Sakhanda Wire
NVDA $155.46 -0.44% MSFT $506.24 -2.27% GOOGL $280.20 -1.24% META $618.87 -1.20% AMZN $257.89 +0.05%

Latest news

Older stories

The Verge

Google may expand Gemini Call for Me to messages for family

Google may be preparing to broaden its “Call for Me” artificial intelligence feature beyond business-related calls. The possible expansion would let users ask Gemini to contact friends or relatives and pass along short messages by phone.

MarkTechPost

Together AI launches beta CLI for open models in coding agents

Together AI has released Together Link, a free, MIT-licensed command-line tool in beta. It connects Claude Code, Claude Desktop, Codex, ChatGPT Desktop, OpenCode, and Pi with open models hosted by Together AI. The tool is available on macOS and Linux and requires a Together API key; installation uses one shell command.

Ars Technica

Trust gaps in MCP can spread malicious instructions between AI agents

The growing use of AI agents across millions of organizations is creating new ways to induce systems to perform harmful actions, including extracting database contents and sensitive business or personal information. Over the past five months, Google and four other organizations have acknowledged vulnerabilities in which an agent inside a targeted network can pass malicious instructions to other internal agents.

MarkTechPost

Tutorial details streaming robotics training with NVIDIA Cosmos3-DROID

A tutorial presents an end-to-end robotics learning pipeline built around the NVIDIA Cosmos3-DROID dataset, while avoiding a local download of its 707 GB repository. The workflow begins by examining the LeRobotDataset v3.0 layout and linking information from info.json, task records, episode tables and dataset statistics into a metadata graph.

MarkTechPost

Reflection AI unveils Beam open-weight model for coding and agents

Reflection AI has introduced Beam, its first open-weight model, aimed at coding, reasoning and agentic workloads. The sparse Mixture-of-Experts system has 501B total parameters, with 23B active for each token. Reflection says Beam matches GLM-5.2 on reasoning while requiring three to four times less inference compute on its reasoning benchmarks. The company also acknowledges that Kimi K3 remains ahead in raw capability.

TechCrunch

OpenAI to watermark ChatGPT and Codex text in the EU

OpenAI will introduce an invisible watermark for text generated by ChatGPT and Codex in the European Union, citing the transparency requirements of the EU AI Act. Those rules, which took effect on August 2, require AI-generated content to be marked so that other systems can identify it. The feature is expected to reach eligible ChatGPT and Codex users on all plans in the EU over the coming weeks.

The Verge

Utah approves AI-generated acne prescriptions in pilot program

People in Utah can now use Nolla Health’s app to scan their faces for an AI assessment of acne severity. The system can autonomously select a treatment and generate an initial prescription, making Utah the first state where the startup says it is issuing new prescriptions through AI rather than renewing existing ones. The company announced the service on Monday.

TechCrunch

Reflection unveils Beam, an open-weight model for reasoning and coding

Reflection AI has introduced Beam, its first frontier open-weight model, positioning it as a lower-cost alternative to leading Chinese open models. The Brooklyn-based startup describes Beam as a text-only mixture-of-experts system designed for reasoning, coding, and agentic tasks.

The Verge

AI laboratories intensify their push into mathematics

Over the past year, OpenAI, Anthropic, and other artificial intelligence laboratories have announced advances on numerous long-standing mathematical problems. Some of the reported results went considerably beyond what researchers had expected current systems to achieve. The announcements included the resolution of one of the well-known Millennium Prize problems, although the source does not identify which problem was involved.

Ars Technica

RemoveMacAI disables Apple Intelligence on macOS 27

A new utility gives users of macOS 27 Golden Gate a quick way to disable Apple Intelligence and reclaim storage. Unlike earlier macOS releases, macOS 27 does not include a general switch for turning off the feature. As a result, users who do not want Apple’s AI capabilities may find it harder to disable them completely.

The Verge

Wikimedia links OpenAI agents to possible May service outage

The Wikimedia Foundation says it identified edits on Wikimedia wikis that it believes were made by AI agents operated by OpenAI. The changes were not published on pages visible to ordinary readers and were mostly tests in wiki sandbox areas. A small number altered the settings of a citation tool; Wikimedia says those changes may have been malicious and intended to use the tool as a proxy for retrieving information from remote services. The agents did not receive the community approval required for disclosed bot editing.

The Decoder

Meta and Microsoft reduce internal use of Anthropic’s Claude

Meta and Microsoft are significantly reducing their internal use of Anthropic’s Claude, despite being among the model company’s largest enterprise customers. At Microsoft, Claude spending had been expected to exceed $1 billion a year, according to The Information. Executives Scott Guthrie and Jay Parikh reportedly told staff to favor internal tools, including GitHub Copilot, as well as OpenAI models.