Launches

Anthropic opens free AI vulnerability scans to open-source projects

Maintainers can opt in to periodic model-generated security reports, but the findings arrive without human review.

Goodfire launches activation-based cyber monitors for AI agents

The system reads internal model signals during agent work and sends selected exchanges to an AI judge, with cost and detection gains reported in company tests.

Natura previews a smart ring for voice requests to AI agents

Interface is designed to send spoken requests to connected agents, capture notes and return results through Natura's companion app.

Google releases an offline-first AI meeting notes app for Mac

AI Edge Foresight uses on-device Gemma models to turn meeting audio and shorthand notes into summaries and to answer questions from local documents.

Google Cloud announces a universal Gemini agent for work

Google says the enterprise agent can plan tasks, use company tools and context, and return completed work inside documents, inboxes, and developer environments.

Upscale AI unveils Token Fabric to link rival AI chips in data centers

The Nvidia-backed startup announced a hardware and software networking platform for mixed-supplier AI data centers, with delivery planned in stages.

Oxford Economics launches AI assistant for macroeconomic scenarios

AskOE AI Model Assistant turns natural-language assumptions into scenarios run through the firm's Global Economic Model.

Nous Research raises $90 million and introduces Hermes for Businesses

A reported $90 million Series B backs an enterprise version of the Hermes agent, though access and deployment details remain unclear.

Microsoft previews Copilot actions across local files and Windows

Microsoft showed an Autopilot tax-preparation workflow using local files, but says the underlying Copilot features will arrive over the next few months.

Developing 1 source

OpenAI starts rolling out GPT-6 and interactive answers in ChatGPT

A new ChatGPT model can compose answers with interactive charts, controls and task-specific tools; paid tiers get the rollout first.

GitHub previews AI classifier to block unstructured secrets before code is pushed

A ModernBERT-based detector is in private preview for push protection; GitHub's own measurements suggest exposure volume follows rising code activity.

AWS releases Strands Box preview for policy-controlled AI agents

The open-source developer preview pairs operating-system isolation with Dogwood rules for agent actions, but initial local support is limited to Apple-silicon Macs.

Liquid AI releases open-weight d1 decision models for edge devices

The downloadable d1-3B and experimental d1-omni-600M make bounded decisions from multimodal inputs instead of writing free-form answers.

Source-backed 3 sources

Google opens SynthID Detector site for checking AI media watermarks

A dedicated site extends watermark checks beyond an earlier limited test, but a negative result cannot prove that media was made without AI.

Tab emerges from stealth with a messaging-based personal AI assistant

Tab says its assistant can handle errands through messages, but its public site currently offers a private-beta waitlist rather than open access.

Rivercell announces $25 million seed round for AI virtual cell work

The Paris startup plans to expand a wet lab and cell-response data platform, then train a model to predict how cells react to treatments.

Google launches Playground for prompt-built browser games

The experimental service lets eligible users build and share browser games by describing rules and changes in text, with Unity integration still in testing.

OpenAI previews College Planner and new study tools for ChatGPT teens

The planned College Planner would organize applications and financial aid deadlines; an iOS note-capture feature is available now.