AI News

Follow the latest AI and automation news

25 May 20269 Dhuʻl-Hijjah 1447 AH

Machine Learning and Data Analytics Symposium | QCRI

The eighth edition of the MLDAS 2023 conference was held from October 23 to 24, 2023, organized by the Qatar Computing Research Institute (QCRI) in collaboration with Boeing. The event brought together researchers and experts to discuss AI applications in aviation, education, and sports analytics. Specialists from companies such as Amazon and Boeing participated, along with academics from prestigious institutions like Carnegie Mellon University and KAUST. The conference featured 15 technical presentations on ongoing research projects, facilitating knowledge exchange among participants in this multidisciplinary field.

Infrastructure

25 May 20269 Dhuʻl-Hijjah 1447 AH

Core42 Establishes European Headquarters in Dublin to Support Growing Demand for AI Infrastructure

Core42, a G42 company, has established its European headquarters in Dublin, Ireland, announced during Investopia. This new hub aims to enhance Core42's ability to serve European enterprises and governments seeking secure, high-performance AI infrastructure. Founded in 2023, Core42 focuses on sovereign cloud and advanced compute platforms, partnering with leaders like Microsoft and NVIDIA. The company has expanded its European presence through initiatives like a national AI infrastructure in France and the largest AI compute cluster in Italy.

Business

25 May 20269 Dhuʻl-Hijjah 1447 AH

G42 Sets a New Standard in Precision AI Weather Forecasting Leveraging NVIDIA Earth-2 | Invent a Better Everyday | Abu Dhabi, UAE | G42

G42 and NVIDIA have announced a groundbreaking AI-powered weather forecasting system that achieves 200-meter resolution. Developed at the Earth-2 Climate Tech Lab in Abu Dhabi, this system offers faster and more cost-effective simulations, enabling predictions of extreme weather events. The system successfully conducted a comprehensive fog simulation over the UAE, enhancing forecast accuracy and positively impacting sectors like transportation and public safety. G42 aims to improve risk management and efficiency in response to severe weather conditions through this innovation.

💡

ByteDance Open-Sources Lance: A Native Multimodal Model That Runs Locally on 40GB VRAM

What impact will Lance have on local application development?

Models

Alibaba's Qwen 3.7 Max Surprises Overseas Developers with 35-Hour Autonomous Task Run

25 May 20269 Dhuʻl-Hijjah 1447 AH

Models

25 May 20269 Dhuʻl-Hijjah 1447 AH

Alibaba's Qwen 3.7 Max Surprises Overseas Developers with 35-Hour Autonomous Task Run

Alibaba has launched its new model, Qwen 3.7 Max, which has excited international developers by demonstrating an exceptional ability to perform tasks autonomously for 35 continuous hours with 1,158 tool invocations without any performance degradation. The model gained attention after a user on platform X noted its capability to handle long-term tasks independently. This feature positions Qwen 3.7 Max as a leader in managing complex workflows. Its release strengthens Alibaba's position in the large language model race, as developers begin exploring its capabilities in areas such as programming and agent coordination.

Business

How Xiaomi’s push into AI, chips and EVs is future-proofing its hardware empire

25 May 20269 Dhuʻl-Hijjah 1447 AH

Business

25 May 20269 Dhuʻl-Hijjah 1447 AH

How Xiaomi’s push into AI, chips and EVs is future-proofing its hardware empire

Xiaomi is heavily investing in artificial intelligence to ensure it remains competitive in the smartphone and electric vehicle markets. The company recently introduced its new model, MiMo-V2.5-Pro, which was ranked as the top open-source model for agentic capabilities by Artificial Analysis. This move is part of Xiaomi's broader strategy to integrate AI across its product lines. By focusing on open-source models, the company aims to enhance its competitive edge in the rapidly evolving AI landscape.

Business

OpenAI, Grupo Folha and Grupo UOL announce strategic content partnership

25 May 20269 Dhuʻl-Hijjah 1447 AH

Business

25 May 20269 Dhuʻl-Hijjah 1447 AH

OpenAI, Grupo Folha and Grupo UOL announce strategic content partnership

OpenAI has partnered with Grupo Folha and Grupo UOL to deliver trusted Brazilian journalism through ChatGPT. This collaboration aims to enhance news access with a focus on attribution and transparency. The initiative allows users to receive accurate information from reputable news sources, thereby boosting content credibility. It also aims to improve user experience by providing reliable and diverse news coverage.

Business

DeepSeek's Price Cut for V4-Pro Impacts AI Model Market

24 May 20268 Dhuʻl-Hijjah 1447 AH

Business

24 May 20268 Dhuʻl-Hijjah 1447 AH

DeepSeek's Price Cut for V4-Pro Impacts AI Model Market

DeepSeek's decision to reduce the prices of its V4-Pro model has garnered simultaneous coverage from major tech outlets, indicating that this development is significant within the AI model market. This price cut may influence pricing benchmarks for large models, prompting competitors to reassess their pricing strategies. It could also empower startups to enter the market, leveraging lower prices to offer innovative solutions. The anticipated increase in market competition may enhance the quality of available models and foster new avenues for innovation in AI, making it more accessible to users.

Models

ByteDance study finds that asking LMMs questions beats making it transcribe text for long document training

24 May 20268 Dhuʻl-Hijjah 1447 AH

Models

24 May 20268 Dhuʻl-Hijjah 1447 AH

ByteDance study finds that asking LMMs questions beats making it transcribe text for long document training

The ByteDance Seed model, featuring 7 billion parameters, has demonstrated superior reliability in answering questions on long, image-heavy documents compared to larger models. It was tested on documents four times longer than its training data, showcasing its effectiveness in handling complex information. Instead of transcribing text, the model learns by answering questions and identifying relevant passages independently. This approach marks a shift in how large language models are trained, emphasizing comprehension over rote repetition.

Models

Deepmind's Hassabis sees humanity "in the foothills of the singularity" while LeCun says current AI isn't intelligent

24 May 20268 Dhuʻl-Hijjah 1447 AH

Models

24 May 20268 Dhuʻl-Hijjah 1447 AH

Deepmind's Hassabis sees humanity "in the foothills of the singularity" while LeCun says current AI isn't intelligent

Yan LeCun believes that current AI systems are not truly intelligent, lacking the ability to learn from experience. In contrast, Demis Hassabis sees humanity at the threshold of a transformative shift toward general artificial intelligence. Oriol Vinyals adds that models today would have seemed like general AI seven years ago, yet they still need to achieve real breakthroughs. The differing opinions among these AI pioneers address the challenges facing current systems. LeCun emphasizes the absence of genuine intelligence, while Hassabis discusses the potential future advancements.