AI News

Follow the latest AI and automation news

13 Aug 20261 Rabiʻ I 1448 AH

Does Google even want to win at AI?

Last week, Google announced a significant restructuring of its AI division, Google DeepMind, with Jeff Dean, the chief scientist, leaving to start a new lab. Meanwhile, co-founder and CEO Demis Hassabis will take on a new role focusing on long-term research. These changes come at a time when Google is facing challenges in maintaining competitiveness in the AI landscape.

Models

13 Aug 20261 Rabiʻ I 1448 AH

OpenAI Launches ChatGPT Beta for Linux Systems

OpenAI has announced the launch of a beta version of the ChatGPT application for computers running Linux, expanding the app's availability. The new application allows Linux users to access ChatGPT, ChatGPT Work, and Codex through a single interface, facilitating usage and enhancing productivity. This expansion reflects OpenAI's commitment to meeting the needs of the Linux user community, potentially increasing the adoption of AI in various work environments.

Business

13 Aug 20261 Rabiʻ I 1448 AH

Alibaba adds commercial restrictions to open-weight Qwen3.8-Max AI model

Alibaba Group has announced that the core files of its new model Qwen3.8-Max are available for free download. However, companies with revenues exceeding $50 million over any consecutive 12-month period must obtain a separate commercial license to use this model. This move reflects Alibaba's efforts to regulate the use of its AI models, impacting businesses that provide AI as a service or AI work assistant solutions.

💡

Microsoft is combining its Copilot apps ahead of a ‘super app’

How will the Copilot app integration impact user experience?

Business

Deep Seek Launches V4 Pro Model to Boost Competitive Edge

13 Aug 20261 Rabiʻ I 1448 AH

Business

13 Aug 20261 Rabiʻ I 1448 AH

Deep Seek Launches V4 Pro Model to Boost Competitive Edge

Deep Seek, a Chinese AI startup, officially launched its V4 Pro model today, aiming to regain its competitive edge. The new model offers significant enhancements in agent capabilities and is available through APIs and applications, alongside an increase in API pricing. This launch follows the surprising success of the cheaper V4 Flash model over the beta version of V4 Pro in independent tests, indicating rapid technological advancement. The company faces challenges in converting its popularity into a sustainable business amid rising competition from firms like Moonshot AI and Alibaba.

Arab AI Scene

Presight Q2 revenue jumps 36% to AED 713 million

13 Aug 20261 Rabiʻ I 1448 AH

Arab AI Scene

13 Aug 20261 Rabiʻ I 1448 AH

Presight Q2 revenue jumps 36% to AED 713 million

Presight AI Holding PLC, listed on the Abu Dhabi Securities Exchange, reported a 36.1% increase in Q2 2026 revenue, reaching AED 713.2 million. The company's EBITDA rose by 37.4% to AED 143.7 million, with a profit after tax of AED 116.8 million, reflecting a 30.2% increase. The second quarter saw a record domestic order intake of AED 2.5 billion, maintaining a 100% renewal rate across major national programs. The order backlog reached AED 4.9 billion as of June 30, 2026, up 44.3% since the start of the year.

Infrastructure

AI demand drives triple-digit profit growth for Chinese chip foundries SMIC, Hua Hong

13 Aug 20261 Rabiʻ I 1448 AH

Infrastructure

13 Aug 20261 Rabiʻ I 1448 AH

AI demand drives triple-digit profit growth for Chinese chip foundries SMIC, Hua Hong

China's leading chipmakers, SMIC and Hua Hong, reported significant profit increases in Q2, with net profits soaring by 261.7% and 385.9% to $479.2 million and $38.6 million, respectively. This profit surge is driven by a spike in demand for domestic AI chips that are not subject to US export controls. SMIC, the largest foundry in China, noted that the rising demand reflects a growing need for local technology amid ongoing trade tensions.

Agents

The builder’s guide to GPT‑5.6

13 Aug 20261 Rabiʻ I 1448 AH

Agents

13 Aug 20261 Rabiʻ I 1448 AH

The builder’s guide to GPT‑5.6

Startups are utilizing the GPT-5.6 model to develop faster and more cost-efficient AI agents, benefiting from new capabilities in the Responses API. These new API features enhance model selection, improving the agents' effectiveness in executing complex tasks. This development comes at a time when the demand for scalable AI solutions is increasing. This trend highlights the importance of innovation in the AI field, as companies seek to enhance efficiency and reduce costs, potentially leading to significant changes in both the Arab and global markets.

Models

Fable 5's slow adoption suggests corporate willingness to pay for frontier AI has hit a ceiling

13 Aug 20261 Rabiʻ I 1448 AH

Models

13 Aug 20261 Rabiʻ I 1448 AH

Fable 5's slow adoption suggests corporate willingness to pay for frontier AI has hit a ceiling

Anthropic's Fable 5 is regarded as the most powerful AI model available, yet U.S. companies are hesitant to purchase it. According to Ramp data, Fable 5 accounts for only six percent of total tokens sold by Anthropic. The steep price of the model suggests that corporate spending on AI may have reached a ceiling, particularly if performance improvements do not translate into tangible everyday value. The slow adoption of Fable 5 indicates a reluctance among companies to invest heavily in advanced AI, which could impact the future development of AI models in the market.

Models

Top AI lab researchers warned about automated AI research, and several of their predicted milestones have already fallen

13 Aug 20261 Rabiʻ I 1448 AH

Models

13 Aug 20261 Rabiʻ I 1448 AH

Top AI lab researchers warned about automated AI research, and several of their predicted milestones have already fallen

IAPS fellow Severin Field interviewed 25 researchers from OpenAI, Anthropic, Google Deepmind, Meta, and US universities about recursive self-improvement in AI. In a new blog post, he assessed the current state of automated AI research, noting that several milestones identified by these researchers have already been achieved. These milestones include significant advancements in AI, raising concerns about the implications of automated research.