Lionsgate has partnered with AI firm Runway to create an AI model for their film and TV content. Executives claim filmmakers are using the tool, emphasising it will be used within guardrails and may reduce costs. The deal has reportedly received support from industry partners after clarification. This partnership marks a significant step in AI integration in filmmaking.
Amazon is in talks to make another multibillion-dollar investment in AI company Anthropic, following a $4 billion investment last year. The new investment is expected to be of similar scale.
Samsung trails SK Hynix in producing high-bandwidth memory (HBM), crucial for AI chips used by Nvidia. This has impacted Samsung's profits and market value. SK Hynix secured Nvidia's approval for its HBM, while Samsung awaits approval. Samsung claims its HBM3E is in production and aims to mass-produce next-gen HBM4 in late 2025.
OpenAI has hired Meta's former AR hardware lead to spearhead their hardware efforts, initially focusing on robotics and partnerships. This move aligns with OpenAI's collaboration with Jony Ive on AI-powered hardware, aiming to redefine human-technology interaction.
Two AI startups, Anysphere and Mercor, are in funding talks at £1.6 billion+ valuations each. Anysphere developed a code generation assistant, while Mercor provides staffing for AI companies. Both have received investor term sheets.
Google has launched Vids, an AI-powered video creation tool for select Workspace editions. Using the Gemini AI model, it creates workplace and marketing videos from Google Drive files and user descriptions. Vids offers templates, script suggestions, and a virtual recording studio. It compiles slideshows based on source documents or user input, rather than generating videos from scratch. The feature is being rolled out to various Business and Enterprise tier Workspace members.
Andreessen Horowitz, a venture capital firm, has announced a new investment focus on AI-powered parenting tools. They are interested in companies developing AI applications for various parenting tasks, including sleep monitoring and pregnancy guidance. The firm suggests AI could potentially act as a constant parenting companion, though this field is still in its early stages.
Google accidentally revealed Jarvis, an AI program designed to perform web-based tasks by taking over a user's computer. It was briefly available as a Chrome extension before being removed. Jarvis is similar to Anthropic's Claude 3.5 Sonnet in its ability to operate computers like a human. Such AI agents could potentially handle tasks like purchasing tickets or making reservations.
The Washington Post has launched 'Ask The Post AI', a chatbot that answers questions using the newspaper's articles since 2016. It uses curated data, won't respond without sufficient reporting, and may produce inaccurate information. Users are advised to verify results with published articles. This initiative explores new ways of presenting journalism using AI technology.
Apple is developing a feature to prevent voice assistants from responding to unintended triggers in adverts. Using the Shazam API, it matches audio from device microphones against downloaded fingerprints. When matched, the Siri trigger is disabled. Currently in the HomePod's software, it may expand to other Apple devices.
AI is being used by some tattoo designers to create mock-up designs, sparking debate in the industry. Proponents praise AI's design capabilities, while critics view it as 'cheating'. Some clients have been disappointed to learn their designs were AI-generated. AI designs can create unrealistic expectations about tattoo outcomes. This trend reflects AI's broader impact on creative industries, raising questions about human creativity in artistic fields.
OpenAI has launched SearchGPT, an AI-powered search engine integrated into ChatGPT for paid subscribers. It offers real-time data and a conversational search experience. OpenAI has partnered with major news organisations for reliable sources. This positions OpenAI as a competitor to Google in search. The feature will expand to other user types soon, with no current plans for advertising. It may impact future SEO strategies.
Cal Poly Pomona's annual Cybersecurity Fair focused on AI's role in cybersecurity. Demonstrations showcased AI's potential for creating and defending against cyber threats. Industry experts discussed AI threats and quantum computing. The event included interactive demonstrations, research presentations, and a job fair. It emphasised opportunities for women in cybersecurity and featured a poster contest.
AI-powered lending company Upstart reported strong financial results, with loan volumes up 43% quarter-over-quarter. They forecast £180 million Q4 revenue, beating expectations. The stock rose 20% in after-hours trading.
Gray Swan AI, a security startup founded in 2023, focuses on identifying risks in AI systems and developing tools for safe deployment. They've secured partnerships with major AI companies and the UK's AI Safety Institute. Their key products include Cygnet, an AI model resistant to jailbreaking, and Shade, a tool for finding AI system weaknesses. The company hosted a 'jailbreaking arena' competition to test their security measures, has received £5.5 million in seed funding, and is preparing for Series A funding. They aim to build a community of hackers to continually improve AI safety measures.
Anthropic has partnered with AWS and Palantir to provide AI models to U.S. defence and intelligence agencies. AI-related federal contracts in the U.S. have increased significantly, with the Department of Defense as a major player. Other tech companies like Meta and Microsoft have made similar partnerships for defence applications.
French AI startup Mistral AI has launched a content moderation API using their Ministral 8B model. It detects harmful content in 9 categories across 11 languages, analyzing both text and conversations. Available through Mistral's cloud platform, this positions them to compete with major AI companies on safety issues. Recent partnerships with Microsoft, Qualcomm and SAP boost Mistral's enterprise presence. Their focus on edge computing and safety features may appeal to European firms under strict data regulations.
A phishing campaign is impersonating OpenAI, claiming ChatGPT subscription payments were declined and urging recipients to update payment details via a link. While this attack is unsophisticated, future AI-driven attempts may be more advanced. Experts recommend strengthening cybersecurity measures and providing employee training to combat such threats.
NVIDIA AI Summit Japan, an invite-only event in Tokyo, will feature NVIDIA and SoftBank CEOs discussing AI's transformative role. The event will showcase AI innovations, including generative AI and robotics, and explore Japan's AI infrastructure initiative and sovereign AI. Sessions will cover AI's impact on Japanese policy and business strategy, as well as workshops on AI development.
ChatGPT's coding abilities have significantly improved. In a recent test, it quickly created a complex Tic-Tac-Toe variant on a 12x12 grid, providing full HTML and JavaScript code. The AI showed enhanced efficiency, clarity and ability to explain advanced programming concepts compared to its previous version.
