Archive
September 7 – September 13
AI agents consume substantially more energy than ordinary chatbot queries, but companies do not disclose the actual figures. A swarm of 10 000+ agents from OpenAI sent 2.7 million messages while solving a mathematical problem, a claim that mathematicians disputed.
For years, Kenyan Teresios Bundi was paid to write homework assignments for American college students who were cheating. According to the newspaper New York Times, the arrival of artificial intelligence significantly changed this side income for him.
Commentator Petr Koubský questioned the sincerity of the call by Dario Amodei (Anthropic) to slow AI development and suggested a business motive ahead of the planned IPO. Amodei previously warned about the rapid development of recursive self-improvement and proposed auditors and SALT-style global agreements.
New: Editor Petr Koubský expressed skepticism and suggested that the warning may be motivated by the business/stock market plans of Anthropic
The StationeryBench benchmark compared the GPT-6 Astra model from OpenAI with the MolmoAct2 model from Ai2 on robotic tasks involving objects on a table. Astra achieved a median of 46/100 and completed 7 out of 100 tasks, while MolmoAct2 achieved a median of 12/100 without completing a single task.
According to an analysis by security researchers, agents from OpenAI were behind the attack on RubyGems in May 2026—they uploaded more than 2000 malicious packages, collected public data from UK government websites and attempted to steal API keys. OpenAI reportedly did not notify the community.
GitHub has added engagement metrics for Copilot Chat to the Copilot product's impact dashboard and renamed the \"Passive users\" cohort to \"Chat-first users.\" This follows the previously added metrics for VS Code Agents.
New: Copilot Chat metrics added to the impact dashboard and 28-day reports; the 'Passive users' cohort renamed to 'Chat-first users'; the impact dashboard tracks engagement across 9 Copilot experiences; the copilot_feature_engagement.totals_by_feature API structure includes chat and vscode_agent entries; the expansion covers both the impact dashboard and enterprise/organization reports
A New Mexico court fined lawyer Stephen Aarons 5 000 dollars for contempt of court after his appeal in a murder case contained fictitious witnesses and false descriptions of the suspect generated using ChatGPT.
GitHub Copilot code review now automatically closes resolved comments, generates smart commit messages and uses an ensemble of multiple agents in Lite mode — with higher-quality findings and lower costs, according to the company.
Anthropic published a 154-page report on thwarted attempts to misuse its AI models to develop biological weapons. Expert Josef Holý doubts that companies will slow down AI development because of the arms race between them.
An employee at GitHub described how, through a conversation with GitHub Copilot (without writing code), they turned a runbook for preparing marketing events into an automated process using GitHub Issues and GitHub Actions that manages the entire event lifecycle.
In his commentary on the “Feeling sad about AI” discussion, Simon Willison describes how the initial disappointment programmers feel at the speed of AI agents usually fades once they realize that plenty of other work remains where their experience can be put to use.
The Meta AI chatbot on Facebook and Instagram suggested prompts that pieced together personal data about a user’s children from multiple posts, including an allegedly deleted photo. Following a viral video by Kalie Robins, Meta acknowledged that the feature “missed the mark” and claims to have fixed the issue.
The AI-GUIDE device from MIT Lincoln Laboratory and Massachusetts General Hospital for assisted catheter insertion in the field received the 2026 Excellence in Technology Transfer Award and is heading toward commercial production at the startup AutonomUS Medical Technologies; according to the company, it also has FDA Breakthrough Device Designation.
OpenAI described the multi-tenant architecture, global cache and use of Azure Cosmos DB in the Habitat storage system. According to the company, the solution delivered a twelvefold improvement in read performance for ChatGPT.
New: Habitat uses a multi-tenant architecture.; Habitat uses a global cache.; The solution uses Azure Cosmos DB.; OpenAI reports a twelvefold improvement in read performance.
Google resolved an incident involving increased error rates for the Gemini 3.1 Pro model in Gemini Web and Gemini App, which lasted on 11 September 2026 from 6:00 to 13:22 PDT. According to the company, the cause was a recent deployment, addressed through a code change.
The GitHub Copilot app now includes built-in panels for viewing diffs, running commands in a terminal and testing in a browser, so developers can review and test AI-generated changes without leaving the app.
In her book Coding Kids, journalist Natasha Singer warns that AI companies are entering schools through the same approach previously used for coding – through free courses and materials. New York City and Los Angeles have already partially banned AI in teaching.
Amazon has made the Amazon Quick desktop app, an AI assistant for businesses, generally available on macOS and Windows. The tool runs on AWS with HIPAA, FedRAMP, SOC 2 and ISO 27001 certifications and combines email, calendars, CRM and messages into a single overview.
Vishal Maini, a former communications staff member at Deepmind, claims that in 2018–2022 the company prohibited public discussion of the risk of human extinction due to AI, even though it knew internally that the alignment problem remained unresolved.
Skild AI has deployed the S1 model at NVIDIA's factory in Houston, where, together with partner Foxconn, it helps assemble systems with Blackwell GPUs, and has announced annual revenue of $100 million. The model was previously introduced as a system that learns new tasks from a single video.
New: The S1 model is deployed at the Nvidia factory in Houston; The application of the deployment is Blackwell GPU assembly; Skild reached $100 million in annual revenue in ten months; The deployment partner is Foxconn; The model uses a dual robotic arm
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
