AI News

Follow the latest AI and automation news

14 Jun 202629 Dhuʻl-Hijjah 1447 AH

Glass Core Substrates Are Booming: How Far Are They From True Commercialization in Chip Packaging?

Glass core substrates are emerging as a hot topic in semiconductor packaging, driven by the demand for higher performance in AI chips. Major companies like Intel and Samsung are heavily investing in R&D, with BOE Technology Group planning to produce glass substrates for advanced chip packaging. These substrates offer superior dimensional stability and better electrical performance, making them attractive for large-scale AI chip packaging. However, challenges remain, including managing the brittle nature of glass and developing reliable processes for through-glass via formation.

Agents

14 Jun 202629 Dhuʻl-Hijjah 1447 AH

Huawei Unveils HarmonyOS 7: AI Agents Reshape the OS as Xiaoyi Becomes the System's Intelligent Core

Huawei officially launched HarmonyOS 7 at the HDC 2026 developer conference, marking a significant shift in its operating system. The Xiaoyi assistant has evolved into a system-level intelligence agent, controlling over 2,100 capabilities and accessing more than 200 data categories. The new 'intent-as-service' model simplifies app navigation into single natural language commands. Huawei also introduced openPangu 2.0, featuring 505 billion parameters, with support for 30-billion-parameter models on Kirin chips by autumn 2026.

Governance

13 Jun 202628 Dhuʻl-Hijjah 1447 AH

Anthropic Raises Security Concerns with Trump Administration

A source informed Reuters that Amazon CEO Andy Jassy was among tech leaders who raised security concerns regarding Anthropic's advanced AI models. The Trump administration issued a directive preventing the company from allowing foreign citizens to use its Fable 5 and Mythos 5 models. In response, Anthropic announced it would globally disable access to these models, citing U.S. government concerns about potential security breaches. This decision surprised some experts, as it includes allied nations like Canada and the UK.

💡

Amazon CEO reportedly raised Anthropic model concerns before government crackdown

What security concerns might impact AI development moving forward?

Arab AI Scene

G42 Joins Forces with Credo AI To Advance Responsible AI Adoption Across Global South

13 Jun 202628 Dhuʻl-Hijjah 1447 AH

Arab AI Scene

13 Jun 202628 Dhuʻl-Hijjah 1447 AH

G42 Joins Forces with Credo AI To Advance Responsible AI Adoption Across Global South

G42 has signed a memorandum of understanding with Credo AI to promote responsible AI in the Global South. The agreement was finalized during the AI Impact Summit in New Delhi, India, in 2026, where global leaders, policymakers, and innovators convened to discuss transforming AI advancements into tangible outcomes. The partnership emphasizes the need to transition from principle-based commitments to operational practices, particularly in regions with diverse regulatory environments.

Models

New AI model called "Count Anything" does exactly what it says, and that's harder than it sounds

13 Jun 202628 Dhuʻl-Hijjah 1447 AH

Models

13 Jun 202628 Dhuʻl-Hijjah 1447 AH

New AI model called "Count Anything" does exactly what it says, and that's harder than it sounds

The AI model "Count Anything" has been launched as the first of its kind to count objects in any image, whether from crowds or cellular samples under a microscope. In comparative testing, the new model reduced the error rate by half compared to previous systems, demonstrating significant progress in counting accuracy. However, it still struggles with extremely dense objects and ambiguous terms. "Count Anything" uses simple text as input, making it user-friendly for various applications.

Business

AI Reshapes Global Stock and Bond Markets

13 Jun 202628 Dhuʻl-Hijjah 1447 AH

Business

13 Jun 202628 Dhuʻl-Hijjah 1447 AH

AI Reshapes Global Stock and Bond Markets

Investors are focused on identifying winners and losers in the AI era, with significant impacts on investment portfolios. The recent IPO of SpaceX has drawn exceptional interest, with expectations that companies like OpenAI could follow with valuations reaching a trillion dollars. Tech giants are also injecting new shares into the market, reversing a two-decade trend of declining listed companies in U.S. exchanges. The influence of AI extends to the bond market, as firms like Amazon and Alphabet have raised their debt levels to around $300 billion to fund AI infrastructure investments.

Models

Microsoft's SkillOpt boosts GPT-5.5 by using nothing but a trained Markdown file

13 Jun 202628 Dhuʻl-Hijjah 1447 AH

Models

13 Jun 202628 Dhuʻl-Hijjah 1447 AH

Microsoft's SkillOpt boosts GPT-5.5 by using nothing but a trained Markdown file

Microsoft, in collaboration with three Chinese universities, has developed SkillOpt, a method aimed at enhancing educational documentation for intelligent agents. This approach leverages traditional model training principles, utilizing a simple Markdown file to boost GPT-5.5 performance by 23 points on procedural tasks. SkillOpt excels in transferring improvements across various models and environments, such as Codex and Claude Code.

Models

Two Turing Award Winners Convene at Beijing Zhiyuan Conference, Confronting the Theoretical Black Hole Behind AGI

13 Jun 202628 Dhuʻl-Hijjah 1447 AH

Models

13 Jun 202628 Dhuʻl-Hijjah 1447 AH

Two Turing Award Winners Convene at Beijing Zhiyuan Conference, Confronting the Theoretical Black Hole Behind AGI

The 8th Beijing Zhiyuan Conference began on June 12, 2026, focusing on cutting-edge AI research. The opening featured Turing Award winners Whitfield Diffie and Andrew Barto, who discussed critical aspects of AI development. Diffie highlighted the paradox of defining safety in AI agents, while Barto emphasized the challenges of reward functions in reinforcement learning, indicating a gap in theoretical foundations compared to industry enthusiasm.

Models

Claude Fable 5 outpaces GPT-5.5 by 13 points on FrontierMath's toughest problems

13 Jun 202628 Dhuʻl-Hijjah 1447 AH

Models

13 Jun 202628 Dhuʻl-Hijjah 1447 AH

Claude Fable 5 outpaces GPT-5.5 by 13 points on FrontierMath's toughest problems

Anthropic's Claude Fable 5 achieved an impressive 88% accuracy on the toughest FrontierMath tier, a significant leap from Opus 4.5, which was below 10% in early 2026. In comparison, OpenAI's GPT-5.5 reached about 75% on the same tier. This rapid improvement in AI math accuracy indicates an accelerating pace of advancements in the field. This progress poses a considerable challenge to competitors, suggesting that such results will influence future model development strategies, emphasizing the need for enhanced performance in complex mathematical tasks.