AI News

Follow the latest AI and automation news

14 Aug 20262 Rabiʻ I 1448 AH

Chain-of-Thought Prompting: Techniques and When To Use Them

Large language models often struggle to provide direct answers to complex queries. Chain-of-thought (CoT) prompting addresses this issue by enhancing logical transparency, aiding teams in debugging incorrect outputs and verifying the model's conclusions. CoT prompting elicits a sequence of intermediate reasoning steps that the model uses to arrive at its response. A study published in Frontiers Media SA revealed that CoT had the lowest hallucination rate at 18.1%, compared to 34.5% for other prompting methods, demonstrating its effectiveness in improving answer accuracy.

Infrastructure

14 Aug 20262 Rabiʻ I 1448 AH

Universitas Gadjah Mada, Indosat and NVIDIA Open Indonesia’s First University AI Center to Develop Local AI Talent

Indonesia has launched the AI Technology Center at Gadjah Mada University in collaboration with Indosat and NVIDIA. This center aims to develop AI solutions that address the country's urgent national priorities. The center is part of Indonesia's AI Center of Excellence initiative, bringing together government, industry, and academia. It reflects a long-term vision for Indonesia to lead in developing AI innovations and contributing to the global AI landscape.

Models

14 Aug 20262 Rabiʻ I 1448 AH

Alibaba's Qwen team releases Qwen 3.8 models with open weights under the Apache 2.0 license

Alibaba's AI team Qwen has announced the release of new open model weights for Qwen 3.8 under the Apache 2.0 license. The new model, featuring 27 billion parameters, is designed to outperform its predecessor, Qwen 3.7 Plus, in coding and office tasks, processing up to 262,000 tokens of context. This release targets developers building local AI-based applications, enhancing the model's usability across various environments.

💡

OpenAI's Computer History turns your clicks and keystrokes into a searchable ChatGPT memory timeline

What are your thoughts on the privacy implications of this computer history feature?

Agents

Study contradicts Anthropic and OpenAI claims that autonomous AI research is within reach

14 Aug 20262 Rabiʻ I 1448 AH

Agents

14 Aug 20262 Rabiʻ I 1448 AH

Study contradicts Anthropic and OpenAI claims that autonomous AI research is within reach

A new study reveals that AI models like Claude Opus 4.8 and GPT-5.6 Sol, given six days and $3,000 in API credits, failed to produce acceptable research papers. The original authors of unpublished NeurIPS papers rated the results as "Reject," indicating these models' inability to make sound research decisions. Conducted in collaboration with Princeton University and the UK AI Security Institute, the study found that frontier models can manage the entire research engineering process but lack research judgment and creative problem-solving skills. These findings contradict claims by Anthropic and OpenAI that autonomous AI research is within reach.

Models

Does Mark Zuckerberg really believe AI is ‘for everyone’?

14 Aug 20262 Rabiʻ I 1448 AH

Models

14 Aug 20262 Rabiʻ I 1448 AH

Does Mark Zuckerberg really believe AI is ‘for everyone’?

Meta has launched Glimmer, an open-weight AI model available for anyone to download and run on their hardware. This release contrasts with the more powerful Muse Spark, which remains locked behind proprietary APIs. In a letter from Mark Zuckerberg, he emphasized that AI should be accessible to everyone rather than controlled by a few labs. This approach reflects Meta's vision of promoting broader access to AI technologies.

Agents

Building AI Agent Observability for Production Workflows

14 Aug 20262 Rabiʻ I 1448 AH

Agents

14 Aug 20262 Rabiʻ I 1448 AH

Building AI Agent Observability for Production Workflows

AI agents are evolving to handle complex tasks, but debugging them is becoming more challenging. Understanding where and why a failure occurred is crucial, making observability a key aspect of AI workflows. AI agent observability provides a comprehensive view of an agent's execution, including model calls and tool usage. This visibility allows engineering teams to investigate failures and enhance system reliability over time. Observability tools typically consist of three components: traces, metrics, and logs. These tools help teams understand agent behavior, making it easier to identify failure points and analyze performance effectively.

Models

GPT-5.6 Sol goes 14x faster as OpenAI launches Ultrafast mode powered by Cerebras

14 Aug 20262 Rabiʻ I 1448 AH

Models

14 Aug 20262 Rabiʻ I 1448 AH

GPT-5.6 Sol goes 14x faster as OpenAI launches Ultrafast mode powered by Cerebras

OpenAI has announced the launch of a new inference mode called 'Ultrafast,' which delivers speeds of up to 750 tokens per second. This mode is powered by Cerebras technology, part of a $10 billion partnership. Ultrafast is part of a three-tier pricing structure that includes 'Standard' and 'Fast' modes, allowing inference speed to be treated as a standalone product. GPT-5.6 Sol is now 14 times faster than previous versions.

Business

Hyperscalers might regret embracing natural gas if new forecast proves correct

14 Aug 20262 Rabiʻ I 1448 AH

Business

14 Aug 20262 Rabiʻ I 1448 AH

Hyperscalers might regret embracing natural gas if new forecast proves correct

Natural gas prices are expected to triple in some areas of the U.S., significantly impacting the operational costs of AI data centers. These data centers are crucial for running advanced AI models, relying on substantial energy consumption. The rising prices could lead to massive energy bills, threatening the profitability of major companies. If these trends continue, companies may need to reassess their energy consumption strategies, potentially affecting their investments in AI technology. This could result in a slowdown in innovation within the sector.

Models

Meta’s ‘open’ AI, and a $250M deal gone very wrong

14 Aug 20262 Rabiʻ I 1448 AH

Models

14 Aug 20262 Rabiʻ I 1448 AH

Meta’s ‘open’ AI, and a $250M deal gone very wrong

Meta has released Glimmer this week, an open-weight AI model that anyone can download and run on their own hardware. This release coincides with a letter from Mark Zuckerberg emphasizing that AI should be 'for everyone' rather than controlled by a few labs, highlighting the growing discourse on open access to AI technologies. The launch of Glimmer signifies a shift towards greater transparency in the AI field, potentially fostering innovation and allowing developers and users to leverage this technology more effectively.