LFM2.5-24B-A2B builds on LFM2-24B-A2B, expands the context from 32K to 131K tokens and always produces a reasoning trace before answering. According to the authors, it achieves 184 tokens/s on Apple M5 Max. The available excerpt does not confirm a public release.
Buttons for expressing interest or lack of interest may also appear below posts on Facebook. Meta may then ask follow-up questions that users answer in their own words. This is part of changes to recommendations in Feed and Reels.
Ask Play in the Play Store now lets users search for apps and games by asking questions. Recommendations may include detailed comparison tables. According to the article, the expansion is available from the day the article was published.
Part of an overview of multiple AI topics; only this event has been covered.
Google has enabled users to run personal apps created using Google AI Studio through Google Play without connecting their phone to a computer. Initial access is available to users in the USA aged 18 and older; according to Google, the apps remain outside the public catalog.
According to the article, an AI-generated image of Mia Goth circulated on Instagram from the account miagoth_, which had a blue verification badge. It received hundreds of thousands of “likes”, and both OK! and People covered it as authentic.
GitHub now detects anthropic_service_api_key and anthropic_user_api_key. Both types are included in push protection by default: in repositories with secret scanning enabled, it blocks pushes containing these keys, including in free public repositories.
Discovery AI on Amazon Bedrock converts questions into security queries in Sonar. According to AWS, the deployment reduced query costs by 90 % and response times from approximately 60 to 15 seconds compared with the original prototype.
Google DeepMind presents demonstrations of the Gemini 4 Argon model for controlling robots, including a simulation of a trash can in MuJoCo and verification of movements using cameras. According to the company, the model achieves 87 % on ERQA; the available excerpt does not specify access to the model.
The Deckhand agent at ShipStation Global answers IT questions in Slack, reads screenshots of error messages, and manages tickets in Jira Service Management. According to the company, it resolved 97.5 % of queries in the support channel and received 96 % positive ratings from employees.
AWS demonstrated how to build an agent with Grok 4.7 on Amazon Bedrock. According to its test, the agent succeeded in 67 of 72 runs; the low setting consumed approximately one-eighth as many output tokens as xhigh, but had a lower success rate.
Administrators can require current policies to be loaded when GitHub Copilot starts. With forceRemoteSettingsRefresh enabled, the client does not continue unless it confirms that the policies are current. The managed sandbox setting now also enforces a minimum client version.
The AWS blog describes an optimization of fused MoE for the Amazon SageMaker AI service. According to the company, transposing matrices is intended to improve utilization of GPUs based on the NVIDIA Hopper architecture during generation, when small token counts do not fill the fixed dimension of WGMMA instructions.
In five US surveys from 2026, 55–77 % of adults described the pace of AI development or use as too fast or supported slowing or halting development. The questions differed, so the results are not directly comparable.
The article explains the difference between getting the correct answer from AI and learning. It recommends trying on your own first, then asking for a hint or an analysis of your mistake. It mentions studies on the use of ChatGPT, but the available excerpt does not include the results of the last experiment described.
SpaceXSI is introducing The X Subscription with X Premium, Grok and Cursor. It offers a free plan and paid tiers at 8, 30, 100 and 200 dollars per month. The latest Grok models are available on the Premium plan and above; the option for existing customers to switch is being rolled out gradually.
Geoffrey Huntley describes a pre-commit hook that checks git diff and blocks the commit when it finds a direct call to observability modules outside the diagnostic library. The AI agent then receives an instruction to use the required library.
In his commentary, Konstantinos Komaitis proposes building independent capabilities for AI evaluation and connecting them through shared standards. He recommends giving further thought to a possible new international institution; he cites the distributed governance of the internet as inspiration.
Philosopher Henry Shevlin joined Google DeepMind in May and devotes 80 % of his working time to the company. In an interview with IEEE Spectrum, he explains his move from academia and discusses people's relationship with AI and questions about whether it could be conscious.
American Express has introduced Intent Intelligence Framework in ACE Developer Toolkit. The customer is to verify their identity and confirm the purchase instructions before payment. The agent then uses a digital payment credential instead of the actual card number.
Apple has opened preorders for the HomeView display at 299 dollars. The Siri AI assistant is intended to control the home and use personal context, which requires a compatible iPhone on the same network and facial recognition. Shipping will begin on 13 November; the Czech Republic is not among the initial markets.
Meta is now emphasizing the specialist knowledge of Muse Wonder and Muse Wonder Full for work tasks. It has yet to integrate them into Muse. Both models are available to Muse Code subscribers; only Muse Wonder is available through the API.
GitHub has made AI Scan metrics available in public preview: finding counts, their resolution, and the use of suggested automatic fixes. The feature is available on github.com to customers with GitHub Code Security or GitHub Advanced Security; GitHub Enterprise Server is not supported.
According to the article, Higgsfield uses the GPT-6 Astra model to let creators make videos longer than ten minutes instead of one minute. The available excerpt does not specify the conditions for access.
The Ministry of Health is preparing guidance for the public and professionals on using AI for mental health difficulties. The project is expected to run until 2028. According to research by PAQ Research, 13.4 % of users postponed professional help or did not seek it because of AI.
AI Radar monitors Czech and international sources every day, looking for changes that truly deserve attention.
MonitorsOfficial AI company blogs, specialist media, and research sources.
Selects and combinesFilters out information noise and combines articles about the same change into a single event.
Summarizes and explainsExplains significant events in English: what happened, why it matters and where the information comes from.
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.
Everything you need to navigate the AI world
Today’s briefingAn overview of news from the past 24 hours and a main story when available. Beside them is Live · AI Flash, followed by research and links to other Radar sections. On mobile, the blocks appear one below another.
AI FlashAn ongoing feed of brief updates with an evidence status. Links lead to a Radar detail page when one is ready, otherwise to the original source. You can also find reset and outage histories here.
Practical applicationsWhat new tools and features can do, what you can try and what their actual impact could be.
Model selectionModel comparison by type of work, capabilities, price and speed.
Research and archiveA separate research overview, topic search and older events by date.
One event, everything that matters
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.
Importance: ▮▮▮ majorOpenAIModels✓ 6
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.
▮▮▮ major · ▮▮ important · ▮ we're tracking = how significant the change is✓ 6 = six independent publishers, not the number of articles or feeds✓ official = a clear release, law or incident is substantiated by the relevant authority1 source = no independent confirmation yetbold = who is behind the changegray text = a brief summary of what happened
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.
The same news, two practical uses
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.
For individualsWhat you can use or try, how the change can help you at work and what to watch out for.
For businessesWhat impact the change could have on processes, costs, risks and other business decisions.
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.