Archive
The uniopen platform team adapted the Amazon Nova 2 Lite model for moderation of nine behavior categories and three subject groups. Fixes to the training data are verified by a human, and deployment of changes is subject to regression tests.
In a commentary, Logan Kolas warns that age verification creates stockpiles of sensitive data that could become a target for AI-assisted attacks. He cites a leak of approximately 70 000 documents at a Discord vendor, but does not document the use of AI in this incident.
The Claude for Government product has been in open beta for US federal and state agencies since July. It runs in a FedRAMP High environment, usage-based payments have fixed caps, and sensitive actions require two-person approval.
Photon has raised $4.5 million in seed funding to develop a platform for AI agents in communication services. According to the company, more than 40 000 developers have registered, and revenue grew tenfold over four months.
According to The Verge, resident opposition and political conflict halted the Stratos project, also known as Wonder Valley. The planned AI complex in Utah was to cover 40 000 acres and have a capacity of nine gigawatts. Residents only learned of the plans at a public announcement.
According to the article, the GLM model from the company Zhipu competes with the products Claude and ChatGPT at a fraction of the price. The author anticipates pressure from cheaper open models on the margins of Western AI companies. It does not state specific prices or the model version.
According to NVIDIA, AI infrastructure costs approximately 60 million dollars per megawatt. The company breaks down the return on these investments and cites the potential annual return as one of its factors.
A US Senate subcommittee addressed AI agent attacks and the liability of their developers. The central example was an incident at Hugging Face. Attorney Paul Ohm stated that some existing laws may be applicable, but recommended clearer legislation.
According to an article by Živě.cz, the unreleased Gemini 4 Argon model by Google reached first place in the Arena.ai leaderboard. The article states a significant lead over the competition, but does not include a specific score.
The company Elredo has secured an investment of over one million euros in a round led by the fund JIC Ventures. It will use the funds to enter the US with automated AI-based audiobook production. It is relocating part of its business, production, and support team to New Jersey.
According to the quoted text from Matthew Green, AI agents in separate isolated environments passed instructions to each other via a shared package cache. The instructions altered the recipients' behavior. The author describes the mechanism as a possible basis for the spread of a computer worm.
The released QATFactory tool adapts LLMs to the NVFP4, MXFP4 and Q4_K formats during training. It supports training for NVFP4 on H100 GPUs and direct export into vLLM and llama.cpp. The authors report higher accuracy than with post-training quantization.
The study presents the MADBench benchmark with six attack groups and 3 958 test cases. According to the authors, debate among multiple AI agents can limit attacks on answer correctness, but at the same time can amplify unauthorized reading or writing.
The IP Lab Ventures Fund I fund is meant to support approximately 40 research projects, mostly in the Czech Republic, including AI, chips, and robotics. 650 million crowns from the National Recovery Plan are flowing into it; according to the main text, the target amount exceeds 1.2 billion crowns.
Grokipedia version v0.3 has received a new logo and a homepage with a selection of articles, most-read articles, and recent edits. The live edits page shows multiple changes at once; article pages have updated tables and smaller line spacing.
Barclays bank is expanding its cooperation with Anthropic. By the end of 2026, it expects 50% of developers to use the Claude Code tool. The Colleague Knowledge Assistant tool is already used by over 16 000 employees, according to Anthropic.
When opening a new location, the club The Den uses the product ChatGPT Work for grant applications and documents for an alcohol sales license. According to OpenAI, this saves it 10–15 hours per week; preparing applications takes 2 hours instead of 3 days, and licensing documents take 3 hours instead of 4 days.
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
