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
1135 items in the selected feed
How to read evidence statuses ↓
Ode with Anthropic acquired Casper Studios, a company specializing in integrating Claude models into business processes. This is the first acquisition by Ode with Anthropic since its founding in May 2026; the financial terms were not disclosed.
IBM announced the interconnection and cooling of modular cryogenic cells for the fault-tolerant quantum computer Starling. The modular units can be connected side by side and operate below 15 millikelvin. It plans, by 2027, to connect multiple chips to reach at least 1000 programmable qubits and by 2029 to build…
Part of an overview of multiple AI topics; only this event has been covered.
AI companies buy older physical books, scan them to train models and then destroy the original. Unlike piracy, destroying a legally purchased book in this way is permissible according to the source.
Guidelight AI Standards evaluated five leading AI labs (OpenAI, Anthropic, Google, Meta, xAI) based on publicly available plans for containing a model that escapes control. OpenAI ranked highest, Anthropic and Meta lowest; most companies do not have a published, specific intervention plan.
Simon Willison argues that a key skill when working with coding agents is being able to give them clear instructions and then verify the correctness of the changes made – and that reading every line of code is not the most effective way to do this.
Scientist Matyáš Boháček from Stanford described his AI models for deepfake detection and sign language translation in an interview for Radio Wave. According to him, people around the world use these models. He said that AI helps science but carries risks.
RayNeo iO glasses without a camera display text in the field of view (waveguide, 97% transparency). Four microphones are used for listening, summarizing and extracting tasks from conversations. Default suite: RayNeo AI and Gemini 3.1 Flash Lite. For $9.99/month, access to ChatGPT 5.1, Gemini 2.5 Pro, Claude, DeepSeek.
A study by Villanova University found that readers cannot distinguish shorter AI-written texts from human-written ones and often rate them as more readable. Following the withdrawal of the novel by Jerry Falade (Macmillan) and the cancellation of the book by Mia Ballard (Hachette Book), publishers are introducing Human Authored certifications.
Over a million users have clicked the “Seems like AI slop” button on the social network LinkedIn since its launch on 30 July 2026. According to the company LinkedIn, the visibility of content classified as AI slop fell by 40 percent over a few weeks.
Anthropic has deployed Claude Mythos 5 in Claude Security, which scans code for vulnerabilities and suggests fixes approved by a human. The beta is for Enterprise customers, and security partners serving hospitals, utilities and banks are also adopting the model.
AWS introduced the Agentic Data Operations Platform (ADOP) reference architecture, which uses AI agents in Claude Code on Amazon Bedrock to automate ETL, data quality checks and compliance across the Bronze-Silver-Gold stages.
GitHub has launched a public preview of an integration that brings GitHub Copilot agent mode to Slack. The team assigns a task by mentioning @GitHub, the agent works asynchronously in a dedicated code channel, and the PR can be made subject to approval before merging.
GitHub Copilot can now be launched as a cloud agent session directly from Microsoft Teams by mentioning @GitHub. A dedicated code channel is created for tracking progress, and administrators can require additional approval for pull requests. Public preview, paid plans, consumes AI credits.
Developer Matt Webb described how ChatGPT, acting as an interactive tutor, helped him understand quaternions for the AR mode in the app Galactic Compass 2 - something he had been unable to achieve through books or consultations with mathematicians.
A university lecturer argues that it is critical to teach students to distinguish between the capabilities of AI and its limitations. As an example, they cite an exercise in which the product ChatGPT failed to analyze a literary scene — it produced sterile, abstract assessments rather than a nuanced understanding of complex emotional and…
Every trained an AI agent for copy-editing on 30 000 historical edits by its editor-in-chief Kate Lee and tested it retrospectively against her older texts — according to CEO Dan Shipper, the goal is to distribute one person's expertise throughout the company.
OpenAI has launched a plugin connecting Apple Messages to ChatGPT – it allows messages to be sorted, deleted, searched and sent on behalf of the user. According to the company, it runs locally without indexing messages, but the company recommends reviewing them before sending.
Google is launching a chatbot interface in the Google app for describing Discover feed preferences in natural language, alongside personalized audio briefings in Google News and a Preferred Sources button for publisher websites.
Meta has released a beta version of the Meta AI app for Mac, free for users with a Meta account, aimed at small business owners and content creators, with connections to Google Workspace and Meta Ads; the company is also testing a paid subscription for more demanding features.
Amazon Bedrock AgentCore has expanded Policy Authoring, a tool that converts written rules in natural language into formal Dogwood policies for governing AI agent behavior – including time constraints, tool call order, and integration with Amazon Bedrock Guardrails.
Amazon published the second installment of a blog series on scaling agentic AI systems, describing architectural principles for managing heterogeneous environments with multiple frameworks, models, and providers without creating vendor lock-in.
Following a wave of executive departures from April 2026 (including Fidji Simo, Brad Lightcap and Denise Dresser), OpenAI president Greg Brockman took over leadership of product and the entire commercial division, effectively becoming the second most powerful person at the company after Sam Altman, just before the planned IPO.
A commentary in MIT Technology Review argues that rhetoric about the consciousness and autonomy of AI systems (J-space from Anthropic, statements by Sam Altman about the singularity, calls for AI rights) in practice helps companies avoid liability for harm caused by their AI agents.
Generalist AI introduced GEN-1.5, a model that, according to the company, learns a new task from a single 3 to 12-second demonstration without further training. Without training, it achieves a success rate of 59 %, rising to 83 % with ten training steps on five minutes of data. The results have not been independently verified.
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
