
Claude adds Motion and Dashboards
Anthropic's beta features now allow the creation of animated explainer videos and live data dashboards from BigQuery/Snowflake.
ddNews Β· a daily AI briefing for Dirk (generative video/ddStudio, Lit

Anthropic's beta features now allow the creation of animated explainer videos and live data dashboards from BigQuery/Snowflake.

The new agentic AI can plan, execute tasks across business apps, and delegate work to specialized sub-agents.

A cost-effective monitoring system that inspects model internals to detect rogue agent behavior.

A wearable device allowing users to summon AI agents for task completion and device control.
A five-year investment aimed at advancing scientific research through AI hardware and software.
New findings show agents can write faster CUDA kernels, though benchmarking remains a challenge.

Researchers warn that AI-driven mathematics could potentially break wallet signature schemes.

Significant funding round for the agent startup following its split from Meta.

A single prompt could allow attackers to take over all AI agents within the same AWS account/region.
The title launches on the cloud with GeForce RTX-powered performance.

An experimental note-taking app that performs meeting transcription entirely offline.

Technical breakdown of production-grade multi-agent systems using Google ADK Java.

Google is launching a background agent that works across apps and devices for work tasks.

New top-level domains are being introduced to cater to the agentic AI ecosystem.

New 9B and 27B open-weight models designed for option selection rather than text generation.

New hardware specifically optimized for the AI era and large-scale model deployment.
Record-breaking quarterly earnings driven by AI-related hardware demand.
Backed by Nvidia and Samsung, the developer of Hermes Agent is scaling business-user agent tools.

Technical advancement in diffusion-based models to improve generation quality during coarse transitions.
A consortium including Google DeepMind and Meta is building AI models to predict cellular behavior for disease combat.
A high-performance small model that significantly outperforms its predecessor in computer-use tests at a fraction of the cost.

The Hermes Agent developer raised $90M in Series B to expand business AI agents.
A major scientific milestone using an unreleased frontier model to prove long-standing hypotheses via GitHub.
Integration of AI agents directly into team conversations and workplace management.

New AI PCs designed specifically to run local AI models and agents.
A networking shift that treats the network as the backplane to support massive AI clusters.

The new model generates interactive interfaces, including charts, buttons, and forms, rather than just text.
A strategic partnership to optimize Windows PCs for local AI agents, featuring the RTX Spark.

The AI agent is rapidly scaling its ecosystem following a successful mobile debut.

A high-performance small model that significantly outperforms its predecessor in computer-use tests at a lower price point.
Using domain-specific context to reduce hallucinations in enterprise agent workflows.

New hardware powered by Nvidia RTX, targeting high-performance AI workloads.

A survey indicates that while velocity is up, AI coding agents are creating a "comprehension gap" and increasing debugging needs.

A consortium including Meta and Google DeepMind is training AI to predict cellular behavior.

A public tool to verify if images, videos, or audio were generated by Google, OpenAI, or Nvidia.

Copilot can now access local files and take direct actions across the Windows operating system.

A specialized interface for classification and probability tasks that is 10x faster than the standard Responses API.

A high-end Nvidia-powered developer machine priced at $5,999 for local AI development.

Rolling out alongside GPT-6, this feature allows the chatbot to answer queries with interactive visuals.

A browser-based platform allowing users to create custom games using Gemini via simple text prompts.
An optimization engine designed to tune enterprise AI agents continuously after they have gone live.

Google's new tool covers images, video, and audio clips for public verification.

A $300M joint investment to create a digital cell model for disease combat.
A new framework for mapping structural variance to estimate coding agent reliability without ground truth.

A collection of Lean-formalized mathematical proofs has been released on GitHub for machine verification.

Using frontier models in a harness to automatically generate and test new attack variations against its own firewall.

The distributed SQL database now runs on-premises and across clouds, removing the need for proprietary atomic clocks.

A lightweight, open-weight decision model optimized for real-time content moderation.

The Nvidia-backed AI computing startup is scaling infrastructure with a $14.5B valuation.

Efforts are underway to create a standard that allows personal agents to bypass anti-bot defenses for bookings and shopping.

A highly efficient 740M parameter open embedding model that runs on-device with only 191MB RAM.

A collection of AI-generated mathematical proofs and Lean formalizations released on GitHub.

A new model designed to predict human behavior for market research, moving beyond LLM role-play.

New features added for data context, preparation, and storage to support enterprise AI pipelines.
Telecom operators are prioritizing open models for greater control, trust, and customization of critical infrastructure.
A PyTorch implementation that reduces data usage by 94% by syncing a remote machine with a cloud-based world model.

A new open-source CLI (`cf`) designed specifically for interacting with AI agents on Cloudflare services.

New rubrics for training generalist LLM agents across diverse interactive environments.