AI
Firefox's new Smart Window feature now integrates real-time web information and provides source links in AI chat responses through a partnership with Exa. The update also introduces automatic tab grouping and visual previews for browsing history.
Aug 18, 2026 · 2 min read
5 connections in the Atlas
AI
Z.ai has announced a two-week delay in the public release of its GLM-5.3 model's open weights, citing unexpectedly strong cybersecurity capabilities. The company intends to implement additional safety controls before making the model broadly available, acknowledging its potential for both defensive and offensive applications.
Aug 18, 2026 · 2 min read
7 connections in the Atlas
AI
Google now permits users to disable visible watermarks on AI-generated images, videos, and music from Gemini and Flow. Invisible SynthID watermarks and C2PA metadata will continue to be embedded for provenance tracking.
Aug 16, 2026 · 2 min read
8 connections in the Atlas
AI
OpenAI has introduced an opt-in "Computer History" feature for its ChatGPT macOS desktop application, enabling the AI to track user activity across applications and websites. This functionality aims to help ChatGPT learn user workflows, suggest automations, provide context for ongoing tasks by building a searchable timeline of interactions.
Aug 16, 2026 · 2 min read
5 connections in the Atlas
AI
Twitch now allows users to opt out of Amazon using their channel content for generative AI model training. The platform faced significant user criticism after initially enabling the setting by default.
Aug 15, 2026 · 2 min read
AI
Google DeepMind has launched Gemini 3.7 Flash, an updated version of its efficient AI model, just three weeks after its predecessor. The new model offers enhanced performance in software engineering, web development, and complex knowledge work, alongside an introductory price reduction.
Aug 14, 2026 · 3 min read
24 connections in the Atlas
AI
The White House is reportedly planning to broaden its artificial intelligence policy to encompass open models, indicating a potential shift in its regulatory strategy. This expansion suggests the administration is continuing to address the governance of AI, including systems with publicly available components.
Aug 13, 2026 · 3 min read
4 connections in the Atlas
AI
Twitch has introduced a new setting allowing streamers to prevent their content from being used to train Amazon's generative artificial intelligence models. This opt-out option, found in account security settings, covers streams, VODs, clips, and channel text, though other AI-powered features will remain active.
Aug 12, 2026 · 2 min read
11 connections in the Atlas
AI
Google DeepMind has released its sign-language-to-text (SL2T) model, enabling American Sign Language (ASL) translation to English text on Pixel 11 phones. This integration allows Deaf and hard-of-hearing users to sign directly into Gboard and Live Transcribe for various communication tasks.
Aug 12, 2026 · 3 min read
1 connection in the Atlas
AI
SpaceXAI has launched Grok Bot, an AI agent service designed to operate as independent "AI teammates" that can execute multi-step workplace tasks. These bots can sign into user accounts and applications, working autonomously in a cloud-based environment. The service aims to shift AI assistants beyond simple prompt responses to continuous work execution, returning to users only for approvals or upon task completion.
Aug 12, 2026 · 3 min read
7 connections in the Atlas
AI
Google's research AI system, AMIE, has shown expert-level performance in real-time clinical video consultations during simulated patient interactions. The system matched or exceeded primary care physicians across several clinical competencies and patient-reported metrics.
Aug 11, 2026 · 2 min read
4 connections in the Atlas
AI
Anthropic will embed invisible watermarks into text generated by its Claude AI models and digitally sign image files. This global implementation aligns with European Union regulations requiring AI transparency. The markings aim to signal AI processing but do not definitively prove AI authorship.
Aug 11, 2026 · 2 min read
12 connections in the Atlas