Sakhanda Wire
NVDA $154.73 -0.47% MSFT $505.75 -0.10% GOOGL $283.93 +1.33% META $619.12 +0.04% AMZN $264.02 +2.38%

Latest news

Older stories

Ars Technica

AI prior-authorization pilot draws scrutiny over Medicare care denials

In January, the Trump administration launched WISeR, a pilot program using artificial intelligence to approve or reject certain types of care for Medicare patients. The federal health program for seniors had not previously required doctors to obtain this kind of prior authorization for those services.

The Decoder

Anthropic signs $11.6 billion, seven-year cloud deal with Akamai

Anthropic has agreed to a seven-year cloud-computing contract with Akamai Technologies valued at $11.6 billion, Reuters reported. The arrangement includes a warrant allowing Anthropic to acquire as much as 5 percent of Akamai’s shares. Two percentage points are connected to the current agreement, while the remaining three could apply if the contract grows by up to another $9 billion.

The Decoder

DHH says AI has ended his practice of writing code by hand

David Heinemeier Hansson, co-founder of Basecamp and creator of Ruby on Rails, says he has stopped writing code by hand after 25 years as a professional programmer. He made the announcement during a keynote at Rails World 2026, saying he had not typed a line himself since around March 2026.

MIT Tech Review

Pentagon Seeks $30.3 Million for AI-Assisted Lie Detector

The US government has requested $30.3 million over five years for Polygraph+, also called Polygraph Next. The program would develop artificial intelligence and machine-learning algorithms for credibility assessments, as well as standoff sensing, which measures physiological signals without attaching equipment to a person. The budget proposal has not yet been approved by Congress.

The Decoder

White House seeks U.S. review before British testing of new AI models

The White House has asked OpenAI and Anthropic not to provide new artificial-intelligence models to the U.K. AI Safety Institute (AISI) until U.S. agencies have reviewed them, Politico reports. The request came from the Office of the National Cyber Director, putting both companies in the position of delaying access for AISI or risking opposition from the White House.

AI News

Why AI Workloads Queue Beside Apparently Idle GPUs

GPU utilisation usually measures the share of streaming multiprocessors active during a sampling period. It does not show whether a device is allocated or whether its VRAM can accept another workload. A model serving occasional requests may keep compute activity at 5%, while its GPU and memory remain reserved. Kubernetes normally assigns an entire GPU to one pod, so another workload waits even when the device appears mostly idle.

AI News

AI Can Extend Legacy Software Without Replacing Its Core

Companies do not necessarily need to rebuild long-running software to introduce artificial intelligence. A full rewrite can be costly and risky because older systems contain accumulated business rules, exceptions and regulatory logic that may be poorly documented. Keeping the existing foundation also preserves processes tested through years of use. The article presents AI as an additional layer around that foundation rather than a replacement for it.

AI News

Seven AI Tools Target Document Fraud in Enterprise Onboarding

Enterprise onboarding increasingly relies on passports, business registrations, bank statements, tax filings and other documents that can be altered or fabricated. The article identifies four major fraud types: physical counterfeits or presentation attacks, edits to genuine files, template-based serial fraud and fully AI-generated documents. Detecting reused templates and images across applications is important because single-document checks can miss organized fraud networks.

TechCrunch

Lightspeed targets $250 million for new India AI fund

Lightspeed is seeking $250 million for Lightspeed India Partners V, a new early-stage fund focused on artificial intelligence in India and Southeast Asia. The vehicle will be half the size of the firm’s $500 million predecessor, raised in 2022, and has secured commitments covering 80% of its target, according to a letter to investors.

MarkTechPost

Fastino releases 340M open-weight model for structured decisions

Fastino Labs has released GLiNER2.5-Decide, a 340M-parameter open-weight model for operational decisions in AI agent pipelines. It accepts text and a schema of typed questions, then returns structured answers with probability distributions, confidence scores and constraint-feasibility metadata. The weights are available under Apache 2.0, can run on CPU or GPU, and are suitable for air-gapped environments. Hosted inference and fine-tuning are also available through the GLiNER API.

MarkTechPost

FLUX 3 Action leads RoboLab-120 with a 7B robot-control model

Black Forest Labs has released FLUX 3 Action, a 7B open-weights World Action Model for robot control. It processes camera images, robot-state data and a text instruction, then jointly predicts future video frames and the next sequence of robot actions. On RoboLab-120, it recorded a 42.92% task-success rate, ranking ahead of the listed alternatives.

The Verge

Google gives Gemini Live a real-time animated avatar

Google’s Gemini 3.8 Live update adds Live Avatar, an animated persona that responds while users speak with the model. The avatar synchronizes its mouth movements with speech and changes facial expressions during conversations. The feature is currently available only to Gemini Enterprise customers.