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 ↓

Tuesday, October 20

24 items
Business and investment only one source so far

Enterprise AI costs depend on capacity utilization and the cost of a usable result, according to a commentary

The commentary recommends evaluating enterprise AI by the total cost of a usable result. It compares in-house infrastructure, pay-as-you-go services and private platforms; it makes the cost advantage of in-house GPUs conditional on sufficient demand and efficient operations.

MIT Technology Review — AI section Original source ↗
Regulation and law confirmed by 2 independent sources

US federal civilian agencies must remove technology from Deepseek

The US Department of Homeland Security has banned federal civilian agencies from using and acquiring technology from Deepseek. They must also remove existing products. The ban also covers affiliated companies and some third-party services.

The Decoder (daily AI news) · and 1 more source Details and sources →
Robotics only one source so far

Eka Robotics raised 150 million dollars for robotic handling

Eka Robotics raised 150 million dollars at a post-money valuation of 1.45 billion dollars. Its architecture combines a force-sensitive gripper with a vision-force-action model. According to the company, the robot achieves a grasping success rate of over 95 %.

TechCrunch AI Original source ↗
Regulation and law only one source so far

Anthropic applies AI content labeling under the AI Act worldwide, according to the article

According to the article, Anthropic labels AI content worldwide because of AI Act requirements. This is reportedly because labeling cannot reliably be restricted by region. The available excerpt does not state when this practice was introduced.

Part of an overview of multiple AI topics; only this event has been covered.

The Conversation — Artificial Intelligence Original source ↗
Security only one source so far

Army Cyber Command is developing autonomous AI agents for cyberattacks

According to documents obtained by The Intercept, Army Cyber Command is developing the ECHOTRIBBLE platform for AI agents that would spread autonomously through networks and disable targets. The project is at an early stage; the experts quoted warn of collateral damage that would be difficult to control.

The Decoder (daily AI news) Original source ↗
AI agents only one source so far

Saint-Gobain used Microsoft Discovery to research cement admixtures

Saint-Gobain says it used the Microsoft Discovery platform to create a knowledge base from 62 000 patents and 44 000 abstracts in a few hours. It estimated that manual preparation would take a year; the team developed hypotheses about new admixture components in approximately six weeks.

Microsoft AI blog (Microsoft Source — AI topic) Details and sources →
Productivity only one source so far

Smartsheet assesses benefits for employees as it rolls out AI

According to Markus McKay-Fleisch, Smartsheet has shifted its assessment of AI adoption from frequency of use to benefits for work and company goals. He says unauthorized AI use signals a lack of tools; making those tools available is subject to approval through IT.

AI agents only one source so far

Yelp releases Smart Lead Reply to respond to customer inquiries

Yelp is releasing Smart Lead Reply for 30 dollars per month. The AI agent is intended to respond to customers and conduct conversations to qualify leads. Business owners can configure how it should conduct those conversations.

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

Hardware only one source so far

Samsung expands memory production in Pyeongtaek to meet AI demand

Samsung is expanding its complex in Pyeongtaek: three factories are operating, and a fourth is ramping up operations. Construction of a fifth is continuing after a pause in 2024, driven by growing demand for memory for AI. The company expects demand to grow until 2028.

MIT Technology Review — AI section Original source ↗
AI agents only one source so far

Terminal X achieves 88.6% agreement with the judgment of experienced investors, according to the source

According to the source, Terminal X achieves 88.6% agreement with the judgment of experienced investors. Its Eval Agent, using the GPT-6.1 Sol model, evaluates outputs against 54 criteria and detects 25 % more deviations from the fund’s mandate; the available excerpt provides neither the methodology nor the basis for comparison.

Regulation and law only one source so far

Ed Husic calls for slower construction of AI data centers in Australia

Former Australian industry minister Ed Husic criticizes the pace of data center construction and the water and energy demands of these facilities. He says it is unclear whether the planned mandatory standards will also cover centers currently under construction.

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

The Conversation — Artificial Intelligence Original source ↗