Sakhanda Wire
NVDA $152.04 -2.20% MSFT $504.96 -0.25% GOOGL $281.53 +0.47% META $616.55 -0.37% AMZN $266.35 +3.28%

Latest news

Older stories

MarkTechPost

PrismML releases 5.93 GB ternary Qwen3.8 27B model

PrismML has released Ternary Bonsai 2 27B, a ternary-weight version of Qwen3.8 27B with an Apache 2.0 license. The language model occupies 5.93 GB, compared with 53.80 GB for the FP16 version. It handles text and images and supports a 262K-token context. PrismML says it retained 98.2% of the parent model’s average score across 20 benchmarks.

TechCrunch

Disney appoints former Character.AI CEO as its first CTO

Disney has appointed Karandeep Anand as its first chief technology officer. Anand previously served as CEO of Character.AI, a generative artificial intelligence company founded in 2021 that lets users create virtual characters and interact with them through chat. He became the company’s CEO in May 2025 after serving as a board adviser, according to his LinkedIn profile.

The Decoder

Newsom orders AI oversight panel and calls for model kill switches

California Governor Gavin Newsom has signed an executive order aimed at accelerating independent oversight of artificial intelligence companies and promoting a “kill switch” for AI models. The measure follows recent incidents, including an attack involving Hugging Face, according to the order’s announcement.

TechCrunch

Google tests CC as an AI household coordination agent

Google is testing a new version of CC, an AI agent intended to help families organize daily life. The service works across email, calendars, chats and tasks, and can support activities such as tracking school schedules, sports practices, bills and meal plans. It can also create shopping lists, prepare weekly meal plans and handle some family-specific paperwork, including permission slips and activity registration forms.

Ars Technica

US officials remove Chinese AI search tool from Federal Register

US officials removed a Chinese AI search tool from the Federal Register website on Wednesday, Reuters reported. The tool had been offered to visitors as an option for searching public comments on proposed regulations. The National Archives, which operates the website, had used one of Alibaba’s Qwen open-source AI models. It is not clear when the search feature was introduced or how long it remained available.

The Decoder

Researchers used Claude to breach OpenAI systems in under 72 hours

Three security researchers from Hacktron say they used Anthropic’s Claude models to chain two vulnerabilities in OpenAI’s community forum and reach employee ChatGPT and Codex accounts. From there, they accessed OpenAI’s internal GitHub code repository. The team said it did not view sensitive information and demonstrated access by creating a harmless pull request in the internal monorepo.

TechCrunch

Dario Amodei proposes a plan to pace AI development

Anthropic CEO Dario Amodei has presented a plan for “pacing the frontier” of artificial intelligence development. His proposal comes about a week after a warning from an Anthropic researcher about a possible doomsday scenario unsettled the AI industry.

TechCrunch

AI safety coordination and Automattic power struggle discussed on Equity

A new episode of the Equity podcast examines Anthropic CEO Dario Amodei’s proposal to “pace the frontier” of artificial intelligence development. The plan followed a warning from an Anthropic researcher that had unsettled the AI industry. It calls for independent safety evaluators and cooperation among AI laboratories in democratic countries. The episode also considers whether companies can agree on what slowing AI progress should mean and who would oversee compliance. The proposal has received some support in the industry, while Nvidia CEO Jensen Huang has offered pointed criticism.

The Verge

Newsom Orders Review of AI Safety Rules and Possible Kill Switch

California Governor Gavin Newsom has ordered the state to convene a group of experts to recommend stronger AI safety measures in state law. The group is expected to submit its recommendations within two months. The order places California in a position to consider additional oversight of advanced AI systems.

TechCrunch

Manus discusses $500 million raise at $4 billion valuation

Chinese artificial intelligence startup Manus is discussing a $500 million funding round that could value the company at $4 billion, The Wall Street Journal reported, citing anonymous sources. The talks follow the company’s return to independent operations after its planned combination with Meta was blocked in China.

The Verge

Hollywood Labor Groups Shift Focus to AI’s Current Impact

As technology companies warn that artificial intelligence could pose an existential threat to humanity, entertainment labor groups are directing attention to effects already visible in the industry. Models that create digital facsimiles of actors and seek to reproduce conventional filmmaking tools may endanger some human jobs, even though they differ from systems developed by Anthropic and OpenAI.

The Verge

Researchers used Claude to breach OpenAI employee accounts

Three independent security researchers from Hacktron say they gained access to OpenAI employee accounts in less than 72 hours with help from Anthropic’s Claude Opus 4.8 and 5, the Wall Street Journal reports. Their work involved a corrupted HEIF image and software used by Discourse, the third-party platform hosting OpenAI’s community forums.