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

TechCrunch

Atoms may expand into autonomous taxis under Travis Kalanick

Travis Kalanick’s startup Atoms is reportedly considering a move into autonomous ride-hailing. The Financial Times reported that the company is preparing to hire extensively and pursue acquisitions that could give it a significant position in the autonomous-vehicle industry. Atoms has also discussed with Uber how the ride-hailing company might use its robotaxi technology.

The Decoder

WeatherNext 3 forecasts weather from live satellite data

Google Research and DeepMind have released WeatherNext 3, an AI weather model that uses live geostationary satellite observations instead of relying on traditional physics-based simulations. It refreshes forecasts every hour and produces them on grids as fine as five kilometers, compared with the 25-kilometer, six-hour grid used by WeatherNext 2. Temperature and humidity are modeled at five kilometers, while other variables use resolutions of 10 or 25 kilometers.

The Decoder

OpenAI developer says Astra moved some plans forward by six months

OpenAI developer Thibault Sottiaux said Astra may have been the company’s “biggest competitive advantage” before it became publicly available. In a post on X, he attributed the assessment to the system’s internal use and said the resulting productivity increase was large enough for some plans to be brought forward by six months.

The Decoder

Google adds Lyria 3.5 music generation to Gemini

Google has released its Lyria 3.5 music generation model in the Gemini app and through an API. The company says the updated model produces more expressive vocals and richer musical arrangements than its predecessor. Users can select a genre and style, decide whether to create a vocal or instrumental track, and generate either short or longer pieces.

The Decoder

Meta releases Muse Voice Transcribe for real-time multilingual speech

Meta has released Muse Voice Transcribe, a real-time audio model developed by its Superintelligence Labs. It transcribes speech, identifies up to 20 speakers and marks sentence boundaries within one system. Incoming audio is processed in 80-millisecond segments, while the model decides how long to listen before producing each word. Meta says reinforcement learning taught it to give easy words shorter delays and difficult ones more processing time.

The Decoder

Abliteration.ai sells access to open models with safety refusals removed

US startup Abliteration.ai is offering commercial API access to modified open-weight AI models whose trained refusal mechanisms have been suppressed. Its current service is based on Z.AI’s GLM-5.3, altered through a technique called abliteration, which changes model weights rather than relying on a prompt jailbreak. The company says coding, cyber and agentic abilities remain largely intact.

MarkTechPost

UC Berkeley team releases unified platform for computer-use agents

Researchers led by UC Berkeley have released CUA-Lite, an open platform designed to bring computer-use agents, environments, interaction traces, and evaluation and training tools under common interfaces. It uses one action space and one data schema across desktop, browser, and mobile settings. The software installs on Python 3.12 and its sandboxes run on Docker hosts without /dev/kvm, including cloud instances, continuous-integration runners, and nested containers.

MarkTechPost

GPU stack reuses LLM kernels for faster embedding service

Perplexity has described the serving infrastructure behind pplx-embed and ranking models used in Perplexity Search, Computer and the API Platform. The company says GPU embedding inference is largely standardized on mature Hopper and Blackwell hardware, making runtime software and request handling the main areas for improvement.

TechCrunch

Seattle Times and Newsday sue OpenAI and Microsoft over AI training

The Seattle Times and Newsday have filed a lawsuit against OpenAI and Microsoft, alleging that journalism from their publications was used to train artificial intelligence systems. The organizations argued that the spread of generative AI could leave the news industry “broken beyond repair.”

MarkTechPost

GitHub launches HydraFusion for task-specific model workflows

GitHub has released Project HydraFusion as a research preview in GitHub Copilot CLI. Instead of selecting one model for an entire request, the system creates an execution plan for each coding task and can use models from multiple providers. It is available to users on all GitHub Copilot plans. Developers can enable it with /update, /experimental on, and /model, then select HydraFusion (Research Preview). There are no open weights or self-hosted option, and usage is billed according to the standard token rates of the models involved.

TechCrunch

Hikers rescued after using Gemini to plan Mount Shasta trip

Three young hikers were rescued from California’s Mount Shasta after using Google’s Gemini chatbot to help plan their expedition, according to a report from the Siskiyou County sheriff’s office. The group began its climb at 3am and reached the summit at 7pm, despite guidance that hikers should turn back if they have not reached the top by noon.

MarkTechPost

Hermes Desktop adds one-click setup for local AI models

Hermes Desktop now offers a one-click workflow for running open-weights models locally. The application examines the computer’s hardware, checks available models against the GPU, selects the highest-quality build that fits, downloads its weights and configures the inference runtime. The flow appears on first launch and can later be opened through Settings → Providers → Local Models.