OpenAI introduces Control Room for monitoring and controlling AI workloads
OpenAI announced Control Room, a tool for monitoring model workloads within its infrastructure. It allows workloads to be stopped in case of safety concerns.
AI Flash · live feed
A timeline of everything that has passed through the radar — newest first, with an evidence status for each item.
Continuously updated
2526 items in the selected feed
How to read evidence statuses ↓
OpenAI announced Control Room, a tool for monitoring model workloads within its infrastructure. It allows workloads to be stopped in case of safety concerns.
Specialty chemicals manufacturer Lubrizol states that it is building a digital platform with AI modeling and cloud infrastructure for R&D in order to speed up the testing of new formulas and the development of more sustainable materials.
Google has released two new eighth-generation TPU variants: TPU 8tx for encrypting data during processing, and TPU 8i Max with doubled performance and memory thanks to combining two chips into a single package. These are configurations of an existing chip, aimed at expanding the offering for external customers in competition with Nvidia.
Meta is expanding worldwide a technology that uses AI to detect teenagers even when a false (adult) date of birth has been entered, placing them into restricted Teen Account protection on Instagram, and later also on Facebook and Threads.
Following the Shy Girl scandal, in which the detection tool Pangram flagged 78% of the text as AI-generated, developers are offering software such as Workings, Blindproof, Wintertext, Human Ink or ObraVera that tracks the writing process as evidence of human authorship.
The company Kuken Industries deployed search via the OpenAI API over a 70-year-old archive of cooling tower drawings. The time needed to find a drawing for a repair dropped from nearly an hour to 4.5 minutes.
Hacker Chema Alonso at the Check Point conference in Prague demonstrated how a cleverly phrased question, without technical knowledge, makes a corporate chatbot act outside its assigned topic and bypass its restrictions.
NVIDIA states that it supports the development of physical AI in Korea – the deployment of accelerated computing and simulations in semiconductor manufacturing, robotics, and autonomous vehicles – with a mention of Samsung as one of the players involved.
Part of an overview of multiple AI topics; only this event has been covered.
Hyundai Motor Group is exploring the use of NemoClaw blueprints and the OpenShell runtime by NVIDIA, while AIM Intelligence is integrating its Starfort platform with OpenShell. The goal is to secure AI agents for enterprise operations.
A research study shows that inserting fake text mimicking the model's own reasoning (thought injection) achieves up to a 50% success rate in attacks such as reverse shells or backdoors in a repository against agentic LRMs.
Engineers from MIT and NYU have developed FLEX, a platform for training robots to understand the hidden intent of natural language commands. It uses an LLM for deduction: the robot understands "don't step on the laptop" as "avoid all electronics." The study will be presented at the Conference on Robot Learning.
NVIDIA, Rebellions, and SK Telecom announce a collaboration on sovereign AI infrastructure in Korea using NVIDIA NVLink Fusion. The initiative combines NVIDIA's open infrastructure, Korean-designed AI processors from Rebellions, and SK Telecom's data center and AI services.
Simon Willison had the Claude Opus 5.5 Medium model port his 3D web application from the Three.js library to Vanilla JavaScript with WebGL. According to him it succeeded, the result is functional, although the details differ from the original.
Simon Willison recommends hiding detailed AI-generated pull request descriptions behind a details/summary HTML element and putting only a short human summary at the top, so PRs don't come across as low-quality AI text.
Anthropic's documentation describes that thinking blocks from the Claude Fable 5.2 and Claude Mythos 5.2 models are silently discarded when switching to another model within a conversation, without the API returning an error. In the opposite direction — when switching to these two models — reasoning from other models is preserved.
Anthropic has established that the Claude Fable 5.1 and Claude Mythos 5.1 models require 30-day data retention by default; zero retention mode is only available after explicit approval from Anthropic.
Anthropic has documented that in the beta advisor tool feature in Claude Platform, an executor with the Claude Fable 5.2 or Claude Mythos 5.2 model accepts only one of these two models as the advisor. Any other combination returns an HTTP 400 error.
According to Bloomberg, Apple CEO John Ternus is pushing to speed up the development of Siri AI so it can compete with the Dots, Muse, and Grok agents. Fine-grained control of Apple's own apps arrives with iOS 27.3, while integration of third-party apps (Grubhub, OpenTable) is planned for spring 2027.
Rights holders to Elvis Presley's music and the publisher of Eminem are suing Suno over alleged illegal imitation of artists through AI music generation and violation of Tennessee's ELVIS Act against voice cloning.
Andrej Karpathy built the HackerTube website over a weekend using the Claude Opus 5.5 model, which filtered 43 million Hacker News comments into 6454 videos, and the TristanTube website with 47 000 videos, which the Claude Code tool "one-shotted" in three minutes.
New York Times Cooking launched an iOS beta version of an AI tool on October 23, 2025 that adjusts recipes based on available ingredients, cookware, or diets. Reader reactions were largely negative, while the Wired author rated the feature positively.
Nearly half of users discuss personal relationships and problems with AI chatbots, and a tenth trust them more than family. Experts warn of data leakage and the reinforcement of dangerous delusions.
World models – technology generating 3D simulations of the physical world for training robots and autonomous vehicles – are attracting major investment. Niantic has reoriented itself toward this field after selling Pokémon Go, AMD acquired World Labs for $8+ billion, and Yann LeCun raised a billion dollars for his own startup.
According to The New York Times, US authorities approved the company New Matrix to receive Nvidia H200 chips in China, even though its headquarters is an empty office and its owners have ties to a sanctioned Chinese private equity firm.
AI Radar monitors Czech and international sources every day, looking for changes that truly deserve attention.
The result is a quick overview of what has actually changed in the AI world, rather than another stream of articles.
Use the CS/EN switch to read the same Radar in Czech or English. English content is published after its translation has been checked, so new and older items may appear later.
Each row represents one event — not one article. At a glance, you can see its significance, credibility and main point.
Illustrative example, not a current news item.
Agent mode is available to all paying users
Until now, the mode was available only on the highest plan; it is now available on all paid tiers without a waitlist.
The detail page contains a fuller summary, its significance and original sources. Practical impact appears in the detail and the For individuals and For businesses views. An AI Flash item reaches the main selection only after it has been expanded and meets the publication rules.
We first summarize each event in the same way for everyone. Based on those same facts, we then explain what the change means for your own use and what it could mean for how a company operates.
Today’s briefing is the same for everyone. Pages For individuals and For businesses can be found in the main navigation — they select only events relevant to the given use case.
Reset history · Outage history · Subscribe in a feed reader (RSS) · Methodology and label explanations