Category: AI

Nvidia has entered into a non-exclusive licensing agreement with AI chip competitor Groq, a strategic move that includes hiring Groq’s founder Jonathan Ross and president Sunny Madra, along with other employees. While Nvidia clarified that this is not an acquisition of the company, CNBC reported the acquisition of assets from Groq for $20 billion, potentially marking Nvidia’s largest deal yet. This partnership positions Nvidia to strengthen its presence in the AI chip manufacturing market.

Amid the growing demand for computing power to enhance AI capabilities, Nvidia’s GPUs have established themselves as the industry standard. However, Groq has been focusing on a distinct chip variant known as an LPU (language processing unit). Groq claims that its LPU can operate LLMs (large language models) at speeds 10 times faster and with significantly lower energy consumption compared to traditional solutions.

Jonathan Ross, Groq’s CEO, is recognized for his contributions to the development of Google’s TPU (tensor processing unit), a specialized AI accelerator chip. Groq’s recent valuation of $6.9 billion following a $750 million funding round showcases the company’s rapid growth, with its AI applications now powering the work of over 2 million developers.

Source: TechCrunch

Waymo, the self-driving technology company, is reportedly testing the integration of Google’s Gemini AI chatbot into its fleet of robotaxis. This move aims to introduce an AI assistant that can engage with riders, provide general knowledge, control in-cabin features, and offer reassurance when needed. The findings, uncovered by researcher Jane Manchun Wong, reveal a detailed 1,200-line system prompt outlining the expected behavior of this in-car AI companion. While not yet available in public builds, Waymo’s spokesperson mentioned ongoing efforts to enhance the rider experience through innovative features.

Gemini’s capabilities have been leveraged by Waymo previously to enrich its autonomous driving technology, enabling vehicles to handle complex and uncommon scenarios effectively. The AI assistant is designed to be friendly, helpful, and unobtrusive, focusing on enhancing rider safety and comfort. Clear communication, avoidance of technical language, and a seamless integration into the autonomous vehicle ecosystem are key aspects guiding the development of this AI-powered assistant.

As Waymo continues to explore the possibilities of autonomous transportation, the potential integration of Gemini AI into its robotaxis signifies a step towards delivering a more interactive and informative ride experience for passengers.

Source: TechCrunch

The Italian Competition Authority (AGCM) has ordered Meta to suspend its policy prohibiting companies from deploying their own AI chatbots using WhatsApp’s business tools. The AGCM intervened following concerns of potential abuse of market dominance by Meta in restricting the use of AI chatbots on WhatsApp.

The AGCM highlighted that Meta’s actions could impede competition and technological advancements in the AI chatbot services sector. Meta’s alteration to its API policy, set to exclude general-purpose chatbots, has prompted investigations and regulatory scrutiny from authorities like the European Commission.

While Meta argues that its API is not intended as a chatbot distribution platform, critics fear the policy change may limit user choices and stifle innovation. Notably, businesses leveraging AI for customer service on WhatsApp are unaffected by the ban on AI chatbots from external providers like OpenAI and Perplexity.

The European Commission’s involvement underscores the broader implications of Meta’s policy adjustments, particularly in the European Economic Area. The ongoing investigations reflect a growing focus on ensuring fair competition and consumer choice within the evolving landscape of AI-powered services.

Source: TechCrunch

Lemon Slice, a digital avatar generation company, is expanding the capabilities of AI chatbots by introducing a video layer through its new diffusion model. This technology allows for the creation of customizable digital avatars from a single image, enabling more interactive AI agents within apps.

The Lemon Slice-2 model enables the generation of dynamic digital avatars that can engage with users, addressing tasks ranging from customer queries to mental health support. Co-founder Lina Colucci highlighted the importance of video interactivity, drawing inspiration from tools like ChatGPT.

Lemon Slice’s solution, offered through an API and embeddable widget, features a 20-billion-parameter model capable of livestreaming videos at 20 frames per second on a single GPU. Users can customize avatar backgrounds, styles, and appearances to cater to diverse requirements.

Furthermore, Lemon Slice focuses on creating both human-like and non-human avatars, leveraging ElevenLabs’ technology for generating avatar voices. By utilizing a general-purpose diffusion model, the company aims to differentiate itself from competitors and deliver superior avatar solutions.

Source: TechCrunch

Amazon has announced the expansion of its AI-powered digital assistant Alexa+ with four new integrations in 2026. The company has partnered with Angi, Expedia, Square, and Yelp, enabling users to conveniently book hotels, request home service quotes, and schedule salon appointments, among other functions.

For instance, customers can now utilize Expedia through Alexa+ to compare, book, and manage hotel reservations or receive personalized recommendations based on their preferences, such as finding pet-friendly hotels in specific locations. The addition of these services complements Alexa+’s existing partnerships with platforms like Fodor, OpenTable, Suno, Ticketmaster, Thumbtack, and Uber.

By enabling seamless interactions with popular online services, Amazon aims to simplify consumer experiences through its digital assistant. Users can now leverage Alexa+ to perform tasks like calling an Uber or making OpenTable reservations, all while engaging in natural language conversations with the AI assistant.

The evolving model of integrating AI into everyday consumer activities emphasizes the importance of user-friendly experiences to drive adoption. Early adopters of Alexa+ have shown strong engagement, particularly with home and personal service providers like Thumbtack and Vagaro.

Source: TechCrunch

A group of authors, including John Carreyrou, have filed a lawsuit against six major AI companies – Anthropic, Google, OpenAI, Meta, xAI, and Perplexity. The authors allege that the companies have trained their AI models using pirated copies of their books without authorization.

The dispute centers around a proposed class action settlement offered by Anthropic, which the authors believe does not adequately address the core issue of intellectual property infringement. The plaintiffs argue that the settlement primarily benefits the AI companies, allowing them to evade the true consequences of their alleged infringement practices.

This lawsuit underscores the ongoing debate surrounding intellectual property rights in the AI industry and raises questions about the ethical implications of using copyrighted material without authorization. It also highlights the complexities of balancing innovation with legal and ethical responsibilities in the development of AI technologies.

Source: TechCrunch

OpenAI has acknowledged the persistent risk of prompt injections in AI browsers, particularly those with agentic capabilities like Atlas. Despite efforts to enhance cybersecurity using an ‘LLM-based automated attacker,’ the company recognizes that prompt injections remain a significant challenge in ensuring the safe operation of AI agents on the web.

In a recent blog post, OpenAI highlighted the complexity of prompt injections, which involve manipulating AI agents to execute malicious commands embedded in web content or emails. The company emphasized that prompt injections, similar to scams and social engineering tactics, are unlikely to be fully eradicated, underscoring the continuous need for robust security measures.

Since the launch of its ChatGPT Atlas browser, OpenAI has been vigilant in fortifying defenses against prompt injection attacks. Security researchers have demonstrated how AI-powered browsers can be compromised through subtle manipulations, prompting concerns about the security vulnerabilities introduced by ‘agent mode’ in ChatGPT Atlas.

Notably, the U.K.’s National Cyber Security Centre has also highlighted the persistent threat posed by prompt injection attacks on generative AI applications, cautioning that complete mitigation may not be achievable. As a proactive measure, OpenAI has adopted a rapid-response strategy to proactively identify and address novel attack vectors, aiming to preemptively counter emerging threats.

Source: TechCrunch

OpenAI’s ChatGPT, known for its conversational AI capabilities, is launching a new feature reminiscent of Spotify Wrapped, called ‘Your Year with ChatGPT.’ This annual review, available to eligible users in select markets like the U.S., Canada, the U.K., Australia, and New Zealand, provides a personalized summary of users’ interactions with the chatbot throughout the year.

The feature, accessible through the ChatGPT web and mobile apps for iOS and Android, offers awards, poems, and images based on individual usage patterns. For example, users may receive awards like the ‘Creative Debugger’ for problem-solving interactions. The experience is designed to be user-controlled and privacy-focused.

Similar to Spotify Wrapped, which offers a personalized music listening summary, ‘Your Year with ChatGPT’ leverages engaging visuals and tailored content to reflect each user’s unique chatbot interactions. While the feature is opt-in and not intrusive, it provides a fun and nostalgic way for users to reflect on their year in conversations with ChatGPT.

Source: TechCrunch

OpenAI’s ChatGPT, the text-generating AI chatbot, has gained significant traction since its debut in November 2022, attracting a user base of 300 million weekly active users. Originally designed to enhance productivity through writing assistance, ChatGPT has evolved into a prominent force in the AI landscape.

In 2025, OpenAI faced challenges amid competition from Chinese AI firms like DeepSeek, prompting strategic shifts and heightened focus on its flagship chatbot. The company also engaged in significant initiatives such as data center expansions and fundraising endeavors, positioning itself for future growth.

Reflecting on 2024, OpenAI forged key partnerships, launched innovative products like GPT-4 featuring voice capabilities, and introduced groundbreaking models like Sora for text-to-video generation. However, the company also experienced internal dynamics, including the departure of key executives and legal conflicts, such as copyright disputes and regulatory interventions.

Throughout the year, OpenAI continued to update ChatGPT, unveiling new features and improvements to enhance the user experience. The chatbot’s evolution underscores OpenAI’s commitment to advancing AI capabilities and maintaining relevance in a competitive market.

Source: TechCrunch

Waymo, the autonomous driving technology company, temporarily halted its robotaxi service in San Francisco following a significant power outage that left many of its vehicles stranded on city streets. The blackout, caused by a fire at a Pacific Gas & Electric substation, affected around 120,000 customers, disrupting Waymo’s operations.

Social media posts showed Waymo robotaxis stopped at roads and intersections during the blackout. The company cited the power failure as the reason for the service suspension, with spokesperson Suzanne Philion assuring customers that they are working closely with local officials to restore service promptly.

While Waymo did not provide specific details on the disruption, the incident highlights the vulnerability of autonomous vehicle operations to infrastructure failures and underscores the importance of robust contingency plans. The blackout also affected Muni mass transit and cell service, prompting San Francisco Mayor Daniel Lurie to advise residents to avoid unnecessary travel.

Despite efforts to restore power, thousands of customers remained without electricity on Sunday morning, prolonging the disruption. Waymo’s response to the blackout illustrates the challenges faced by autonomous vehicle companies in ensuring uninterrupted service delivery amidst unforeseen circumstances.

Source: TechCrunch

OpenAI has introduced a new feature that allows ChatGPT users to adjust the chatbot’s tone and expression. This update, announced on social media, enables users to access customization options within the Personalization menu, offering settings to control the levels of warmth, enthusiasm, and emoji usage.

In addition to these adjustments, users can already define a ‘base style and tone,’ which includes options like Professional, Candid, and Quirky tones introduced earlier. This move comes after ChatGPT faced tone-related controversies, prompting OpenAI to reverse an overly flattering update and modify GPT-5 to exude a ‘warmer and friendlier’ demeanor based on user feedback.

While some users have welcomed these customization features as enhancing the experience, concerns have been raised by academics and AI critics regarding chatbots’ tendency to excessively praise and validate users, potentially fostering addictive behaviors and negatively impacting mental well-being.

Source: TechCrunch

New York Governor Kathy Hochul has signed the RAISE Act into law, solidifying the state’s position as a leader in AI safety regulations. The act mandates that major AI developers disclose details about their safety procedures and promptly report any safety incidents to the state within 72 hours. Additionally, a new oversight office will be established within the Department of Financial Services to supervise AI advancements.

The bill’s enforcement includes fines of up to $1 million for initial violations and up to $3 million for subsequent breaches if companies fail to comply with safety reporting requirements or provide false information. California has already enacted a similar law, which Governor Hochul referenced as a model for New York’s legislation.

Despite initial industry concerns, Governor Hochul’s decision to sign the bill demonstrates a commitment to prioritizing public safety amid the rapid expansion of AI technologies. The move aligns New York with California’s regulatory framework and underscores the need for consistent AI safety standards across tech-forward states in the absence of federal mandates.

Source: TechCrunch

Meta, the parent company of Facebook, is expanding its AI capabilities with the development of two new models set for a 2026 release. The AI models, codenamed ‘Mango’ for image and video processing and ‘Avocado’ for text-based functions, are part of Meta’s efforts to enhance its AI offerings. These developments stem from Meta’s AI research division, led by Scale AI co-founder, Alexandr Wang.

According to a report by The Wall Street Journal, Meta plans to launch these new models in the first half of 2026, as revealed during an internal Q&A session led by Wang and Meta’s chief product officer, Chris Cox. The focus is on refining the text-based model to improve coding abilities and exploring novel world models capable of comprehending visual data, reasoning, planning, and acting autonomously without extensive training.

Despite facing competition from industry rivals like OpenAI and Google, Meta aims to strengthen its AI division with these upcoming releases. Recent restructuring within the AI division, including leadership changes and talent acquisitions, underscores Meta’s commitment to AI innovation. The departure of key figures like Yann LeCun to pursue independent ventures reflects the evolving landscape within Meta’s AI ecosystem.

As Meta continues to develop its AI capabilities, the success of ‘Mango’ and ‘Avocado’ will be crucial in shaping the company’s future AI endeavors and maintaining its competitive edge in the tech landscape.

Source: TechCrunch