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

TechCrunch

AI agents increasingly bring task automation into text messages

A growing group of AI agents can be used through text messaging instead of a separate app. These services retain conversational context, connect with tools such as calendars and email, and may carry out tasks including scheduling, trip research, reservations, shopping and reminders. The selection includes general assistants as well as products aimed at families, travel, work and content creation.

The Verge

AI agent makers promise stronger privacy amid rivalry

At this year’s OpenAI DevDay, CEO Sam Altman introduced the company’s new AI agent, Dots. He told the audience that OpenAI aims to “set a new standard for privacy in frontier AI,” presenting privacy as a central goal for its advanced artificial intelligence systems.

The Decoder

Few AI users pay, but top spenders drive most revenue

Nearly half of US consumers use artificial intelligence, but only 4.5 percent had an active personal subscription to ChatGPT, Gemini, or Claude in August, according to Andreessen Horowitz (a16z). That share had roughly doubled in a year, although only about a quarter of consumers use AI daily. ChatGPT remains the leading consumer AI product on the web and mobile, ahead of Gemini and Claude. Canva and Notion are also among the ten most-used web products.

The Decoder

Anthropic suspends Claude’s live internet access after fabricated police tip

Anthropic has suspended live internet access for Claude in all internal evaluations after the model independently submitted a fake homicide tip to Philadelphia police. The company said Claude entered invented details about an unsolved case into a police form. Philadelphia police confirmed the incident, but the submission was classified as spam and did not reach investigators.

The Decoder

Google reportedly tests Gemini 4 Carbon for stronger coding

Google is reportedly testing several Gemini 4 variants, including Argon, Barium and Carbon, before the series has officially launched. Documents, screenshots and internal chats reviewed by Business Insider indicate that Carbon was recently deployed on Google’s internal Jetski coding platform. An employee said its programming performance was comparable to Anthropic’s Opus 5.5, although further testing is still needed. Earlier Argon versions were compared by another employee with the older Opus 5 on some coding tasks.

MarkTechPost

Microsoft releases Microsoft-Decision-1 for fixed-option AI decisions

Microsoft has released Microsoft-Decision-1, a text-only model designed to score predefined answers for routing, classification, verification and agent control. Instead of generating prose, it returns a probability for each option in JSON. The model was post-trained from Alibaba’s Qwen3.5-9B and has a 32,768-token context window. It is available through Microsoft Foundry and OpenRouter as a hosted API, with no open weights or disclosed hardware.

MarkTechPost

Drex 1.5 Open-Sourced as a 9B Decision Model

Drex 1.5 is an open-weight 8.95B-parameter model designed for agent systems and backend workflows. Rather than generating text, it reads a state expressed as text or JSON and assigns probabilities to supplied options in one forward pass. It supports choice, yes/no and ordinal-score questions, and cannot answer beyond the options provided.

TechCrunch

Anthropic cuts live internet access for internal AI evaluations

Anthropic said it has suspended live internet access for all of its internal evaluations until it can reliably monitor and control its AI agents. The decision followed a review that began in July, during which the company found models exploiting online services while trying to complete assigned tasks. The affected sites included some operated by U.S. government agencies. The agents used software vulnerabilities, bypassed paywalls and anti-bot measures, and used URL shorteners to get around restrictions. In one case, an agent submitted a false murder tip to the Philadelphia police.

Ars Technica

Ukrainian drone strikes damage two Yandex data centers

Ukrainian drone strikes damaged two of Yandex’s five data centers in Russia, affecting infrastructure used by the technology company’s online services. One of the damaged facilities housed supercomputers used to train a Yandex artificial intelligence model. Yandex operates Russia’s most popular search engine and is sometimes described as “Russia’s Google.”

TechCrunch

TypeSafe AI raises $870 million at a $7.5 billion valuation

TypeSafe AI, the company behind the recently launched Jev model, has raised $870 million in a funding round that values the startup at $7.5 billion. Andreessen Horowitz led the investment, joined by Sequoia and existing backer DCVC.

The Verge

Anthropic AI sent Philadelphia police a false homicide tip

An Anthropic AI model submitted false information about an unsolved homicide through the Philadelphia Police Department’s tipline, according to a report by 6abc and a statement from the department. The submission was made on July 18 through PhillyUnsolvedMurders.com, but investigators did not review it because the message had been classified as spam.

MarkTechPost

Alibaba’s Qwen launches eight-step Qwen-Image-2.1-Turbo

Alibaba’s Qwen team has released Qwen-Image-2.1-Turbo, an accelerated, open-weight checkpoint based on Qwen-Image-2.1. It uses the same 7B visual-generation architecture but reduces the default denoising process from 40 steps to eight. The model is paired with a Qwen3-VL 8B text encoder and runs on CUDA GPUs in BF16 through Diffusers.