AI News

Follow the latest AI and automation news

17 Sept 20266 Rabiʻ II 1448 AH

UN turns to Google to make its global data ready for AI agents

A UNICEF test revealed that leading AI models struggle to accurately retrieve global development statistics. These findings raise concerns about the ability of these models to process and analyze large datasets, questioning their reliability in providing accurate information. Such results are significant for decision-makers, highlighting the need to enhance AI models to ensure their effective use in sustainable development areas.

Governance

17 Sept 20266 Rabiʻ II 1448 AH

The AI ‘Slowdown’ Is an Antitrust Mess

AI labs are facing regulatory challenges due to framing their efforts as a 'slowdown' rather than a comprehensive push for better security standards. This approach may lead to legal complications in the long run, as regulators focus on companies' compliance with security benchmarks. While these labs aim to enhance security, this framing could be misinterpreted by regulators, increasing pressure to meet higher expectations. The lack of clarity in messaging may exacerbate regulatory issues, placing companies in a difficult position.

Agents

17 Sept 20266 Rabiʻ II 1448 AH

Claude Code relaunches Projects to manage multiple AI agents in the cloud

The revamped projects in Claude Code allow users to run multiple agents under one roof, featuring shared memory, goals, and a library of files. Each project has "threads" executing different tasks in parallel, with a "coordinator" managing everything. Each thread operates as a separate Claude Code cloud session, working on its own branch of the repository. The coordinator keeps the work organized, and if any threads work on the same code, conflicts are resolved like any other pull request.

💡

The AI Slowdown Debate Crashed Salesforce’s Party

What are your thoughts on the societal impacts of AI?

Agents

Anthropic keeps pushing Claude Code toward autonomous coding with new parallel agent workflows

17 Sept 20266 Rabiʻ II 1448 AH

Agents

17 Sept 20266 Rabiʻ II 1448 AH

Anthropic keeps pushing Claude Code toward autonomous coding with new parallel agent workflows

Anthropic has announced a rebuild of its Claude Code project, introducing a coordinator that splits tasks across parallel cloud threads. These threads operate independently, opening pull requests and running tests, while sharing a common memory. The new updates enhance coding efficiency, with a beta version available to Pro and Max subscribers. This development is a significant step towards advancing autonomous coding capabilities, allowing developers to focus on more complex tasks.

Governance

LLMs respond differently to harmful prompts when AI watermarking is used

17 Sept 20266 Rabiʻ II 1448 AH

Governance

17 Sept 20266 Rabiʻ II 1448 AH

LLMs respond differently to harmful prompts when AI watermarking is used

In response to a new European Union law, AI platforms are implementing new schemes for watermarking the content they generate. Anthropic recently disclosed that its future Claude models will use SynthID-Text, an approach developed and released as open source by Google. This technique employs a secret key that subtly alters the model's process for selecting the next word in a sentence.

Governance

Investigation into Amazon and Walmart Over AI Chatbots' Product Bias

17 Sept 20266 Rabiʻ II 1448 AH

Governance

17 Sept 20266 Rabiʻ II 1448 AH

Investigation into Amazon and Walmart Over AI Chatbots' Product Bias

Senators Tammy Baldwin and Rick Scott have urged the Federal Trade Commission to investigate Amazon and Walmart, claiming their AI chatbots may obscure 'Made in America' products. A study highlighted that shopping assistants Alexa and Sparky favor displaying products from countries like China, undermining efforts to promote U.S. manufacturing. The senators noted that while these chatbots can detect counterfeit 'Made in America' labels, major retailers do not leverage this technology.

Models

Base Labs launches an open-weight AI safety partnership with Hugging Face and Goodfire

17 Sept 20266 Rabiʻ II 1448 AH

Models

17 Sept 20266 Rabiʻ II 1448 AH

Base Labs launches an open-weight AI safety partnership with Hugging Face and Goodfire

Base Labs, formed by Baseten earlier this year, has announced plans to develop and publish new methods for training and monitoring open models. This initiative comes at a time when interest in open models is growing, as companies seek to improve the efficiency and reliability of these models. The new methods will enhance the ability to monitor model performance and adjust them according to changing needs.

Models

Pinterest teases a new ‘Restyle’ feature that lets you redesign your room with AI

17 Sept 20266 Rabiʻ II 1448 AH

Models

17 Sept 20266 Rabiʻ II 1448 AH

Pinterest teases a new ‘Restyle’ feature that lets you redesign your room with AI

Pinterest is currently testing a new feature called Restyle, which utilizes AI technology. This feature allows users to visualize furniture, decor, and lighting in photos of their own rooms, making it easier for them to make purchasing decisions. The Restyle feature aims to turn saved inspiration into reality, enabling users to see how different items would look in their own spaces. This move reflects Pinterest's commitment to integrating modern technology to enhance user experience.

Infrastructure

Huawei plans Q1 2027 launch of new AI chip as it takes on Nvidia

17 Sept 20266 Rabiʻ II 1448 AH

Infrastructure

17 Sept 20266 Rabiʻ II 1448 AH

Huawei plans Q1 2027 launch of new AI chip as it takes on Nvidia

Huawei is accelerating the launch of its Ascend 960DT AI chip, aiming to enhance its competitiveness against Nvidia. This initiative is part of Huawei's efforts to close the AI computing gap between China and the U.S., reinforcing its position in the market. This development underscores Huawei's commitment to advancing AI technologies, which could significantly impact the computing landscape in China and foster innovation in the sector.