Sakhanda Wire
NVDA $150.96 -2.89% MSFT $508.29 +0.40% GOOGL $282.44 +0.80% META $616.38 -0.40% AMZN $267.74 +3.82%

Latest news

Older stories

The Verge

OpenAI introduces a dedicated ChatGPT mode for teenagers

OpenAI is introducing a dedicated ChatGPT mode for teenagers that brings existing youth safeguards together with additional safety features and parental controls. The company says the new experience is intended for teens who use the chatbot to learn, think critically, deepen their understanding and use artificial intelligence with confidence.

The Verge

FAA outages disrupt flights and air traffic operations

On August 6, the Minneapolis Air Route Traffic Control Center experienced a failure involving both radar and communications. The disruption lasted about two hours and affected more than 1,100 flights moving through the facility’s airspace. The center manages a sector covering 330,000 square miles across nine states.

The Decoder

Anthropic took 65.1% of Vercel Gateway spending in July

Anthropic generated 65.1 percent of spending on Vercel’s AI Gateway in July, despite representing 30 percent of the tokens processed. According to Vercel’s figures, Anthropic’s average price per token was 4.4 times higher than the average for all other providers. This gave the company substantially more revenue per use than competing services.

MarkTechPost

Sonic-3.6 Leads Two Speech Synthesis Leaderboards

Cartesia has released Sonic-3.6, the latest version of its real-time text-to-speech model, about three months after Sonic-3.5. The system is ranked first on both Artificial Analysis speech leaderboards, with 1,283 Elo on Provider Voice and 1,123 on Controlled Voice. The latter evaluates models using the same eight reference voices; Sonic-3.5 ranks second and ElevenLabs Eleven v3 third.

MIT Tech Review

AI Observatory seeks clearer evidence of how people use AI

AI companies including Anthropic and OpenAI publish usage reports for products such as Claude and ChatGPT, but researchers say those reports provide only a partial view. The new AI Observatory, co-led by Stanford researcher Anka Reuel and Shayne Longpre, aggregates and analyzes consented conversations from seven existing datasets to offer an independent source for researchers and policymakers.

The Decoder

Claude Code adds /design for terminal-based UI mockups

Anthropic has released an early preview of the /design command for Claude Code. The feature allows developers to create interface mockups in the terminal or desktop app before beginning implementation. A prompt such as "/design a few options for {feature}" asks Claude to produce several drafts as artboards.

AI News

GLM-5.3 leads one cyber benchmark but trails on two others

Zhipu, also known as Z.ai, launched its coding-focused GLM-5.3 on August 14 and published results against competing models. On CyberGym, which tests whether a model can identify and validate software vulnerabilities from source code, GLM-5.3 scored 84.5%. That was slightly above Anthropic’s Mythos 5 at 83.8% and OpenAI’s GPT-5.6 Sol at 83.6%.

The Verge

Leaked video appears to show Apple’s camera-equipped AirPods

A short video found by MacRumors in the macOS Tahoe 26.7 Release Candidate may offer the first look at Apple’s rumored camera-equipped AirPods. The clip shows a man wearing the earbuds while holding up a book, with its cover facing the camera. Visual Intelligence identifies the book title in the demonstration.

MIT Tech Review

Study questions how soon AI can improve its own research

A study led by Peter Kirgis and Sayash Kapoor of Princeton University suggests that AI agents may be farther from recursive self-improvement than some industry forecasts imply. The researchers found that agents could handle many technical tasks involved in AI research, but struggled with the judgment and creativity needed for open-ended investigation.

The Decoder

Anthropic’s annualized revenue surpasses $65 billion

Anthropic’s annualized revenue rate exceeded $65 billion based on its performance through the end of July 2026, according to people familiar with the matter who spoke to Bloomberg. The figure represents a sevenfold increase compared with the same point a year earlier. One source said the company included the figure in a regular update for investors.

MarkTechPost

Hermes Agent adds named bots with separate profiles and collaboration

Nous Research has released Bot Mode for Hermes Agent, its MIT-licensed open-source agent. The feature replaces the single-agent session list with a roster of named bots. Each bot is a Hermes profile with its own chat history, memory, skills, credentials and pinned model. Profiles are stored separately under ~/.hermes/profiles/<name>/.

MarkTechPost

ByteDance and Tsinghua present CUDA Agent for faster GPU kernels

ByteDance Seed and Tsinghua AIR have introduced CUDA Agent, a reinforcement-learning system designed to generate CUDA kernels that outperform compiler-generated code. On the 250-task KernelBench benchmark, the underlying Seed1.6 model passed 74.0% of tasks but produced kernels faster than torch.compile in only 27.2% of cases, with a 0.69× geometric-mean speedup.