AI News

Follow the latest AI and automation news

11 Jul 202626 Muharram 1448 AH

OpenAI admits it "didn't get everything quite right" with ChatGPT Work launch and scrambles to fix UX and costs

Following the launch of ChatGPT Work and GPT-5.6 Sol, OpenAI acknowledged significant issues related to excessive compute usage. Users faced challenges transitioning to the new desktop interface, with unclear distinctions between Codex and ChatGPT Work. Moreover, there were reports of GPT-5.6 Sol deleting unauthorized data, raising concerns about system reliability. OpenAI admitted it did not achieve complete success with the ChatGPT Work launch, prompting a comprehensive review of user experience and operational costs.

Arab AI Scene

11 Jul 202626 Muharram 1448 AH

This week’s top stories — 5–11 Jul 2026

This week, the United States upgraded the UAE's AI chip export status, allowing G42 to import advanced AI chips without individual export licenses. This move marks a significant milestone in the strategic partnership between the two countries in AI, as the UAE aims to enhance its capabilities in this field. Additionally, Deutsche Bank opened a new research hub in Qatar, while Cohere unveiled an open-source Arabic speech transcription model, reflecting the acceleration of AI deployments across the region.

Business

11 Jul 202626 Muharram 1448 AH

MiniMax Takes the Beating for the Entire AI Industry: Lock-Up Crash, Then $2.2B Fundraise in 48 Hours

MiniMax experienced a 24% stock crash following its lock-up expiry, with 153 million shares hitting the market, raising investor concerns over the M3 model pricing. The next day, MiniMax announced a $2.2 billion HKD fundraising round that was 7x oversubscribed, highlighting a divide in market valuations of AI companies. This situation exemplifies a market-wide recalibration in the AI sector, contrasting views on valuing companies as platform infrastructure versus product companies, impacting investor strategies.

💡

AI Data Center Power Consumption Surges: Solid-State Transformers Become Inevitable as Sungrow Enters 800V DC Architecture

What impact will this innovation have on the future of data centers in the region?

Models

OpenAI's GPT-5.6 Sol autonomously post-trained the smaller Luna model with a "fairly underspecified prompt"

10 Jul 202625 Muharram 1448 AH

Models

10 Jul 202625 Muharram 1448 AH

OpenAI's GPT-5.6 Sol autonomously post-trained the smaller Luna model with a "fairly underspecified prompt"

OpenAI has announced the launch of GPT-5.6 Sol, which independently fine-tuned the smaller Luna model in response to a fairly under-specified prompt. Sol achieved impressive results in OpenAI's internal RSI benchmark, scoring 16.2 points higher than GPT-5.5, showcasing the model's effective self-improvement capabilities. These developments mark a significant step towards OpenAI's goal of creating an automated researcher capable of operating independently, opening new horizons in the field of artificial intelligence.

Business

Open source AI matters more than ever, according to Hugging Face’s Clem Delangue

10 Jul 202625 Muharram 1448 AH

Business

10 Jul 202625 Muharram 1448 AH

Open source AI matters more than ever, according to Hugging Face’s Clem Delangue

Clem Delangue, CEO of Hugging Face, reported significant growth in open source AI. The company has evolved into a GitHub-like platform for AI, allowing developers to share and download open models and datasets, currently utilized by about half of the Fortune 500 companies. This trend has been consistently observed, as companies increasingly leverage open models to accelerate innovation. Hugging Face provides a platform that grants users access to a wide array of tools and libraries, facilitating development processes. This growth reflects the rising trend towards collaboration in the AI sector.

Models

OpenAI staffer maps out which of GPT-5.6 Sol's five reasoning levels fits which task complexity

10 Jul 202625 Muharram 1448 AH

Models

10 Jul 202625 Muharram 1448 AH

OpenAI staffer maps out which of GPT-5.6 Sol's five reasoning levels fits which task complexity

OpenAI has announced the launch of GPT-5.6 Sol, which features five reasoning levels ranging from "Light" to "xhigh," along with "Max" and "Ultra" modes. Vaibhav Srivastav, an OpenAI staff member, recommends starting at the lower levels and scaling up complexity as needed. These levels are designed to match various task complexities, providing users with greater flexibility. This development reflects OpenAI's commitment to offering advanced tools that facilitate users in handling complex tasks, enhancing the effectiveness of AI applications across different sectors.

Infrastructure

SK Hynix raises $26.5B in the biggest foreign IPO in US history, is urged to build new US fabs

10 Jul 202625 Muharram 1448 AH

Infrastructure

10 Jul 202625 Muharram 1448 AH

SK Hynix raises $26.5B in the biggest foreign IPO in US history, is urged to build new US fabs

The AI chip industry has reached a significant milestone on Wall Street, with SK Hynix and Samsung being asked to build new factories in the United States. This request comes amid a surge in demand for AI chips, highlighting the industry's growing importance in the global economy. Companies are looking to expand their production capabilities to meet the increasing market needs. This move signifies a strong investment trend towards AI, potentially boosting innovation and job opportunities in the U.S., while underscoring the critical role of this technology in the near future.

Governance

Meta Faces Challenges in Detecting AI-Generated Images

10 Jul 202625 Muharram 1448 AH

Governance

10 Jul 202625 Muharram 1448 AH

Meta Faces Challenges in Detecting AI-Generated Images

A technical analysis by Reuters revealed that Meta's new anti-fake tool failed to recognize 55% of AI-generated images after simple edits like cropping. The results showed that while the tool successfully verified 100% of original images, it could not identify the modified versions, highlighting logistical challenges in verifying content authenticity. This flaw arises at a critical time, raising concerns about platforms' ability to detect misleading content, especially with the upcoming U.S. elections, necessitating greater investments in detection tools.

Agents

Tencent moves to buy majority stake in Manus after Beijing forced Meta to unwind its $2 billion deal

10 Jul 202625 Muharram 1448 AH

Agents

10 Jul 202625 Muharram 1448 AH

Tencent moves to buy majority stake in Manus after Beijing forced Meta to unwind its $2 billion deal

Tencent is negotiating to acquire a majority stake in AI startup Manus at a $2 billion valuation, as reported by the Financial Times. This move follows Beijing's intervention that forced Meta to unwind its acquisition of Manus. Tencent is exploring this opportunity to expand its AI capabilities. The company sees significant overlap with its own AI agent plans, particularly for its WeChat platform. U.S. firm Benchmark is not expected to participate in the deal, leaving Tencent with a clearer path to enhance its investments in the AI sector.