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 Decoder

Anthropic will make Auto Mode the default in Claude Code

Anthropic will enable Auto Mode by default in Claude Code for Pro, Max, and Team subscribers starting August 14. Enterprise customers will still have to activate the feature themselves. Auto Mode allows the coding assistant to proceed without requesting approval for every operation. A classifier evaluates actions and asks for confirmation when it identifies a potentially dangerous or irreversible step.

The Decoder

Claude Code sessions can now exchange messages across terminals

Claude Code has added communication between sessions running in parallel on macOS and Linux. The feature allows one session to send a text summary to another, reducing the need to copy information manually between terminals. Anthropic says the exchange works in both directions: a session can ask another one a question and receive the answer directly.

The Decoder

Backflip AI converts 3D scans into editable CAD models

Backflip AI has released its second AI model for computer-aided design. The system converts 3D scans and mesh files into fully editable, parametric CAD models. The company says a task that usually requires substantial time and advanced CAD expertise can be completed in a few minutes.

The Decoder

Fields Medalist Jacob Tsimerman joins OpenAI to work on AI safety

Jacob Tsimerman, a newly awarded Fields Medalist and number theorist at the University of Toronto, is joining OpenAI to work on artificial intelligence safety. He has described AI as an extremely transformative technology and said society should devote substantially more effort to understanding and managing its risks.

Ars Technica

WeatherNext gave forecasters an extra day to predict Hurricane Melissa

In October 2025, a storm forming over the Caribbean Sea presented weather models with sharply different possibilities. Some projections kept it weak and directed it toward Haiti, while others indicated that it could strengthen and move toward Jamaica. WeatherNext, an artificial intelligence model developed by Google’s DeepMind and Google Research, supported the second scenario.

The Decoder

AI agents consume far more energy than simple chat prompts

Climate scientist Zeke Hausfather tracked eight weeks of his work with Claude Code and estimated the electricity used by the agent. His 1,138 typed prompts triggered more than 14,000 model calls and processed 3.2 billion tokens. The estimated data-center consumption was about 170 kWh, with a range of 70 to 330 kWh. That equals roughly 150 Wh per prompt, or about 600 times the 0.24 Wh reported for a median Gemini text prompt.

MarkTechPost

XY tutorial demonstrates scalable interactive visualization workflows

A tutorial on the XY Python library presents techniques for building interactive charts that remain usable with complex or large datasets. It shows how multiple marks, dual axes, annotations, tooltips, legends, themes and navigation controls can be combined in one chart declaration. Examples also use Pandas DataFrames, named data channels, regional facets, shared axes and linked selections.

The Decoder

xAI launches Imagine Image 2.0 for Grok with new editing tools

xAI has launched Imagine Image 2.0 as a new Quality Mode on grok.com/imagine and in Grok’s iOS and Android apps. The company says the image generator is built to follow detailed instructions, preserve clean typography and layouts in complex images, and maintain consistency across repeated generations. API access is planned for a later release.

The Decoder

OpenAI pauses Astra work over possible critical cyber capabilities

OpenAI has paused some development activities for Astra after internal evaluations found major progress in agentic coding and cybersecurity. The company said it could no longer rule out the model reaching the highest, Critical, level in its Preparedness Framework. This is the first time OpenAI has identified that possibility for one of its models; previous systems, including GPT-5.6-Sol, were rated no higher than High.

MarkTechPost

Mistral AI releases 3B multimodal safety classifier Shieldstral 1.0

Mistral AI has released Shieldstral 1.0 3B, an open-weights multimodal safety classifier that lets operators define moderation policies at inference time. Instead of relying on a fixed list of harm categories, the system evaluates a plain-language yes-or-no question and produces a calibrated safety score in one forward pass. It is built on Ministral-3-3B-Base-2512 and includes a native Pixtral vision encoder, under the Apache 2.0 license.

TechCrunch

OpenAI pauses parts of Astra development over cybersecurity concerns

OpenAI has paused some work on Astra, an unreleased model, after an internal review found major progress in agentic coding and cybersecurity. The company said Astra reached its “critical cybersecurity threshold,” a designation meaning the model could independently identify and conduct cyberattacks against traditionally well-protected real-world systems.

MarkTechPost

Tencent Cloud open-sources team memory hub for AI coding agents

Tencent Cloud has released TencentDB Agent Memory v2.0 as an MIT-licensed, self-hosted memory hub for teams using AI agents. The stable 2.0.0 release was published on August 3, 2026. It converts conversations, documents and code into four governed asset types: Chat Memory, Skill, LLM-Wiki and Code-Graph. These assets carry ownership, versions, status and visibility settings, and can be equipped for specific agents.