Skip to content

AI Flash · live feed

What is happening in AI right now

A timeline of everything that has passed through the radar — newest first, with an evidence status for each item.

1 new news item today 2526 across the entire feed ↗ details, original sources and our own analyses

Continuously updated

Latest items

2526 items in the selected feed

How to read evidence statuses ↓

Wednesday, September 30

24 items
Business and investment only one source so far

OpenAI and Synopsys have entered into a partnership to develop the GPT-Synopsys model for chip design

OpenAI and Synopsys have entered into a multi-year partnership to develop the GPT-Synopsys model. The model is intended to help with chip design and verification and to directly operate Synopsys's tools. The first tests with customers from the semiconductor industry are already underway.

The Decoder (daily AI news) Details and sources →
Security only one source so far

Anthropic suspended accounts after attempts to bypass the safeguards of the Claude product

According to a report by the company Anthropic, foreign researchers attempted in five cases to bypass the safeguards of the Claude product when querying about pathogen modification. The company did not provide the answers and suspended the accounts; it acknowledged that the intentions might not have been malicious.

Part of a broader thematic overview; only this AI event has been covered.

Wired — AI section Original source ↗
AI agents only one source so far

DoorDash has announced an AI agent for ordering through Apple Messages and opened a waitlist in the US

DoorDash has announced an AI agent for ordering food through Apple Messages. According to the company, it is meant to recognize a usual order, recommend local businesses, and take into account a group's dietary preferences. The waitlist is open to users in the US.

Part of a broader thematic overview; only this AI event has been covered.

Security only one source so far

Google has found a way to label AI-designed proteins

According to the headline, Google has found a way to watermark AI-designed proteins. The excerpt describes the risk that current DNA screening checks fail to detect dangerous proteins designed using AI; it does not state the labeling method or results.

Ars Technica (AI) Original source ↗
AI agents only one source so far

AWS describes deploying a music workflow with three agents using the Amazon Bedrock AgentCore Runtime Instances service

AWS's guide demonstrates a music workflow with three agents: one generates audio on a GPU, another edits it, and a third reviews it. They share files on a single instance; according to the company, the Amazon Bedrock AgentCore Runtime Instances service supports sessions of up to 14 days.

AWS Machine Learning Blog Details and sources →
Security only one source so far

Google DeepMind introduces SynthID Bio watermarking tool

Google DeepMind has introduced SynthID Bio, a watermarking tool that embeds a verification mark into biological designs generated by AI. It allows DNA synthesis screening providers to verify that a sequence originates from a trusted model. It reduces manual screening and helps flag AI-generated records in…

Google DeepMind Blog Original source ↗