The article explores how Wall Street firms are increasingly embracing open-source AI technology, marking a significant shift from their traditionally secretive approach to technology. Major financial institutions are now actively contributing to and utilizing open-source AI tools, particularly in areas like large language models and machine learning frameworks. This transformation is driven by several factors, including cost efficiency, talent attraction, and the need to keep pace with rapid technological advancement. The article highlights how firms like Goldman Sachs, Morgan Stanley, and JPMorgan are participating in open-source AI projects and sharing some of their tools with the broader community. This collaborative approach helps banks avoid duplicating efforts and allows them to focus on developing proprietary applications that provide competitive advantages. The piece also discusses how this trend is expected to accelerate through 2025, with financial institutions becoming more prominent players in the open-source AI ecosystem. However, banks are maintaining a balanced approach, carefully deciding which technologies to share and which to keep proprietary. The article emphasizes that this shift represents a cultural change in financial technology, where collaboration and community contribution are becoming as important as internal innovation. It concludes by noting that open-source AI is becoming a crucial element in financial institutions’ technology strategies, though they continue to maintain proprietary control over their most critical competitive advantages.

A new AI model from DeepSeek has analyzed potential market impacts of artificial intelligence, identifying companies that could face significant challenges as AI technology advances. The analysis suggests that certain sectors and companies heavily dependent on traditional business models could see substantial value erosion by 2025. The model specifically highlights concerns for companies in sectors like traditional software, consulting services, and digital advertising, where AI automation could reduce human workforce needs and transform service delivery methods. DeepSeek’s analysis indicates that companies slow to adapt to AI integration might lose market share to more technologically agile competitors. The report emphasizes that businesses with high labor costs in areas that can be automated by AI are particularly vulnerable. However, the model also notes that companies actively investing in AI transformation and developing AI-enhanced products and services are better positioned to maintain market value. The analysis suggests a potential market redistribution where early AI adopters gain at the expense of slower-moving competitors. Key takeaways include the urgency for businesses to develop comprehensive AI strategies, the importance of workforce adaptation to AI technologies, and the potential for significant market value shifts based on AI adoption rates. The model’s predictions underscore the transformative impact of AI on business valuations and the critical nature of technological adaptation in maintaining market competitiveness through 2025.

The article profiles Anima Anandkumar, a prominent AI researcher and Caltech professor, who received recognition at TIME100 Impact Awards for her contributions to artificial intelligence. Anandkumar’s work focuses on making AI more responsible, ethical, and inclusive. She emphasizes the importance of developing AI systems that are environmentally sustainable and that serve diverse global populations. As the director of machine learning research at NVIDIA, she has been instrumental in advancing AI technologies while advocating for increased representation of women and underrepresented groups in the field. The article highlights her innovative approach to AI development, including her work on “tensor algorithms,” which have become fundamental to modern machine learning. Anandkumar strongly advocates for democratizing AI access and ensuring its benefits reach beyond just wealthy nations and corporations. She warns against the concentration of AI power in the hands of a few companies and emphasizes the need for open-source alternatives. The piece also discusses her personal journey as a woman in tech and her efforts to mentor the next generation of AI researchers. Her vision for AI’s future includes making it more energy-efficient, accessible to developing nations, and capable of serving diverse global needs. The article concludes by highlighting her role in shaping both the technical and ethical dimensions of AI development.

A groundbreaking initiative combining artificial intelligence and citizen science is making progress in decoding ancient scrolls that were carbonized by Mount Vesuvius’s eruption in 79 CE. The Vesuvius Challenge, launched with a $1 million prize pool, has successfully revealed readable text from these delicate papyrus scrolls using AI-powered imaging analysis. Three computer scientists recently claimed the grand prize by deciphering four passages containing at least 140 legible Greek characters. Their breakthrough involved using machine learning algorithms to analyze high-resolution CT scans of the scrolls, detecting subtle variations in the papyrus surface that indicate the presence of ink. The team developed specialized AI models to enhance the contrast between the ink and the charred papyrus, making previously invisible text visible. This success has opened new possibilities for reading hundreds of other scrolls from the same collection, which were discovered in a villa near ancient Herculaneum. The decoded text appears to be a philosophical discussion about pleasure and luxury, possibly written by Epicurean philosopher Philodemus. This project demonstrates the powerful combination of modern technology and historical preservation, showing how AI can help unlock ancient knowledge that was previously thought lost forever. The success has encouraged further research, with additional prizes being offered for decoding more text from these valuable historical artifacts.

Legendary short seller Jim Chanos has issued a stark warning about the current state of financial markets, particularly focusing on AI-driven speculation and market bubbles. Chanos points to concerning parallels between today’s AI investment frenzy and previous market bubbles, specifically highlighting how AI has become a catalyst for speculative behavior similar to the crypto and meme stock phenomena. He expresses particular concern about the market’s reaction to AI developments, citing the example of how an AI chatbot called Deepseek caused significant market movements despite lacking substantial verification. The investor emphasizes that while artificial intelligence technology itself has merit, the current market valuations and investor behavior suggest a speculative bubble rather than rational investment decisions. Chanos predicts that by 2025, markets will likely face a reality check regarding AI valuations, suggesting that many current AI-related investments are overvalued and driven by hype rather than fundamental value. He draws attention to how companies are increasingly using AI-related announcements to boost their stock prices, similar to how blockchain mentions affected stocks in 2017. The analysis concludes with a warning about the risks of speculative excess in AI investments and the potential for significant market corrections when reality fails to meet inflated expectations.

The article discusses how AI startup Deepseek’s decision to switch from using Chinese AI models to developing its own highlights broader challenges faced by Western companies trying to access Chinese AI technology. Deepseek initially planned to build applications using existing Chinese language models but faced difficulties due to China’s strict data regulations and export controls. The company’s pivot to developing proprietary models reflects growing concerns about reliance on Chinese AI technology amid geopolitical tensions. The article emphasizes how China’s regulatory environment, including requirements for government approval of AI exports and data transfer restrictions, creates significant barriers for international collaboration in AI development. This situation has led to a “decoupling” in the AI industry between China and the West, with companies increasingly forced to choose sides or develop independent capabilities. The piece also notes that while Chinese AI models have shown impressive capabilities, particularly in Chinese language processing, access limitations are pushing companies to seek alternatives. Deepseek’s experience serves as a case study of how geopolitical tensions and regulatory constraints are reshaping the global AI landscape, potentially leading to parallel development paths in Eastern and Western markets. The article concludes by suggesting this trend could continue, with companies increasingly developing region-specific AI solutions rather than relying on cross-border collaboration.

A new bipartisan bill introduced in Congress aims to ban Chinese AI company DeepSeek and other foreign AI applications from U.S. government devices by 2025. The legislation, known as the Artificial Intelligence Security Protocol (AISP) Act, targets AI companies from China, Russia, North Korea, and Iran. The bill specifically mentions DeepSeek, which has gained attention for its advanced language models that rival those of OpenAI and Anthropic. The legislation comes amid growing concerns about national security risks posed by foreign AI technologies, particularly those with connections to countries considered adversarial to U.S. interests. DeepSeek, founded by former ByteDance employees, has raised significant funding and developed models that perform competitively against leading Western AI systems. The bill’s sponsors argue that foreign AI applications could potentially access sensitive government data and pose cybersecurity risks. The legislation would require federal agencies to remove these AI applications from government devices and prevent future installations. This move follows similar actions taken against TikTok and other Chinese-owned apps on government devices. The bill also mandates the creation of guidelines for identifying and evaluating potential security risks from foreign AI applications. Industry experts note that this legislation reflects broader efforts to secure government infrastructure against potential foreign technological threats while promoting domestic AI development.

DeepSeek AI emerges as a formidable competitor in the AI chatbot landscape, challenging established players like ChatGPT and Claude. The article examines DeepSeek’s capabilities, particularly highlighting its impressive coding abilities and mathematical problem-solving skills. DeepSeek’s free version outperforms ChatGPT 3.5 in several aspects, including generating more detailed responses and handling complex coding tasks. The AI model demonstrates superior performance in mathematical reasoning and shows a remarkable ability to maintain context in lengthy conversations. A key differentiator is DeepSeek’s transparent approach to its training data, which includes publicly available internet information and open-source code repositories. The platform offers both a free version and a paid pro version, with the latter providing access to more advanced features and larger context windows. Notable strengths include its ability to generate and explain code, solve complex mathematical problems, and provide detailed, nuanced responses to queries. However, the article notes that DeepSeek still faces challenges in certain areas and may occasionally produce hallucinations or incorrect information, similar to other AI models. The emergence of DeepSeek represents a significant development in the AI landscape, potentially offering users a powerful alternative to existing chatbots while maintaining competitive pricing and accessibility.

The article discusses France’s upcoming AI Safety Summit in Paris, led by Anne Bouverot, who emphasizes the importance of international cooperation in AI governance. The summit aims to build upon previous AI discussions in Bletchley Park and focuses on establishing concrete measures for responsible AI development. Bouverot highlights three main priorities: creating an international panel of scientific experts to assess AI risks, developing shared testing and evaluation methods for AI systems, and establishing common principles for AI governance. The summit specifically addresses concerns about frontier AI models and their potential risks, while acknowledging the need to balance innovation with safety. A key aspect is the involvement of both Western nations and Global South countries, ensuring diverse perspectives in AI governance discussions. The article emphasizes France’s role in promoting responsible AI development while maintaining competitiveness in the AI sector. Bouverot stresses the importance of including private sector stakeholders and addressing concerns about AI’s impact on jobs and society. The summit represents a significant step in creating a coordinated international approach to AI governance, with France positioning itself as a bridge between different global perspectives on AI regulation. The ultimate goal is to establish practical frameworks for AI development that protect society while fostering innovation.

A bipartisan group of House lawmakers is urging the Biden administration to ban the Chinese AI application DeepSeek, citing national security concerns. The legislators argue that DeepSeek, developed by Beijing Xingzhiyi Technology, poses similar risks to those that led to actions against Chinese-owned TikTok. The lawmakers, led by Reps. Mike Gallagher and Raja Krishnamoorthi, claim the app could collect sensitive data from American users and potentially share it with the Chinese government. They emphasize that DeepSeek’s AI capabilities, including its large language model that rivals ChatGPT, could be used to gather intelligence and manipulate information. The representatives point out that Chinese laws require companies to share data with their government when requested, making DeepSeek a potential conduit for surveillance and data collection. The lawmakers’ letter to Commerce Secretary Gina Raimondo requests that DeepSeek be added to the Commerce Department’s Entity List, which would effectively ban U.S. companies from doing business with it. This push reflects growing concerns about Chinese AI applications and their potential impact on national security, particularly as AI technology becomes more sophisticated and capable of processing vast amounts of user data. The initiative aligns with broader efforts to scrutinize and regulate Chinese technology companies operating in the United States.