Archive
October 5 – October 11
According to OpenAI, developers can define custom annotation controls in the ChatGPT app browser and programmatically trigger suggested prompts. Integration with WebMCP tools allows context from selected page elements to be passed to the agent.
According to AWS, Amazon Quick and Amazon Bedrock Knowledge Bases supplement document filtering based on stored permissions with a check directly with the source system at query time. The approach addresses the risk of outdated permissions between synchronizations.
Liquid AI has released the open weights of d1-3B and d1-omni-600M. The models return decisions in a single pass without generating tokens. They support text with images, and the experimental d1-omni-600M also supports text with audio.
Meta is introducing an LLM system intended to detect seemingly harmless ads that direct users to child sexual abuse material. According to the company, it also checks landing pages and enables destinations that violate its rules to be blocked and action to be taken against accounts.
The Dot feature in Codex may fail to create new threads. OpenAI is investigating the issue; according to the status page, the dots component is experiencing reduced performance.
Starting on the day of the announcement, GitHub is automatically moving existing AI credential checks to a new model. They remain available to GHSP and GHAS customers at no additional charge. According to the company, the model uses the context of the surrounding code to recognize passwords that lack a typical format.
The Agent Lightning v1.0 framework is available as open-source. During training, it uses the same agent harness as during deployment and records model requests and responses through an LLM proxy. The agent continues to control interactions with the environment.
According to Qlik, Qlik Answers retrieves source material from unstructured content using Amazon OpenSearch Service and adds citations. Amazon Bedrock Guardrails then compares generated answers with the sources used.
GitHub has made local sandboxing generally available in GitHub Copilot CLI, the GitHub Copilot app and VS Code editor sessions with Agent Host. The feature, available at no additional charge, restricts access by running tools to system resources according to policies set by the developer or organization.
Starting with version 1.0.94-0, GitHub Copilot CLI discovers supported local models from a running Ollama service using the /model command. Once confirmed, they can be used without restarting. Selecting a local model does not enable offline mode or disable telemetry.
According to AWS, four customer-facing employees built the WealthWise prototype during a six-week program with mentors. The financial tool uses multiple AI agents, Node.js, Python Flask, and Amazon Nova models.
Cornerstone OnDemand deployed Orion AI, which coordinates specialized agents for operational tasks. According to internal measurements, it shortened database diagnosis time from 45 to 10 minutes. The system uses Amazon Bedrock and Strands Agents; a three-person team built it in six months.
The update to Muse from Meta adds native support for iPad tablets and more app integrations, including Granola and Notion. According to the company, the agent can independently plan marketing, client proposals and the preparation of product pages based on a business goal entered by the user.
Google has opened SynthID.com to the public for checking images, video and audio. According to the company, the detector now also recognizes watermarks from all its partners. Checking media on the standalone website does not require using Gemini.
The operator has mitigated errors in loading usage data on platform.claude.com and via Admin API. Work on a full fix continues; data for approximately the last 30 minutes may appear on the portal with a delay. According to the announcement, API requests are not affected.
According to Wired, videos and models can also be uploaded to Playground as reference material. The author created three playable games in a day but described bugs and running out of the allowance for further changes. Access is currently limited to adult users in the USA.
New: Videos can also be uploaded as reference material.; Models can also be uploaded as reference material.; The author says he created three playable and stable games in one day.
Hugging Face announced ongoing maintenance of the H200 infrastructure for the Inference Endpoints service, with completion expected around 12:30 UTC. The creation of new H200 instances may be temporarily limited.
Radisson Hotel Group created a plugin for ChatGPT in collaboration with Accenture. According to the source, it helps travelers search for, compare and book hotels while planning trips.
OpenAI released a fixed version of the desktop app for Linux, macOS and Windows. It addresses a problem with creating new ChatGPT Work threads. The incident is marked as resolved; affected users should update the app.
Pavel Stupka created the Tandem Commander file manager from the Open Salamander 5.0 project. An early version is available for Windows, and the source code is licensed under GNU GPL v2. According to the author, development involves documented specifications, plans, test designs and human approval.
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
