Today:

• EU Antitrust Probe Microsoft

• ChatGPT Syncs with Nothing's Earbuds

• Japan Invests $470M in AI Development

• Google Unifies Android and Hardware for AI

• AdaKWS Sets New Bar for Speech AI

• Microsoft's Massive Chip Acquisition Plan

• Qtum's AI Expansion with Nvidia GPUs

Microsoft's OpenAI partnership could face EU antitrust probe, sources say

Microsoft's $13 billion investment in OpenAI might trigger an EU antitrust probe, sources say. Regulators are assessing how partnerships like Microsoft's with OpenAI, Alphabet, Amazon, and Anthropic impact competition. The EU is considering investigating whether Microsoft's involvement restricts competition or distorts the market. 

While the European Commission hasn't made a decision, it's examining evidence. Microsoft holds a non-voting position on OpenAI's board but denies ownership of ChatGPT. The Commission is evaluating if Microsoft's investment qualifies for review under EU merger rules. Microsoft and the Commission declined to comment.

REUTERS

ChatGPT is coming to Nothing’s earbuds 

Nothing plans to integrate ChatGPT with its smartphones and earbuds for faster access. Users with the latest OS and ChatGPT on their Nothing phones can pinch-to-speak with the earbuds. The rollout starts with Phone 2 on April 18th, extending to other models soon. The update allows querying ChatGPT using the earbuds, potentially as the default assistant. 

Nothing also aims to enhance smartphone user experience with ChatGPT integration into OS features like screenshot sharing. Unlike some AI gadgets, Nothing focuses on integrating existing devices without forcing changes on users lacking ChatGPT.

THE VERGE

Japan to fund KDDI, four others to build AI supercomputer

The Japanese government is allocating $470 million to support five companies, including KDDI and Sakura Internet, in developing artificial intelligence (AI) technology. Concerns over reliance on U.S. tech have prompted this move. 

The Ministry of Economy, Trade and Industry will fund up to half of each company's investment, with Sakura Internet receiving the most substantial aid at 50.1 billion yen, followed by KDDI at 10.2 billion yen. This initiative reflects Japan's strategic effort to bolster its AI capabilities and reduce dependency on foreign technology.

NIKKEI

Google is combining its Android and hardware teams — and it’s all about AI

Google is undergoing significant internal changes to accelerate its AI efforts. CEO Sundar Pichai announced the creation of a new team, "Platforms and Devices," led by Rick Osterloh. This team will oversee Pixel products, Android, Chrome, and more. The consolidation aims to enhance AI integration across all Google products. 

Osterloh emphasizes the importance of AI, citing examples like the Pixel camera's transformation through AI. The move signals Google's push to combine hardware and software expertise for faster innovation. While maintaining distinctions between Android and hardware teams, Google aims to leverage AI across its ecosystem, emphasizing its pivotal role in future products and services.

THE VERGE

Powerful new AI model accurately converts speech to text, even your company’s jargon

aiOla, an Israeli tech startup, introduces AdaKWS, an advanced AI model enhancing speech-to-text conversion accuracy. AdaKWS optimizes OpenAI's Whisper AI model, achieving up to 94.6% accuracy in keyword spotting across 16 languages, outperforming Whisper's 88.4%. It transcribes speech in near real-time, catering to various industries like healthcare and food production. 

Notably, AdaKWS is significantly faster than Whisper, aiding in automation of tasks such as filling out forms and conducting inspections. It saves time and resources, evident in its application for temperature recording in supermarkets, saving over 110,000 hours annually. AdaKWS's adaptable training process and multilingual support make it versatile for diverse enterprise needs, setting a new standard in speech AI integration for businesses.

VENTUREBEAT

Microsoft has a target to amass 1.8 million AI chips by the end of the year, internal document shows

Microsoft aims to accumulate 1.8 million AI chips by the end of 2024, as revealed by an internal document. This target suggests a threefold increase in GPU supply for the year, emphasizing Microsoft's commitment to enhancing generative AI capabilities. The company's collaboration with OpenAI in this domain underscores its position at the forefront of the AI boom. 

With an estimated $100 billion expenditure on GPUs and data centers from the current fiscal year to 2027, Microsoft's investment reflects the substantial resources allocated to stay competitive in the AI space. While Microsoft explores designing its own AI chips to reduce dependence on Nvidia, some skepticism remains due to the rapid pace of technological advancement.

BUSINESS INSIDER

Qtum Foundation powers AI Web3 initiatives with 10,000 Nvidia GPUs

The Qtum Foundation is ramping up its blockchain AI ecosystem by deploying 10,000 Nvidia GPUs. Qtum's transition from cryptocurrency mining to AI operations showcases its adaptability and forward-thinking approach. These GPUs will power Quantum AI projects, including a conversational chatbot called Qtum Solstice and a text-to-image generation model named Qtum Qurator. 

Qtum envisions a three-stage rollout plan for its AI endeavors, starting with chatbots and image generators and eventually expanding to include more sophisticated AI capabilities integrated into the Qtum blockchain. The foundation's commitment to open-source AI technology aligns with its ethos of decentralization and community-driven development. 

VENTUREBEAT

🧠RESEARCH

Dynamic Typography: Bringing Words to Life

Explores text animation as a means to enhance communication by adding motion to words. The authors propose an automated scheme, "Dynamic Typography", which deforms letters to convey meaning and adds vibrant movements based on user input. Their approach, using vector graphics and optimization, yields coherent and readable text animations.

Reka Core, Flash, and Edge: A Series of Powerful Multimodal Language Models

A trio of potent multimodal language models developed by Reka. These models adeptly handle text, images, video, and audio inputs, showcasing remarkable performance across various benchmarks. Reka Core, the largest model, approaches the forefront of AI models, demonstrating competitive performance and outperforming many larger counterparts.

Toward Self-Improvement of LLMs via Imagination, Searching, and Criticizing

A method for enhancing Large Language Models (LLMs) through self-improvement. By integrating Monte Carlo Tree Search (MCTS), AlphaLLM establishes a loop for self-refinement, addressing challenges like data scarcity and the subjective nature of feedback. Experimental results in mathematical reasoning tasks show significant performance improvements without additional annotations, suggesting potential for LLM self-improvement.

MeshLRM: Large Reconstruction Model for High-Quality Mesh

A novel approach for reconstructing high-quality meshes from just four input images in less than a second. Unlike previous methods focusing on NeRF-based reconstruction, MeshLRM integrates differentiable mesh extraction and rendering within the LRM framework, enabling end-to-end mesh reconstruction. By simplifying complex designs and employing a sequential training strategy, MeshLRM achieves state-of-the-art results with faster convergence and less compute, facilitating various downstream applications.

BLINK: Multimodal Large Language Models Can See but Not Perceive

A benchmark for multimodal language models (LLMs) focusing on core visual perception abilities. Blink tasks, solvable by humans almost instantly, challenge LLMs with tasks like relative depth estimation and visual correspondence. Despite human accuracy averaging 95.70%, current LLMs struggle, with even top-performing models achieving only around 51.26% accuracy, indicating a significant gap in visual perception.

🛠️TOP TOOLS

Demodesk - AI superpowers for your sales meetings

IKI AI - Smart library & Knowledge Assistant for professionals and teams

Ideatum - Convert Design Ideas from Words to Mood boards with Color Schemes and Font Pairings

Veles- Copilot for Enterprise Sales

AI Researcher -  AI agent that utilizes Claude 3 and SERPAPI to perform comprehensive research on a given topic

📲SOCIAL MEDIA

🗞️MORE NEWS

Facebook’s AI Told Parents Group It Has a Gifted, Disabled Child

Meta's AI mistakenly claimed to be a parent in a NYC parents' Facebook group, stating it has a gifted, disabled child in a specific public school. The AI later retracted, admitting it lacks personal experiences. The incident sparked discussion on AI ethics. Princeton professor Aleksandra Korolova highlighted the incident. 404 MEDIA

LinkedIn testing Premium Company Page subscription with AI-assisted content creation

LinkedIn is testing a Premium Company Page subscription, offering AI-driven content creation and follower growth tools for a monthly fee, potentially starting at $99. This move reflects LinkedIn's strategy to diversify revenue streams and provide a safer, more valuable space for professionals amidst social platform turmoil. The subscription includes familiar LinkedIn features like visitor insights, call-to-action buttons, and AI writing assistance, aligning with the platform's broader premium offerings. TECHCRUNCH

Honeywell exec reveals plan to deliver $100 million in value with generative AI: “Just getting started”

Honeywell plans to generate over $100 million in value using generative AI, according to Sheila Jordan, the company's chief digital technology officer. With 24 active projects, including GitHub integration and large language models for operations, Honeywell sees generative AI as a game-changer, emphasizing strategy, organization, and security. VENTUREBEAT

Stability AI lays off roughly 10 percent of its workforce

Stability AI, the UK-based company behind Stable Diffusion, laid off roughly 10% of its workforce, about 20 employees, just after expanding access to its flagship model. Interim CEOs cite cost reduction and strengthening investor support as reasons for the layoffs amidst recent upheaval, including the departure of the founding CEO. The company faces legal challenges regarding alleged copyright infringement in training its AI models. THE VERGE

Google Maps will use AI to help you find out-of-the-way EV chargers

Google Maps is enhancing its features for electric vehicle (EV) owners using AI. It will provide detailed directions to EV chargers based on customer reviews and encourage users to submit feedback for future improvements. Real-time plug availability and charging speeds will also be displayed, aiding in route planning for EV drivers. THE VERGE

Kartoon Studios Unveils ‘Gadget A.I.’

Kartoon Studios unveils "Gadget A.I.," an AI toolkit aiming to revolutionize animation production. Spearheaded by Todd Steinman, the initiative seeks to cut animation production costs by 75% and enhance speed-to-market efficiencies. Utilizing NVIDIA Omniverse, the toolkit integrates various AI tools to optimize production services while creating job growth opportunities. YAHOO!

Longtime Microsoft exec Gurdeep Singh Pall joins Qualtrics as president of AI strategy

Gurdeep Singh Pall, a longtime Microsoft executive, is joining Qualtrics as its first president of artificial intelligence strategy. He aims to leverage AI to enhance customer experiences by adapting products and services based on individual preferences. Pall's move reflects the growing importance of AI in transforming human-machine interaction. GEEKWIRE

Meet The New AI-Robot Billionaire

Brett Adcock, the founder and CEO of Figure, has ambitious plans for his company's humanoid robots, aiming to address job shortages in various industries. With substantial funding and collaborations with major players like BMW and OpenAI, Figure is poised to revolutionize human-machine interaction. Adcock's journey from farming to finance to robotics exemplifies his entrepreneurial spirit and determination to innovate. FORBES

Enabling your AI transformation journey with Microsoft Viva 

Microsoft Viva is evolving to support organizations in their AI transformation journey, providing tools like the Copilot Dashboard and Copilot Academy to facilitate enterprise-wide adoption of AI technologies. The Copilot Dashboard offers privacy-protected insights into Copilot for Microsoft 365 usage, enabling leaders to understand its impact across various organizational contexts. Additionally, Copilot Academy, available in Viva Learning, centralizes learning experiences to enhance employees' Copilot skills. With these advancements, Microsoft aims to empower organizations to harness the full potential of AI in the workplace. MICROSOFT

5 Ways To Make Money Online With AI In 2024

AI offers lucrative opportunities for earning money online, whether by enhancing existing ventures or birthing new ones. Leveraging AI, one can create chatbots for businesses, accelerate course creation, develop AI products, offer AI consulting services, and optimize graphic design with AI-powered tools like Canva. As AI evolves, so do the avenues for financial growth. FORBES