Apple Boosts AI in China, Sony and AMD Revolutionize Gaming, GitHub Offers Free Copilot Tier, Perplexity AI Secures $500M Funding, and more...
11:00 Share

Good morning. It's December 19th and this is your daily brief in AI. Here's everything you need to know. Apple is taking significant steps to enhance its presence in the Chinese smartphone market by pursuing partnerships with Tencent and ByteDance for artificial intelligence integration. With over 900 million smartphone users, China represents the largest smartphone market globally, making these collaborations crucial for Apple.

CEO Tim Cook recently emphasized during an earnings call that integrating AI into iPhones is essential for bolstering Apple's competitive position in China, particularly as domestic rivals like Huawei have increased their market share significantly. Huawei's smartphone sales have surged by 42 percent, while Apple's sales in China have seen a slight decline, highlighting the intensifying competition.

Analysts point out that local companies are heavily investing in AI features, raising customer expectations and creating a challenging environment for Apple. Furthermore, the Chinese government requires foreign firms to partner with local companies to launch AI products, adding another layer of complexity for Apple as it navigates these regulatory requirements.

This strategy aligns with Apple's approach in the U.S., where it collaborates with local firms to ensure regulatory compliance. The new AI features, branded as "Apple Intelligence," aim to encourage consumers to upgrade to newer iPhone models, especially after Apple missed sales estimates in Greater China last quarter. Despite facing skepticism regarding its AI initiatives, Apple plans to introduce additional features by April 2025.

Research indicates a strong consumer interest in AI-powered voice assistants, with 63% of users expressing a desire for such technology. Apple is already managing user data storage in compliance with Chinese regulations through existing partnerships, which is vital for its operations in the region. Without these partnerships, Chinese consumers could miss out on generative AI functionalities becoming available in other markets.

Mark Cerny, the chief architect behind the PlayStation 5 and PS5 Pro, has announced a new collaboration between Sony and AMD focused on advancing machine learning technologies for video games, codenamed Amethyst. This initiative aims to significantly enhance game graphics, particularly through the greater implementation of ray tracing and path tracing techniques.

The PS5 Pro, introduced in September 2024 at a price point of $700, retains the RDNA 2.x architecture for its GPU, but incorporates features from RDNA 3 and is set for future advancements with RDNA 4. Despite some visual glitches in certain titles, such as the Silent Hill 2 remake, the PS5 Pro is performing well in the market, with developers actively working on patches to improve the gaming experience.

Cerny believes that substantial generational leaps in console technology are still possible, even as the differences between iterations may appear less pronounced. He stressed the necessity for games to demonstrate significant performance improvements to justify the release of new consoles. Looking ahead, details regarding the upcoming PlayStation 6 are beginning to surface, with indications that AMD technology will play a vital role in its development.

Future advancements in machine learning are expected to enhance various aspects of game graphics, potentially justifying the need for a next-generation console. The collaboration between Sony and AMD aims to create a "democratized library for machine learning" that will empower developers to craft more interactive gaming experiences. This partnership further solidifies AMD's position in the gaming hardware market, allowing it to compete more effectively with Nvidia.

Cerny also highlighted the challenges associated with developing efficient machine learning hardware, noting the distinctions between different performance metrics. GitHub has made a significant move by launching a free tier for its Co-Pilot AI Assistant, now integrated directly into Visual Studio Code. This new offering is aimed at broadening access for developers, particularly those who may have limited budgets. The free version is designed for occasional users, allowing up to 2,000 code completions per month.

This includes all suggestions made, not just those accepted, providing a robust resource for developers looking to enhance their coding experience. With Copilot Free, users can ask coding questions, identify bugs, and search across multiple files for solutions. While it retains core functionalities, the free version does have limitations, such as usage caps and the absence of enterprise-level features, making it less suitable for larger teams.

This initiative is expected to transform coding practices, optimizing workflows and boosting productivity, especially among students and hobbyists. Previously, Copilot required a subscription starting at $10 per month with free access limited to verified students, teachers, and open-source maintainers.

Users on the free plan will have access to specific AI models, including Anthropic's CLAWD 3.5 Sonnet and OpenAI's GPT-4.0, while paid plans offer broader access to additional models. Upcoming updates in January 2024 are expected to enhance user-friendliness, including clearer usage indicators and notifications for users nearing their limits.

This launch follows GitHub's recent milestone of surpassing 150 million developers on its platform, reflecting significant growth and community engagement. CEO Thomas Domke emphasized that "Co-pilot-free" builds on GitHub's tradition of providing free tools for developers, reinforcing the company's commitment to accessibility.

Perplexity AI Inc. has just made headlines by securing $500 million in funding, elevating its valuation to an impressive $9 billion. This marks a significant leap, tripling its valuation from $3 billion in June 2024, following a major investment from SoftBank. Earlier this year, the company was valued at just $520 million, showcasing remarkable growth over the past months.

In a strategic move to enhance its technology, Perplexity has acquired Carbon, a data connectivity firm, although the financial details of the acquisition have not been disclosed. This acquisition is aimed at improving Perplexity's answer engine, allowing it to access a broader range of information and streamline the research process for its users.

As of March 2024, Perplexity reported over 15 million active users, processing hundreds of millions of queries each month, predominantly in the U.S. The company operates on a subscription-based revenue model, with annual revenues soaring from $5 million in January to $35 million by August 2024.

Despite its rapid growth, Perplexity faces stiff competition from established AI giants like OpenAI, Microsoft, and Google, all of which are enhancing their search capabilities. Additionally, the company is currently dealing with legal challenges, having been sued by two publishers owned by News Corporation for allegedly using their copyrighted content without permission.

In light of these challenges, CEO Aravind Srinivasan has indicated plans to establish revenue-sharing partnerships with news publishers. Perplexity's strong backing from notable investors, including Jeff Bezos and NVIDIA Corporation, reflects significant industry support for the startup. The recent funding round highlights a growing investor interest in transforming online search through generative AI technology.

GenBioAI is at the forefront of a revolutionary shift in biological research with its AI-driven digital organism, or AIDO. This innovative technology leverages multi-scale foundation models to simulate, predict, and program biological outcomes, moving beyond traditional methodologies.

The AIDO models incorporate a wide array of biological components, including DNA, RNA, proteins, and evolutionary data, facilitating a comprehensive understanding of biological systems.

Supported by experts from prestigious institutions like Carnegie Mellon University and Stanford, GenBio AI showcased its advancements at the NeurIPS conference. Notably, the AIDO protein foundation model is highlighted for its computational efficiency, enabling in-depth exploration of protein functionalities.

Additionally, AIDO accelerates drug discovery processes by allowing researchers to rapidly simulate and evaluate millions of compounds, significantly cutting down on both development time and costs. The AIDOR-NA Foundation model, with 1.6 billion parameters, is the largest of its type, improving speed and precision in RNA structure prediction and drug design.

Meanwhile, the AIDO single-cell model processes the entire human transcriptomy, learning detailed representations of human cells while avoiding data loss, which enhances our understanding of cellular transcriptional landscapes. The AIDODNA Foundation model, trained on data from 796 species and consisting of 7 billion parameters, advances the capabilities of DNA language models without needing new data.

This technology also aids in comprehending complex diseases by examining systemic interactions and identifying new pathways for intervention. Furthermore, AIDO enhances personalized medicine by creating digital patient twins, which assist in designing tailored treatments and reducing adverse drug reactions.

Dr. LeSong, co-founder and CTO of GenBioAI, envisions a future where interactive digital models of biological systems enable safe experimentation and precise modifications, fundamentally transforming the landscape of biological research. This has been your daily brief in AI. To read more about these stories, follow the links in the episode bio.

You can also subscribe to these updates via email at www.brief.news. For more daily podcasts about the topics you love, visit brief.news forward slash podcasts. Tune in tomorrow. We'll be back with everything you need to know.