AI News

Follow the latest AI and automation news

3 Jun 202618 Dhuʻl-Hijjah 1447 AH

Walmart’s AI workflows meet the realities of the balance sheet

Walmart has started limiting employee use of its AI assistant, Code Puppy, due to higher-than-expected demands on the underlying LLM. Initially, employees were encouraged to use Code Puppy freely, but the company is now assigning a fixed number of AI tokens per employee to control costs. This shift reflects a broader trend towards pay-per-use models for LLMs, moving away from unlimited access.

Infrastructure

3 Jun 202618 Dhuʻl-Hijjah 1447 AH

AI has a water problem — Google thinks it has a fix

Amid widespread backlash against data center construction in the U.S., Google is promoting its efforts to reduce environmental impact. The company outlined five commitments regarding water use, including a goal to replenish more water than it consumes at its data centers by 2030. Google plans to invest in local water infrastructure, identify alternative water sources for its facilities, and maintain transparency about its overall water usage. These initiatives come at a time of increasing concern over water consumption in the tech industry.

Agents

3 Jun 202618 Dhuʻl-Hijjah 1447 AH

Cisco unveils agentic platform to operate and defend critical IT infrastructure

Cisco unveiled its Cloud Control platform at Cisco Live, designed for managing critical IT infrastructure. This unified platform allows users to oversee networking, security, compute, and collaboration through a single interface, facilitating interaction between humans and AI agents. Cloud Control serves as a bridge for users to build applications using natural language, connecting to a broad ecosystem including AWS, ServiceNow, and Slack. This integration enhances its capability to interact with various systems, making it a versatile tool for enterprises.

💡

Google must let publishers opt out of AI Search features, rules UK

What implications do these rules have for publishers in the Arab world?

Infrastructure

Core42 advances US AI infrastructure strategy with New York deployment

3 Jun 202618 Dhuʻl-Hijjah 1447 AH

Infrastructure

3 Jun 202618 Dhuʻl-Hijjah 1447 AH

Core42 advances US AI infrastructure strategy with New York deployment

Core42, a G42 company, announced a significant expansion of its AI infrastructure at the Lake Mariner site in New York, increasing total capacity from 18MW to 60MW. This expansion adds 42MW of AI compute power, reinforcing Core42's heterogeneous AI infrastructure strategy. The existing Maximus cluster, based on AMD technology, ranks 20th on the TOP500 list, showcasing Core42's capability to deliver high-performance AI solutions globally.

Business

Censys powers SOC with AI, SIEM, SOAR, and threat intelligence ecosystems

3 Jun 202618 Dhuʻl-Hijjah 1447 AH

Business

3 Jun 202618 Dhuʻl-Hijjah 1447 AH

Censys powers SOC with AI, SIEM, SOAR, and threat intelligence ecosystems

Censys, a leading Internet intelligence platform, has announced new integrations across AI and security platforms. These integrations allow organizations to quickly understand the context and risks of IPs, domains, and services, enhancing security operations effectiveness. By embedding external intelligence into existing security tools, Censys facilitates faster investigations and threat responses. New releases include native integrations with Cisco Splunk SOAR, Microsoft Sentinel, and Google SecOps, expanding automation and visibility across modern security operations.

Models

Boundless World Model: How an Open-Source Chinese World Model Topped the Global Rankings

3 Jun 202618 Dhuʻl-Hijjah 1447 AH

Models

3 Jun 202618 Dhuʻl-Hijjah 1447 AH

Boundless World Model: How an Open-Source Chinese World Model Topped the Global Rankings

The BWM (Boundless World Model) from China has achieved a remarkable score of 64.54 in WorldArena Track-1, surpassing models from Google and NVIDIA. Developed by a team led by Prof. Shen Hengtao at Tongji University, BWM was fine-tuned from Alibaba's Wan2.2-TI2V-5B video generation model. BWM's performance is driven by three architectural innovations: DiT (Diffusion Transformer) for enhanced spatial reasoning, a Dynamic Memory Mechanism for temporal coherence, and First-Frame Guidance for precise control. These innovations significantly improve its ability to simulate physics realistically.

Infrastructure

Co-Packaged Optics: The Interconnect Backbone for AI Data Centers

3 Jun 202618 Dhuʻl-Hijjah 1447 AH

Infrastructure

3 Jun 202618 Dhuʻl-Hijjah 1447 AH

Co-Packaged Optics: The Interconnect Backbone for AI Data Centers

Data centers are experiencing a massive increase in data traffic due to large language models. However, the connections between graphics processing units (GPUs) have become a critical bottleneck that threatens AI growth. Co-Packaged Optics (CPO) technology offers a promising solution by integrating optical engines directly with ASIC units. The roadmap for electronics and photonics integration consists of three phases. The first phase involves plugable optical modules that consume 30 watts, while the second phase, Near-Package Optics (NPO), reduces consumption to 9 watts.

Models

[AINews] Microsoft Build: MAI-Thinking-1 and MAI Family models

3 Jun 202618 Dhuʻl-Hijjah 1447 AH

Models

3 Jun 202618 Dhuʻl-Hijjah 1447 AH

[AINews] Microsoft Build: MAI-Thinking-1 and MAI Family models

Today, Microsoft announced seven new MAI models at the Build conference, including MAI-Thinking-1 and MAI-Code-1-Flash. These models were developed from scratch, showcasing Microsoft's rapid advancements in AI over the past two years. The new models cover diverse areas such as reasoning, coding, image processing, and speech transcription. Microsoft also released a detailed technical report exceeding 100 pages, reflecting a high level of transparency in the development of these models.

Business

Deep Seek Secures $7.4 Billion in New Funding Round

3 Jun 202618 Dhuʻl-Hijjah 1447 AH

Business

3 Jun 202618 Dhuʻl-Hijjah 1447 AH

Deep Seek Secures $7.4 Billion in New Funding Round

Deep Seek, a Chinese AI startup, is set to raise approximately 50 billion yuan ($7.4 billion) in its first funding round, with major investors including Tencent and CATL. This funding could elevate the company's valuation to between 350 and 400 billion yuan ($52 to $59 billion) following the new investments. The company has emerged as a leading figure in China's AI landscape, with its models 'V3' and 'R1' receiving significant acclaim. This strong performance has solidified Deep Seek's position as a key competitor in the global AI race.