major positive outcome of the AI era.

Alexis Ohanian sees AI boosting live entertainment

Alexis Ohanian, co-founder of Reddit, spoke at the BRIDGE Summit in Abu Dhabi. He believes artificial intelligence makes online content less human, with about 20% of social media content being fake. This shift will make people want more real-world experiences. Ohanian feels live events will become a major positive outcome of the AI era.

Apple's Johny Srouji remains, securing chip and AI future

Johny Srouji, Apple's senior vice president of hardware technologies, will stay with the company. There was speculation he might leave, causing worry about Apple's future in custom chips and AI. Srouji joined Apple in 2008 and led the creation of A-series and M-series chips for iPhones, iPads, and Macs. His decision to remain helps Apple keep its strong position in chip design and continue its work in AI.

Apple faces executive changes amid AI challenges

Apple is seeing many senior executives leave or consider leaving, including two direct reports to CEO Tim Cook. The head of AI and the top lawyer have departed, and the Vision Pro software lead moved to Meta. Hardware chief Johny Srouji initially thought about leaving but decided to stay. These changes come as Apple struggles in the fast-moving AI race, especially after its Siri revamp failed. The company now plans to partner with outside AI developers and has major product launches, like AI glasses and a foldable iPhone, rumored for 2026.

ByteDance AI phone raises security concerns for rivals

ByteDance, the company behind TikTok, launched an AI-powered phone called Nubia M153 with ZTE. This phone features the Doubao Mobile Assistant, which has wide access to the device's functions. However, rival companies like Tencent and Alibaba quickly blocked Doubao from their apps, citing security worries. They are concerned about Doubao's broad permissions and how it handles user data, especially with financial transactions. ByteDance has temporarily stopped some AI features like gaming and payments to address these concerns and plans to create clear guidelines with hardware partners.

Andrej Karpathy advises new way to ask AI questions

Andrej Karpathy, former Tesla AI Director, suggests a new way to talk to Large Language Models or LLMs. He advises people not to ask LLMs what "you" think. Instead, he recommends asking what a group of experts would think about a topic. Karpathy explains that LLMs are like simulators and do not have their own thoughts or opinions. This approach helps users get more useful information from AI.

Many AI companies struggle to make a profit

Many companies focused on AI are finding it hard to make money, even though their products are valuable. Generative AI is very costly to create and operate, with high expenses for each response it generates. A Stanford report shows AI helps businesses save money, but only 3% of US adults pay for AI services. OpenAI, a major AI company, reported billions in losses in 2024 and the first half of 2025. The current subscription model does not work well for AI due to these high costs, so the industry needs to find a better way to sell its products.

Roblox CEO finds AI research hard to follow

Roblox CEO David Baszucki admits it is challenging to keep up with the rapid pace of new AI research. He described the experience as "humbling" because the field is growing so quickly with complex topics. Baszucki noted that AI research has expanded greatly, with major companies like Google now keeping their findings private. He believes pure research is making a comeback, but also pointed out that AI is still in its early stages for 3D applications. This is because AI models are mostly trained on human-made text and images, not real-world 3D data.

AI generated music raises questions for artists

A reporter listened to music created by artificial intelligence, including artists like The Velvet Sundown and Breaking Rust. The Velvet Sundown's 70s rock sound was almost convincing, but the lyrics were uninspired. Breaking Rust's country music was described as bad, with strange instruments and unnatural vocals. While some AI music is getting harder to tell apart from human-made songs, it raises important questions. This includes the changing role of artists and the need for new quality and ethics standards in the music industry.

Stay creative and human in the AI era

An article from the Rolling Stone Culture Council shares advice on staying creative and human in the age of AI. It suggests resisting autopilot thinking and taking time to form your own ideas, as AI can generate content too quickly. The author emphasizes that emotional accuracy and lived experience are things AI cannot replicate. It also advises anchoring innovation in personal values like integrity and imagination, rather than just chasing speed. In a world becoming more automated, building trust through honesty and lived experience will become a key advantage for creative professionals.

AI training copyright rules are splitting globally

A German court recently made a landmark ruling that AI training requires permission from copyright holders. This decision is creating a split in how different parts of the world handle AI copyright. In the US, courts often look at whether AI training is "fair use" or causes market harm. This global difference means AI companies might need to create different training systems for various regions. For music rights holders, this raises questions about where to pursue legal action and how to enforce their rights. If the German ruling stands, it will likely push more AI developers in Europe to license content rather than using unlicensed data.

Sources

• Why AI Makes Alexis Ohanian ‘Bullish’ About Live Entertainment
• Why AI Makes Alexis Ohanian ‘Bullish’ About Live Entertainment
• Apple’s Johny Srouji Stays, Securing Future in Custom Chips and AI
• Apple Faces Leadership Turmoil As AI Setbacks And 2026 Product Plans Raise Pressure
• ByteDance AI phone sparks security fears from Tencent and Alibaba
• Andrej Karpathy On Why Instead Of Asking LLMs What “You” Think, People Should Ask What Experts Think
• Are AI companies sustainable?
• Finding it hard to keep up with AI research? Roblox's CEO says you're not alone
• I listened to AI-generated music so you don't have to
• Culture Council: Staying Creative and Human in the Age of AI
• Licensed or lost? In the future of AI training, "the world is splintering"