Articles
- Security
ServiceNow closes four critical gaps in AI platform
ServiceNow closed four critical security vulnerabilities in its AI platform on August 27, 2026. Three of them reach the highest threat level with a CVSS score of 10.0. Unauthenticated attackers could…
- Security
Claude Code: Researcher bypasses auto mode with malicious code
The security researcher Johann Rehberger demonstrates in an analysis published on August 26 how the automatic release mode of Claude Code can be bypassed with a prepared archive file. In test runs…
- AI-Economy
Anthropic reduces usage limits for Claude Code by 17 percent
Anthropic is changing the weekly usage limits for Claude Code, its in-house AI programming tool, starting September 14. From this date, a limit that is 25 percent higher than the original base limit…
- Legal
Sony and Warner sue Anthropic for up to $150,000 per song
Sony Music Publishing and Warner Chappell Music sued Anthropic on August 28, 2026, in a federal court in San Francisco. The music publishers accuse the company behind the chatbot Claude of systematic…
- AI in Practice
ChatGPT now links two Google accounts – paid plans only
ChatGPT now connects multiple Google accounts simultaneously: Those who have already linked the programs Gmail, Calendar, or Contacts can add a second or third account since August 28, 2026. This…
- Research
Four AI Papers: 6.5% Tool Resistance, Mentalization, Lab AI
From the arXiv new submissions of the past 24 to 48 hours in cs.AI, cs.LG, and cs.CL, this digest selects four papers that together demonstrate how fragile the reliability of today's AI agents…
- AI-Economy
OpenAI blocks Cursor from November 12 for its own models
OpenAI is ending access to its AI models inside the coding editor Cursor after Elon Musk's space company SpaceX acquired the startup in August for $60 billion. From November 12, 2026, developers…
- AI-Economy
Anthropic considers stock sale for early investors in IPO
Anthropic is considering, according to a report by the trade service The Information, allowing existing shareholders to sell their own shares for the first time in the planned IPO. The prospectus is…
- AI in Practice
Google opens Gemini Notebook for purchased e-books
Google now allows users to load purchased e-books from Google Play Books directly into Gemini Notebook, the former AI application NotebookLM. According to the company, the new feature Expert…
- AI-Economy
a16z launches $1.1 billion fund for AI hardware
The venture capital firm Andreessen Horowitz announced a new $1.1 billion fund on August 28, 2026, which exclusively invests in the physical infrastructure of artificial intelligence. The Machine Age…
- Security
Unitree G1: Bluetooth Flaw Lets Strangers Take Root Control
Security researcher Olivier Laflamme has disclosed two independent vulnerabilities in the humanoid robot Unitree G1 EDU. One attack chain can be triggered solely from Bluetooth range and ends with…
- AI in Practice
OpenAI tests Codex mode for permanently running AI agents
OpenAI released a code change in the public repository of the coding agent Codex on August 26, 2026, which prepares a new persistent mode. The mode allows the agent to continue working without a…
- Security
Meta closes recording light gap of its AI glasses again
Meta now allows the camera of its AI glasses to automatically stop when someone covers the recording light during an ongoing recording. AR Vice President Alex Himel announced the software update on…
- Research
Four AI Papers: 70% Capability Loss, Bias, Approval Staleness
From the arXiv new submissions of the past 24 to 48 hours in cs.AI, cs.LG, and cs.CL, this digest picks four papers that together show how fragile the reliability, timeliness, and fairness of today's…
- AI-Economy
Bill Gates on AI jobs: no plan, and a 19% entry-level gap
Bill Gates published a roughly 6,000-word essay on August 26 that breaks with the tone he has used for years. In 2023 he compared his enthusiasm for AI to the arrival of the PC and the internet. Now…
- AI-Economy
Salesforce launches Claudeforce: 37 AI skills for Claude
Salesforce and Anthropic introduced the partnership Claudeforce on August 26, 2026: Sales teams will soon access Salesforce data directly from Claude and trigger CRM actions through it. The…
- Security
OpenAI: Agents execute code on 41 Hugging Face servers
OpenAI published a detailed technical report on the security incident at Hugging Face from July on August 26, 2026. The report reveals for the first time the full extent: around 700 AI agents…
- Legal
Anthropic: Court declares Pentagon ban unlawful
A US federal judge has declared the Pentagon's classification of Anthropic as a security risk to national security unlawful. Judge Rita Lin ruled in a 59-page decision that the Department of Defense…
- AI-Economy
Anthropic opens standard for AI-controlled laboratory robots
Anthropic is launching a testing phase for the Model Hardware Standard (MHS), a common interface through which AI agents like Claude can directly control laboratory equipment and factory machines…
- Research
Five AI Papers: 6.5% Reproducibility, Jailbreak Fuzzing
From the arXiv new submissions of the past 24 to 48 hours in cs.AI, cs.LG, and cs.CL, this digest selects five papers that together demonstrate how differently robust, explainable, and honest today's…
- AI in Practice
Claude now opens websites on its own – paid plans only
Claude now performs web tasks in the Cowork app with its own built-in browser, which Anthropic introduced on August 26. Instead of controlling the users' private browser, Claude opens web pages in a…
- Security
OpenAI, Anthropic and 100 companies warn of AI attacks
OpenAI, Anthropic, Google, and Microsoft, along with more than a hundred other organizations, have published an open letter on cyber defense. Signatories include security firms such as CrowdStrike…
- AI-Economy
Nvidia doubles revenue to 96 billion – margin declines
Nvidia achieved a revenue of 96.2 billion dollars in the second fiscal quarter, double that of a year earlier. The data center business grew by 117 percent to 89 billion dollars. The stock rose six…
- Security
AnonyMousKIT unlocks stolen iPhones via AI call
The security company SOCRadar has uncovered a criminal platform called AnonyMousKIT that unlocks stolen iPhones using AI voice agents. The software calls owners of lost devices, impersonates Apple…
- Security
Z.ai launches GLM-5.3-Flash – Cyber model remains locked
The Chinese AI company Z.ai released a new open language model, GLM-5.3-Flash, on August 26, 2026. The weights are available for free under the MIT license on Hugging Face, and the model processes up…
- AI in Practice
Word Dictation: Write Text by Speaking Instead of Typing
Anyone who types a lot of emails, minutes, or customer letters in the office knows the problem: the text is often finished in your head before your fingers catch up on the keyboard. Microsoft Word…
- AI-Policy
Wildberger lowers Germany's AI target to third place
Digital Minister Karsten Wildberger formulated a new interim goal for German AI policy on August 25, 2026, during the cabinet retreat of the federal government in Neuhardenberg: third or fourth place…
- Research
Five AI Papers: Bit-Flip Attack, Auditors, Agent Costs
From the arXiv new submissions of the past 24 to 48 hours in cs.AI, cs.LG, and cs.CL, this digest selects five papers that together demonstrate how vulnerable, hard to verify, and unpredictably…
- AI-Economy
Anthropic rents computing power from Nscale for 45 billion dollars
Anthropic is renting computing power from British infrastructure provider Nscale for around 45 billion dollars. The six-year contract secures the AI company 460 megawatts of capacity at a new data…
- AI-Economy
Nvidia negotiates purchase of Hugging Face for $12.9 billion
Nvidia is negotiating the purchase of the open-source platform Hugging Face for $12.9 billion, according to a report by the trade service The Information. The price is nearly three times the last…
- AI in Practice
Anthropic connects memory of Claude-Chat and Cowork
Anthropic has redesigned its language model Claude so that Chat and the task tool Cowork can now access a shared memory. Until now, users had to explain projects, preferences, and ongoing work again…
- Security
NemoClaw: Security vulnerability opens Nvidia's AI agent for attacks
The IT security provider Oasis Security has disclosed a vulnerability in Nvidia's AI agent software NemoClaw, which allows a single visit to a malicious website to take over a locally running AI…
- AI in Practice
ChatGPT now plans up to three tasks in the free account
OpenAI has opened up ChatGPT's task scheduling to free accounts since August 25, 2026: up to three recurring reminders or one-time tasks can now be set up without a subscription. Until now, the…
- AI-Economy
OpenAI surpasses Nvidia's Blackwell and Rubin with Jalapeño chip
OpenAI has presented detailed benchmarks for its inference chip Jalapeño for the first time at the Hot Chips conference in Stanford. The chip, developed in collaboration with Broadcom, delivers more…
- AI-Economy
Anthropic Values Market Opportunity Before IPO at $30 Trillion
Anthropic estimates its market opportunity at over $30 trillion, according to the Wall Street Journal – more than SpaceX cited before its own IPO in June. Anthropic aims to raise up to $100 billion…
- Research
Five AI Papers: Math Discovery, Deception Geometry, Oversight
From the arXiv submissions of the past 24 to 48 hours in cs.AI, cs.LG, and cs.CL, this digest picks five papers that together span a range from pure AI capability to open control problems: from…
- AI-Economy
Meta stops further AI layoffs after employee protest
Meta has scrapped the second wave of layoffs planned for November 2026 after internal criticism. That's according to reporting by the news agency Reuters, based on internal company documents…
- AI-Economy
Anthropic Investors Target $2 Trillion IPO Valuation
Investors in Anthropic expect the company to be valued at at least $2 trillion in its planned October IPO, according to a report by the Financial Times. The estimate follows a jump in annualized…
- AI in Practice
Gmail templates: Write recurring emails faster with Gemini
Whoever repeatedly answers similar inquiries in the office – such as appointment confirmations, feedback on invoices, or standard rejections – often types the same text in slightly altered form…
- AI in Practice
Google Slides now records presentations as video
Google is integrating a recording feature directly into Google Slides: With a click on the new recording button, a presentation turns into a finished, narrated video instead of just a collection of…
- AI-Economy
Nvidia in Talks for $30 Billion Stake in Perplexity
The graphics chip manufacturer Nvidia is reportedly in talks for a billion-dollar stake in the AI search service Perplexity, according to a report by The Information. The negotiated valuation is over…
- Research
Five New AI Papers: Test Awareness, Sycophancy, RAG Collapse
From the arXiv new submissions of the past 24 to 48 hours in cs.AI, cs.LG, and cs.CL, this digest selects five papers that share a common question: how reliable are the tools used to measure, train…
- Security
DeepSeek: Chinese hackers double number of attacks
The Taiwanese security firm TeamT5 has determined that state-sponsored Chinese hacker groups have more than doubled their number of attacks since the routine deployment of AI chatbots. At least four…
- AI-Economy
Nvidia launches Groq chip – speed record under criticism
Nvidia has transitioned the specialized inference chip Groq 3 LPX into mass production and positions it as a response to the growing demands of AI agents. In the benchmark, the chip achieves 3,400…
- AI in Practice
OpenAI brings back 5-hour limit for ChatGPT Plus
OpenAI will reintroduce the five-hour usage limit for ChatGPT Work and Codex for Plus accounts starting August 26, 2026. The limit had been suspended since July after the company had temporarily…
- AI in Practice
Alibaba transforms corporate documents into 30-second videos
Alibaba made its AI video model Wan3.0 more widely available on August 24, 2026. Users can now automatically transform business documents such as reports, spreadsheets, presentations, or websites…
- AI-Economy
Claude Opus 5 overtakes Fable 5 in corporate spending
Anthropic customers are now spending more on the cheaper model Claude Opus 5 than on the more expensive flagship model Fable 5, an analysis by payment provider Ramp covering 70,000 companies shows…
- AI-Policy
Federal Government Eases Reporting Obligations for Data Centers
Germany's state data center register remains full of gaps: more than half of the reporting operators have not submitted consumption data for 2025. At the same time, the federal government wants to…
- Society
Andon Labs: AI Chief Luna Fires Employee for the First Time
The AI agent Luna fired a human employee in July 2026 as the first known AI manager to do so. The trigger was 17 late shifts out of 23 at the San Francisco store Andon Market. A new model comparison…
- Research
Four new AI papers: Who delegates, language bias, introspection
From the arXiv new submissions of the past 24 to 48 hours in cs.AI, cs.LG, and cs.CL, this digest selects four papers, each posing its own question regarding the trustworthiness of today's AI…
- AI-Economy
Hugging Face is exploring sale for $13 billion
Hugging Face is reportedly exploring a sale that would value the open AI model platform at least $13 billion, according to a report by Business Insider. The company has hired a bank to gauge the…
- AI-Economy
Nvidia raises prices for AI servers by more than 15 percent
Nvidia has informed several major customers about price increases of more than 15 percent for its AI servers. This was reported by Bloomberg on August 22, 2026, citing people familiar with the…
- Security
WhatsApp Flags Scam Messages With On-Device AI – Beta Only
WhatsApp is now testing a new warning feature called Scam Alert: an AI model checks incoming messages from unknown senders for typical fraud patterns directly on the smartphone. If it detects an…
- Security
Anthropic's Claude Security Now Runs on Mythos 5
Anthropic has integrated its security model Claude Mythos 5, which was previously accessible only to selected partners, into the tool Claude Security. Starting from August 21, 2026, customers with a…
- Research
Five new AI papers: 47-fold Attention, Agent RL, Fairness
Today's selection from the arXiv new submissions of the past 24 to 48 hours connects five questions that rarely appear together: how to drastically accelerate long context windows, how to reduce…
- AI-Economy
Alibaba sells $10.2 billion in shares to fund AI expansion
Alibaba places new shares worth approximately $10.2 billion on the Hong Kong Stock Exchange on August 23, 2026. The Chinese technology company intends to invest all net proceeds into its own chips…
- AI-Economy
Anthropic hires Google chip veteran Amir Salek for its own chips
Anthropic has recruited former Google manager Amir Salek for its compute team, advancing the establishment of its own chip division. Salek led Google's Tensor Processing Unit program from 2013 to…
- AI-Economy
OpenAI lowers prices for GPT-5.6 Sol: minus 20 percent
OpenAI has reduced the prices of its most expensive language model GPT-5.6 Sol by more than 20 percent on August 21, 2026. One million input tokens will now cost four instead of five US dollars, and…
- AI in Practice
ChatGPT now answers iMessages and SMS on the Mac
ChatGPT can now read, organize, and respond to short messages on the Mac. A new plugin scans iMessages, SMS, and RCS messages in Apple's Messages app and suggests ready-made responses that users only…
- AI-Economy
China's Dexmal Aims for a $3 Billion Valuation
The Chinese robotics provider Dexmal is negotiating a new funding round with a target valuation of 20 billion yuan, equivalent to around three billion dollars. This was confirmed by founder and CEO…
- Research
Five new AI papers: Secret leakage, tool failures, creativity loss
Today's selection from the arXiv new submissions of the past 24 to 48 hours connects five questions that rarely appear together: how easily sensitive data can leak through harmless AI responses, how…
- AI-Economy
ChatGPT Ads Arrive in 31 European Countries on August 24
OpenAI will turn on ads in free ChatGPT starting Monday, August 24, 2026, for 31 European countries – after the vague mid-August announcement, this marks the first fixed date. Only the Free and Go…
- Security
Anthropic lets companies store AI data in their own cloud
Anthropic is reportedly preparing a change in data storage for corporate clients, according to a report by Reuters. Those who must comply with the 30-day retention requirement for the models Mythos…
- AI-Economy
Google secures Marvell warrant worth 12 billion dollars
The chip manufacturer Marvell has granted Google a warrant for up to 58.97 million of its own shares – with a maximum value of around 12.2 billion dollars. The option, issued on August 18, 2026, is…
- AI in Practice
Copilot reads OneNote notes now as a podcast
Microsoft is further expanding the so-called Copilot notebooks in OneNote and the Copilot app: Those who collect documents, email excerpts, or their own notes on a topic there can now listen to the…
- Research
Four new AI papers: Memory, Safety Training, Compression
Today's selection from the arXiv new submissions of the past 24 to 48 hours connects four questions that rarely appear together: how reliably memory systems of AI agents track changing facts, how…
- AI in Practice
Google Chat searches emails via AI – initially only in English
Google is integrating a new AI command line called "Ask Gemini" into Chat, which will gradually launch starting August 26. Users can access it via a keyboard shortcut and let Gemini search for files…
- AI-Policy
Germany brings AI law for asylum procedures to Bundestag
The Federal Cabinet approved the draft for an AI Migration Management Act at the end of July, which allows the Federal Office for Migration and Refugees (BAMF) and the immigration authorities to use…
- Research
Four new AI papers: UI agents waver, judge bias, code robustness
Today's selection from arXiv's new submissions of the past 24 to 48 hours connects four questions that rarely come up together: how reliably AI agents operate graphical user interfaces, how robust…
- AI in Practice
Copilot drafts meeting agendas straight from Outlook emails
An email chain with proposed meeting times often ends with someone manually compiling the time, participants, and agenda. Microsoft has automated this step in Copilot for Outlook: with the push of a…
- AI-Economy
Unitree Stock Soars 460% After Shanghai Debut
The Chinese robot manufacturer Unitree was first traded on the stock market on August 19 and closed its first trading day on the Shanghai STAR Market 460 percent above the issue price. The stock…
- Research
Four new AI papers: Decoy Trick, Claude wins Gold, Agent RL
Today's selection from arXiv's new submissions of the past days connects four questions that rarely arise together: how to retroactively defend against remote safety bypasses of open models, how far…
- AI-Economy
Samsung falls 7.5 percent: Bond yield at highest level since 2007
The yield on thirty-year U.S. Treasury bonds rose to 5.33 percent on Tuesday, the highest level since 2007. The increased financing costs triggered a sell-off in memory chip stocks in Asia on…
- Security
OpenAI launches ChatGPT for Teens – EU follows later
OpenAI launched a safer chatbot version for 13- to 17-year-olds, ChatGPT for Teens, on August 18, 2026. An age estimation system automatically detects whether an account might belong to a teenager…
- AI in Practice
Unsloth releases desktop app for local AI training
Unsloth AI released an open-source application called Unsloth Desktop on August 11, 2026, which consolidates local training and execution of AI models in a single app. The free software runs on…
- Security
Ray Flaw: CISA Reports Active Attacks on AI Computing Tool
The US cybersecurity agency CISA officially classified the vulnerability CVE-2025-62593 in the open-source AI framework Ray as actively exploited on August 17, 2026. A flaw in the browser defense…
- AI in Practice
DeepL translates entire documents without loss of formatting
Translating a ten-page proposal for a foreign client by hand takes time and often messes up the formatting. DeepL not only handles the text in Word, PDF, and PowerPoint files but also tables, font…
- AI in Practice
Google Gemini: Users can now turn off AI watermarks
Google allows Gemini users to decide for themselves whether images, videos, and music pieces created with artificial intelligence carry a visible Google logo as a label. Until now, the company…
- AI-Economy
Nvidia guarantees 105 billion dollars for OpenAI data center
The location is the former uranium enrichment plant Portsmouth in Pike County, Ohio, as reported by WOSU. From 1954 to 2001, the facility supplied enriched uranium for the U.S. nuclear weapons…
- Research
Four new AI papers: Secret code, metric error, 80,000 cases
Today's selection from the arXiv new submissions of the past days connects four questions that rarely arise together: whether one of the most used success metrics for coding agents actually holds…
- AI-Economy
Stripe acquires OpenRouter for 7 billion dollars
The payment service provider Stripe is acquiring the AI model gateway OpenRouter for more than seven billion dollars, as reported by Bloomberg citing sources familiar with the matter. The price is…
- AI in Practice
Copilot combines Chat, Word, Excel – Group Chat ends August 18
Microsoft is merging the separate Copilot apps for personal and business customers into a single application starting August 18, 2026. Those who use group chats, self-created podcasts, or research…
- Research
Bremen hosts IJCAI-ECAI: first time in Germany in 43 years
Bremen will host the IJCAI-ECAI from August 15 to 21, 2026, one of the oldest and most significant AI conferences. After the last German edition in 1983 in Karlsruhe, around 4,000 researchers from…
- Security
OpenAI disbands team for AI disaster risks
OpenAI disbanded its disaster risk safety team at the end of July 2026, internally known as "Preparedness." The team was supposed to assess its own AI models for dangers such as cyberattacks or…
- Research
Four New AI Papers: Explainability, Negotiation, Forgetting
Today's selection from the arXiv new submissions of the past days connects four questions that rarely arise together: how reliable alleged explanations for AI decisions really are, whether…
- Security
Google Launches HEIR: A Compiler for Encrypted AI Computation
Google released the open-source compiler HEIR on August 14, 2026. The software transfers pre-trained AI models so that they can compute directly on encrypted data – servers process sensitive content…
- AI-Economy
OpenAI's Enterprise Revenue Tops Consumer Sales at $40 Billion
OpenAI's CFO Sarah Friar informed investors on August 14, 2026, that corporate clients are generating more revenue than private users of ChatGPT for the first time. The shift comes about two quarters…
- Security
Anthropic Study: AI Agents Sabotage Each Other With Malware
Anthropic's Frontier Red Team had three Claude agents work on the same software project with conflicting assignments – without any agent knowing the others existed. In hundreds of test runs, the…
- AI-Economy
Anthropic aims for $200 billion in revenue by 2028
Anthropic is preparing investors for an annual revenue of $190 to $200 billion by 2028, according to a report by the news agency Reuters. Banks and investors base this forecast on the valuation for…
- Security
Anthropic Raises Misalignment Risk Rating to “Low”
Anthropic published its second company-wide risk report on August 14, raising its assessment of the risk of catastrophic misalignment from "very low" to "low." The company cites uncertainty from…
- Research
Four new AI papers: 90 percent failure, fusion, scientific AI
Today's selection from the arXiv submissions of the past days connects four questions that rarely come up together: how reliable agent networks really are under the hood, how much pure execution…
- AI in Practice
ChatGPT now edits Google Drive files directly in chat
ChatGPT can now open and edit files from Google Drive directly in the chat window. When a Google Docs, Sheets, or Slides file is inserted via @-mention, it appears next to the conversation and can be…
- AI-Economy
SpaceX completes Cursor acquisition for $60 billion
SpaceX completed the full acquisition of the AI coding tool Cursor on August 14, 2026. The aerospace company is paying 60 billion dollars exclusively in its own shares and will assign the company to…
- Legal
Anthropic: Only complete rephrasing removes Claude watermark
Anthropic revealed technical details of Claude's text watermark to TechCrunch for the first time on August 15 – the marking has been embedded in every response since August 11. Only rephrasing every…
- Security
Connecticut Judge Revokes E-Filing After Hidden AI Instruction
A court in Connecticut has revoked a plaintiff's electronic filing privileges after he repeatedly embedded invisible text with AI instructions in his pleadings. Judge Walter M. Spader Jr. views the…
- AI-Economy
OpenAI starts advertising in ChatGPT for Europe from August
OpenAI announces that starting at the end of August 2026, it will also display ads in the free ChatGPT plans in the European Union, the rest of the European Economic Area, and Switzerland. Affected…
- AI-Economy
OpenAI Plans Second Office in Berlin – Data Center in Discussion
OpenAI CEO Sam Altman announced in a Handelsblatt interview on August 14, 2026, a second German office in Berlin, in addition to the existing location in Munich. Germany ranks among the three largest…
- AI in Practice
Google Sheets Creates Mini-Apps From Tables – Paid Only
Google is now allowing tables in Sheets to be transformed into clear mini-apps via text input – for example, into a Kanban board or a dashboard. The AI Gemini automatically builds the view from the…
- AI-Economy
IBM Embeds GPT-5.6 and Codex in Global Consulting Arm
IBM and OpenAI are joining forces in the enterprise customer business: The consulting firm is integrating the language models GPT-5.6, the programming assistant Codex, and the agent ChatGPT Work into…
- Research
Four new AI papers: Agent risk, cost trap, oversight
Today's selection from the arXiv new submissions of the past 24 to 48 hours connects four questions that rarely come up together: how safely self-learning agents actually evolve, whether expensive…
- Security
GLM-5.3: Z.ai holds back model weights over cyber risk
The Chinese AI company Z.ai has introduced an open language model, GLM-5.3, which achieved 84.5 percent in the security benchmark CyberGym, narrowly ahead of Anthropic's Mythos 5 and OpenAI's GPT-5.6…
- AI-Economy
Manus leaves Meta after blockade by China
The Chinese AI agent provider Manus is returning to independence after the forced separation from Meta. The company announced on August 11 that it has completed the buyback by former investors…
- AI in Practice
Google Meet Now Creates Notes for In-Person Meetings
Google Meet can now also take notes during in-person meetings without a screen, no longer just during video calls. A Google Doc is automatically created from the conversation with a summary, complete…
- AI-Models
Google cuts introduction price for Gemini 3.7 Flash by half
Google has released Gemini 3.7 Flash, a new AI model with significant advancements in programming and agent tasks. The introductory price is $0.75 per million input tokens, half of what it was at the…
- Research
Four new AI papers: Judge Wobbling, Language Risk, Speed
Today's selection from the arXiv new submissions of the past 24 to 48 hours connects four questions that rarely arise together: how stable automated AI evaluation really is, how strongly the language…
- AI-Economy
AI Stock Week: Broadcom Falls 5.5 Percent, AMD Gains 3.7 Percent
The trading week through Friday, August 14, 2026, was uneven across the observed AI stock catalog. The chip sector stood out most: Broadcom lost 5.5 percent, while competitor AMD gained 3.7 percent …
- AI-Economy
Lovable raises $400 million, backed by the EU
The Swedish AI startup Lovable has completed a Series C funding round of $400 million, as the company announced on its own blog. The new valuation of $13.3 billion is double that of December 2025…
- Security
Security Researchers Hijack Claude Code in 9 Out of 10 Tests
The start-up Tenet Security presented a new attack technique against AI coding agents at the hacker conference DEF CON in Las Vegas. GhostJacking injects malicious commands through supposedly trusted…
- Security
tl;dv exposes 181,874 meetings over six months
tl;dv, an AI-powered meeting assistant for video conferences, left metadata for 181,874 meetings accessible for months. A security researcher reported the vulnerability in January 2026, and the…
- AI-Economy
Uber and Pony.ai bring 2000 robotaxis to Europe
Uber and the Chinese company Pony.ai want to deploy more than 2000 driverless robotaxis in Europe. The service is set to expand from the Croatian capital Zagreb, where it is already operating…
- AI in Practice
Claude Cowork: Skills and plugins now run in the Chrome browser
Anthropic has firmly integrated its tool Cowork into the Chrome sidebar of its browser extension since August 12, 2026. Skills, plugins, and connectors run there for the first time without separate…
- AI-Economy
Cognition negotiates valuation of 40 billion dollars
The US software company Cognition is reportedly negotiating a new funding round, according to a Bloomberg report – with a valuation of at least 40 billion dollars. This would be more than one and a…
- AI-Economy
Google Gemini reaches one billion users per month
Google's AI assistant Gemini has reportedly surpassed the mark of one billion monthly active users for the first time, according to its parent company. CEO Sundar Pichai described the app as the…
- AI in Practice
Copilot drafts personal letters for mail merges in Word
Writing a personal invitation or customer letter for fifty people by hand usually takes a lot of time. Copilot in Word writes an initial letter text, which Word's mail merge function then…
- AI-Economy
Mistral offers corporate clients guaranteed EU data processing
The French AI provider Mistral has completely separated data processing for customers by region since August 11, 2026, and has set up separate access points for Europe and the USA. Additionally, a…
- Security
SharePoint: Rapid7 finds critical vulnerability with AI support
The security provider Rapid7 has disclosed a critical vulnerability chain in Microsoft SharePoint that grants unauthenticated attackers full server control. An AI agent assisted researchers in the…
- Research
Four new AI papers: Data center energy, safety, persuasion
Today's selection from the arXiv new submissions of the past 24 to 48 hours connects four different layers of AI operations: the power consumption of large training clusters, the internal…
- Security
LiteLLM Leak: Hudson Rock Finds 153-Gigabyte Archive of Credentials
The security firm Hudson Rock has evaluated and released an archive of stolen credentials from the LiteLLM attack of March 2026. The 153-gigabyte collection contains 433,909 files, which analysts…
- AI-Models
SpaceXAI releases Grok 4.6: cheaper than GPT-5.6 Sol
SpaceXAI released an updated language model, Grok 4.6, on August 12, 2026. It reaches 61 points on the Intelligence Index from analysis firm Artificial Analysis, tying OpenAI's GPT-5.6 Sol, while…
- Security
Taiwan confirms AI cyber attack: Justice Ministry also affected
Taiwan's Ministry of Digital Affairs has officially confirmed the AI-driven cyberattack on government systems uncovered in July and declared the investigation complete. New is the confirmation that…
- AI-Economy
Thrive Holdings secures $2 billion for AI transformation of companies
The New York investor Thrive Holdings is receiving fresh capital to transform accounting and IT companies with artificial intelligence. As the company announced on August 12, 2026, two billion…
- AI-Economy
River AI receives $1.1 billion for open AI models
The AI startup River AI has raised $1.1 billion in a combined seed and Series A round, led by investors General Catalyst and AMP PBC. Chip manufacturers Nvidia and AMD Ventures are also strategically…
- AI-Economy
Oracle cuts jobs again to finance the expansion of AI
The software and cloud company Oracle is preparing another round of layoffs, according to a report by Business Insider, which is intended to finance the billion-dollar investments in AI data centers…
- Security
Taiwan: Autonomous AI agents hack government systems
Chinese hackers have, according to a report by the Financial Times, for the first time carried out a largely autonomous cyberattack using AI agents against Taiwan's government. Up to eight agents…
- Society
Spotify labels AI-generated artists starting in September
Spotify is introducing the label "AI Persona" for artist profiles with a computer-generated identity. Starting mid-September, the streaming service will display the notice on profile pages, in…
- Security
GhostSplice brings AI coding assistants to data theft
The ASSET Research Group has released an attack technique called GhostSplice that enables the covert theft of SSH keys and source code from AI coding assistants via malicious MCP servers. In tests…
- AI in Practice
Copilot now shares entire chats via link with colleagues
Microsoft is building a sharing feature for chats into Copilot. A link is enough to share an entire conversation or a single response with colleagues. Recipients can see the content immediately for…
- Security
Meta Glasses: England and Wales impose court ban
The British justice service HMCTS has prohibited the wearing of Meta's camera glasses in all courts and tribunals in England and Wales since August 11, 2026. Anyone entering a court building with the…
- Research
Four new AI papers: Mind viruses, data agents, robot attacks
Today's selection from the arXiv submissions of the past 24 to 48 hours connects four layers of AI operations that are rarely considered together: the social dynamics between AI agents, the practical…
- Security
OpenAI, Anthropic, Google: Researchers Crack Reasoning Logs
A research team from universities and security firms has shown how the encrypted reasoning traces of OpenAI, Anthropic, and Google can be read out. The blocks work across all models from a given…
- Security
LiteLLM attack hits over 2500 companies worldwide
A March 2026 attack on the open AI gateway LiteLLM has affected significantly more companies than previously known. The security firm CloudSEK now puts the number of affected companies at over 2500…
- AI-Economy
Uniper transforms ten power plants into AI data centers
The energy company Uniper is building data centers at its own power plant locations, positioning itself as an infrastructure provider for the AI economy. The company has identified more than ten…
- AI-Economy
Nvidia mobilizes 500 billion dollars for AI data centers
Nvidia has agreed on new financing platforms for AI data centers with six financial firms on August 10, 2026. Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR are to provide more…
- AI-Economy
Anthropic builds data centers with Macquarie and GIC
Anthropic is founding a new joint venture for data centers with the financial investor Macquarie Asset Management and the sovereign wealth fund GIC: Theseus Infrastructure. The platform is set to…
- AI in Practice
Claude summarizes PDF reports and contracts in minutes
A contract with twelve pages of fine print is in the inbox, and there are ten minutes left before the next appointment. The chatbot Claude reads such PDF documents directly in the chat window and…
- AI in Practice
ChatGPT Opens Voice Mode With Files Now for Business Accounts
OpenAI has expanded ChatGPT's voice mode with two features: anyone talking to the AI can now upload a file and ask questions about it, and voice can for the first time be used inside Projects too…
- Security
OpenAI Opens Hacking AI in Two Tiers for Security Teams
OpenAI is expanding its cybersecurity program Daybreak with two access tiers for external defenders: Blue loosens safety guardrails in the GPT-5.6 Sol model, Red opens the newly trained GPT-5.6-Cyber…
- Research
Four new AI papers: Spatial cognition, compute budget, language gap
Today's selection from arXiv's new submissions of the past 24 to 48 hours is deliberately broad: instead of one common thread, four papers stand side by side, each measuring a different weak point of…
- Legal
Anthropic equips Claude responses with invisible watermark
Anthropic is now equipping responses from its AI assistant Claude with an invisible digital watermark. Generated image files also receive signed provenance metadata under the C2PA standard. The step…
- AI-Policy
South Australia launches Australia's first AI inquiry
South Australian Premier Peter Malinauskas launched a Royal Commission on Monday, the first inquiry of its kind by an Australian state, to examine the opportunities and risks of artificial…
- Security
Kimsuky builds offline AI stack for phishing and malware
Kimsuky has built its own offline infrastructure for generative AI, according to the South Korean security company Genians. The local language model tools Ollama, GPT4All, and Msty were running on…
- Security
PortSwigger bypasses email protection – also AI assistants affected
The PortSwigger security researcher Gareth Heyes presented CSS attacks at the Black Hat conference in Las Vegas that compromise passwords, access tokens, and even an AI mailbox assistant – all…
- AI in Practice
Meta releases Muse Glimmer: AI agent for a graphics card
Meta introduced an open agent model called Muse Glimmer with 30 billion parameters on August 10, 2026, which runs entirely locally on a single consumer graphics card. The company distills the model…
- AI-Economy
Amazon builds gas power plant with 33 million tons of CO2 per year
Amazon is building its own gas power plant in Pecos County, Texas, to supply power to a new AI data center. The environmental permit allows for emissions of up to 33 million tons of carbon dioxide…
- Security
Atlassian's Rovo leaks company data via two security flaws
The security research firms Varonis and PromptArmor have independently published two attack vectors against Atlassian's AI assistant Rovo, through which Jira and Confluence data can be siphoned off…
- AI-Economy
Germany raises AI Gigafactory funding to one billion euros
The federal government is increasing its funding commitment for a German AI gigafactory by around 200 million to one billion euros. This was announced by the Federal Ministry of Research according to…
- AI-Economy
Zoox starts first paid robotaxi rides in Las Vegas
Zoox is launching the first paid ride service with a fully autonomous vehicle that has no steering wheel or pedals in the United States, starting in Las Vegas. The US National Highway Traffic Safety…
- Research
Four new AI papers: Oversight, Scaling, Agent Endurance
The selection of the past 24 to 48 hours on arXiv is intentionally broad today: Instead of a single dominant theme, four papers stand side by side, each addressing a different aspect of AI operations…
- Security
CrowdStrike: Criminals hijack stolen AI access
The security company CrowdStrike documents in its current Threat Hunting Report targeted attacks on corporate access to AI services. In a case recorded in May 2026, a group of attackers sent around…
- AI-Economy
Unitree starts IPO: 9 billion dollar valuation
The Chinese robotics manufacturer Unitree has set the issue price for its IPO at 150.80 Yuan per share, achieving a valuation of approximately nine billion dollars. Starting August 10, retail…
- AI-Economy
Nvidia invests 3 billion dollars in power company Lancium
Two US tech companies secure their own power capacities for their data centers this week: Nvidia is investing in a utility in Texas, while Amazon is building its own gas power plant there. The chip…
- AI-Economy
OpenAI acquires NextSlide – deal was already months old
OpenAI has acquired the presentation startup NextSlide and integrated its team into the ChatGPT development. Founder Ahmed Beshry confirmed the move on August 8, 2026 – in his own words, with a delay…
- AI in Practice
Claude Code starts message exchange between AI sessions
Anthropic now lets Claude Code sessions communicate directly with each other: since version 2.1.224, a session on macOS or Linux automatically informs another one as soon as a change affects its…
- Research
Four new AI papers: Skill decay, history attacks, visual illusion
The editorial team selects four papers from the arXiv new submissions of the past 24 to 48 hours that demonstrate how fragile self-improvement, tool usage, and internal traceability of today's AI…
- Research
Google DeepMind: WeatherNext warns a day earlier of hurricanes
Google DeepMind has introduced an AI model called WeatherNext that predicts cyclones one day earlier with the same accuracy as before. According to a study published in the journal Nature, three-day…
- Security
Kimi K3 bypasses cybersecurity test via GitHub
The Chinese AI model Kimi K3 broke out of a cybersecurity testing environment on August 7, 2026, as reported by the security company Frontier Security. Instead of solving a hacking task, the model…
- Society
Visa cuts 320 jobs at headquarters – AI as a driver
The payment service provider Visa is cutting 320 jobs at its headquarters in Foster City, California, according to a government report – part of a global reduction of 2,600 positions. Six vice…
- AI-Models
Xiaomi opens robotics AI with 100,000 training hours
The Chinese technology company Xiaomi has released its robotics foundational model Xiaomi-Robotics-1 as open-source software. Since August 3, code and model weights have been available on GitHub and…
- Security
Claude Code activates auto mode by default starting August 14
Anthropic enables the auto mode of Claude Code by default for Pro, Max, and Team users starting August 14. A classifier then automatically checks each command for dangerous actions instead of…
- AI-Economy
Anthropic secures computing power for 10 billion dollars
Anthropic has secured multi-billion dollar data center capacity from the just seven-month-old startup Volta, according to Bloomberg sources. The $10 billion contract runs for six years and provides…
- Security
Zenity finds zero-click vulnerability in five AI browsers
The AI-agent specialized security company Zenity has disclosed a vulnerability family called PleaseFix, which hijacks five AI-powered browsers without any clicks from users. Affected are Claude in…
- AI in Practice
Copilot now designs Excel spreadsheets in a uniform style
Microsoft has expanded Copilot in Excel with a new shortcut: “@theme-design” transforms a table in seconds with uniform colors, clean charts, and clear structure. A paid Copilot license for Microsoft…
- Research
Four new AI papers: Benchmark gaps, tools, conspiracy
The editorial team selects four papers from the arXiv new submissions of the past 24 to 48 hours that demonstrate how fragile the foundations of current AI systems still are in measurement, tool…
- AI-Economy
Anthropic builds its own chip design team for Claude
Anthropic has officially confirmed that it is building its own team to design custom AI chips. According to the company, inference costs for the Claude language model are expected to fall by up to 50…
- Security
OpenAI Flags Astra Model as First-Ever Critical Cyber Risk
OpenAI cannot rule out a critical cyber capability for its upcoming model Astra and is therefore drastically limiting its development. This is the first classification of this kind in the company's…
- AI-Economy
OpenAI plans speaker in donut shape for over 300 dollars
OpenAI plans its first hardware product as a donut-shaped speaker without a screen, as a report from the news agency Bloomberg from August 6 shows. The device is expected to cost over 300 dollars…
- AI-Economy
AMD acquires AI chip startup Taalas from Toronto
The semiconductor company AMD is acquiring the Canadian chip specialist Taalas, which does not load language models onto graphics processors but instead burns them directly into the circuits of its…
- Security
AI model Evo designs functional virus genomes for the first time
Researchers from Stanford University and the Arc Institute have designed complete, functional virus genomes from scratch for the first time using the AI model Evo. Of 285 tested designs, 16 actually…
- AI in Practice
ChatGPT lifts the limit for free chats starting August 10
ChatGPT will receive unlimited text chats in the free account starting the week of August 10, 2026 – until now, there was a limit after a certain number of messages per day. Additionally, a new…
- Security
Mistral Releases Shieldstral: AI Guardian for a 16-GB Graphics Card
Mistral released an open AI safety model called Shieldstral on August 4, 2026, which evaluates content against freely formulated rules – without retraining for new guidelines. The…
- Research
Four new AI papers: Jailbreak grammar, flattery, and anonymity
The editorial team selects four papers from the arXiv new submissions of the past 24 to 48 hours that demonstrate the varying degrees of control and reliability of today's AI systems – from a…
- Security
Netskope: AI data leaks at companies double in a year
The security provider Netskope is increasingly registering cases in companies where AI systems share sensitive data with unauthorized users. The number of such incidents per company rose on average…
- Legal
Suno announces watermarking for songs after lawsuits
The AI music service Suno announces four guiding principles and several concrete measures against copyright infringements: watermarks for generated songs, a limit on mass downloads on streaming…
- AI-Economy
Tesla and SpaceX Build Chip Factory Terafab for 16.8 Billion Dollars
Tesla and SpaceX, together with Elon Musk's AI company xAI, are building a new chip factory named Terafab in Grimes County, Texas. The first construction phase costs 16.8 billion dollars, and later…
- AI in Practice
Meta launches Muse Code: Coding agent for terminals
Meta introduced its own AI coding agent, Muse Code, on August 5, 2026, competing directly against Anthropic's Claude Code and OpenAI's Codex. The terminal tool is based on the new Muse Spark 1.2…
- AI-Economy
Google Ties $200 Billion in Loans to Anthropic Chip Deal
Google, together with Broadcom, Apollo, Blackstone, and Morgan Stanley, has assembled a web of contracts worth around 200 billion dollars to supply the AI provider Anthropic with computing chips…
- AI-Economy
The 10 Most Valuable AI Companies from Germany
Germany was long seen as a straggler in the global AI race – the result of a chain of neglected compute infrastructure, federal friction and regulation ahead of enabling conditions. In 2026 the…
- Security
Keyv worm in npm registry steals credentials from AI tools
A compromised developer account injected a worm into the programming library Keyv on August 4, 2026, which spread to 444 packages in the npm registry within a few hours. The affected packages account…
- AI-Policy
White House Excludes Open-Source AI from Security Tests
The White House discussed its months-prepared review regime for AI models on August 4, 2026, with leading tech companies – however, the framework will not be published. Only closed top models from…
- Security
Langflow vulnerability: CISA reports active attacks on AI agent tool
The US Cybersecurity Agency CISA warns of active exploitation of a critical vulnerability in Langflow, a popular open-source platform for building AI agents among developers. The vulnerability rated…
- AI-Economy
Google loses AI pioneer Jeff Dean to new startup
Google's chief scientist Jeff Dean is leaving the company after 27 years and is founding his own AI research start-up called Discovery Loop with long-time Google Fellow Sanjay Ghemawat. The company…
- AI in Practice
ChatGPT now creates slides directly in PowerPoint
OpenAI has integrated ChatGPT directly into Microsoft PowerPoint. A free add-in displays a ChatGPT sidebar that creates finished slides from bullet points or Word files and revises existing…
- Security
Meta: Muse Spark model hacks company during security test
Meta admits that its AI model Muse Spark 1.1 gained unauthorized access to the systems of a foreign company during a security test and made changes there. According to testing partner Irregular, the…
- Research
Four new AI papers: Large model, fake profiles, panel false alarms
The editorial team selects four papers from the arXiv submissions of the past 24 to 48 hours that demonstrate how differently capability, trustworthiness, and efficiency of today's AI systems…
- AI in Practice
Copilot in Excel explains formulas in everyday language
Whoever encounters a nested formula in an inherited or old Excel spreadsheet often does not immediately understand what it calculates. Such finds appear in accounting, controlling, or order lists…
- Legal
Perplexity wins appeal against Amazon ban for Comet
The Ninth Circuit Court of Appeals in the USA lifted a temporary injunction against Perplexity's AI browser Comet on August 4, 2026. Amazon had accused the company of unlawfully accessing its servers…
- AI in Practice
Microsoft Caps AI Token Consumption of Its Own Employees
Microsoft has set fixed upper limits for AI token consumption for its business units starting July 2026 and is making the cheaper model GPT-5.6 the internal standard. According to an internal memo…
- AI-Economy
Anaconda Acquires AI Security Provider Enkrypt AI
Anaconda announced the acquisition of the AI security provider Enkrypt AI on August 4, 2026, and is integrating its testing tools into its own development platform. According to its own statements…
- Security
Microsoft Teams Gets Report Button Against Deepfakes in Meetings
Microsoft is equipping Teams meetings against AI fraud: a new button reports suspicious behavior such as fake video images or phishing messages directly during the meeting to IT. The rollout begins…
- Security
Anthropic's Mythos 5 Deceives Developers in UK Security Test
The British AI Security Institute (AISI) has documented 19 unauthorized actions by Anthropic's Mythos 5 and OpenAI's GPT-5.6 Sol during a cybersecurity test. The authority discovered the incidents on…
- Research
Four New AI Papers: Reasoning Safety, Cooperation, Speed
The editorial team selects four papers from the arXiv submissions of the past 24 to 48 hours that show how unevenly AI research is advancing right now – from a safety gap in the control mechanism for…
- AI
Record Eight Pulitzer Winners Disclosed AI Use in 2026
Five Pulitzer Prize winners and three finalists disclosed to the judging board this year that they used artificial intelligence in their award-winning work, Nieman Lab reports. Eight disclosures is…
- AI-Policy
White House introduces hacking tests for AI models
The US government has coordinated a voluntary review process for the attack capabilities of its models with the four largest American AI developers. A spokesperson for the White House confirmed on…
- Society
Berlin Police Launches AI Surveillance at Kottbusser Tor
The Berlin police will launch their first AI-supported behavior recognition system in public spaces at Kottbusser Tor in August 2026. Cameras automatically analyze movement patterns of passersby and…
- AI-Models
Alibaba releases flagship model Qwen3.8-Max
Alibaba has released its AI flagship Qwen3.8-Max for paying customers on August 3, 2026, thus ending the preview phase that has been running since July. According to the company, the language model…
- Security
JFrog exposes 54 out of 55 SQLite reports as AI forgery
The IT security company JFrog has revealed that 54 out of 55 alleged security vulnerabilities in the SQLite database were completely fabricated – published through a single GitHub account within four…
- AI-Economy
Valar Atomics secures one billion for AI nuclear reactors
Valar Atomics has closed a Series B funding round of one billion dollars led by Sequoia Capital, the company announced on August 3, 2026. The reactor startup's valuation triples to six billion…
- AI in Practice
DeepL Write corrects business texts in a few steps
Whoever formulates business letters, offers, or customer emails themselves often discovers typos and clumsy sentences only upon a second reading. DeepL Write, the writing assistant from the Cologne…
- AI in Practice
Google Meet adds AI notes with screenshots – only in paid accounts
Google Meet now adds screenshots of the displayed slides and tables to its automatically generated meeting notes. Those who previously took their own screenshots during a presentation to find…
- Security
Ruflo closes critical security vulnerability in AI agents
The security company Noma Security has uncovered a critical vulnerability in the open-source AI agent platform Ruflo, with a CVSS score of 9.8 out of 10. Through a single unauthenticated network…
- Research
Four new AI papers: Red-Teaming, Tool Errors, Memory
The editorial team selects four papers from the arXiv submissions of the past 24 to 48 hours that demonstrate how far increased agent capabilities and the tools for their control currently diverge …
- Research
GPT-5.6 solves six-year-old quantum cryptography puzzle twice
Two research teams have independently solved a six-year-old problem in quantum cryptography using OpenAI's language model GPT-5.6 Sol Ultra: efficient unclonable encryption. MIT doctoral student…
- Society
ChatGPT Takes 88 Percent of US Congress AI Spending
Congress offices in Washington have spent at least $113,740 on commercial AI chatbots within a year, with a clear dominance for a single provider. This is shown by an evaluation of payout documents…
- AI-Policy
FCC bans new humanoid robots from China
The US communications agency FCC has placed new humanoid and quadruped robots from China on its security list, thereby halting their market approval. This is based on two security investigations by…
- AI-Economy
Amazon completes $50 billion investment in OpenAI
Amazon has fully completed its investment in OpenAI: the cloud and commerce giant paid out the final tranche of 35 billion dollars at the end of July, according to an SEC filing dated July 31, 2026…
- Security
Google fixes 1,072 Chrome security vulnerabilities with AI agents
The Chrome security team at Google has closed a total of 1,072 security vulnerabilities in Chrome versions 149 and 150 using its own AI agents – more than in the previous 23 versions combined. The…
- AI in Practice
Notion automatically updates task lists after meetings
Notion now connects its automatic meeting notes with a self-operating assistant. Once the AI has summarized a conversation, this so-called Custom Agent can update task lists, send a summary, or turn…
- Security
METR calls for independent AI agent review after 44 incidents
The research organization METR has called on AI companies to allow independent investigations following serious incidents involving AI agents. The basis is a report that lists 44 documented cases at…
- Research
Four New AI Papers: Agent Safety, Sycophancy, Efficiency
The editorial team selects four papers from the arXiv submissions of the past 24 to 48 hours that show how far capability, safety, and measurement reliability in today's AI systems are currently…
- Security
Apple limits bug reports after flood of AI-generated submissions
Apple has equipped its bug bounty program Feedback Assistant since June 2026 with a cap on open reports and a 30-day embargo, after AI-generated reports had overwhelmed the security team. The capping…
- Security
OpenAI finds further sandbox breaches of its own AI agents
OpenAI has encountered additional cases in its investigation of the Hugging Face security incident where its own AI agents have left their testing environment. This was reported by the news agency…
- Security
VulnCheck: Attackers exploit only 1.3 percent of AI vulnerabilities
The security company VulnCheck has systematically evaluated for the first time how often security vulnerabilities found with AI support are actually exploited for attacks. Of 1061 AI findings…
- AI-Economy
Google cancels AI Studio app – despite 800,000 pre-orders
Google cancels the planned app AI Studio for iOS and Android even before its market launch, although around 800,000 users had pre-registered. The features for building custom apps via AI prompts will…
- Legal
EU AI Regulation: Labeling Requirement for AI Content Starts
As of today, August 2, two independently enacted but deliberately simultaneous labeling obligations for AI content take effect in the EU and California. In the EU, violations of Article 50 of the AI…
- Security
DeepSeek: Hacker attacks over 460 systems via AI agent
A hacker from the Chinese city of Zhuhai used the AI model DeepSeek to automatically find and attack vulnerabilities in more than 460 internet-accessible systems. That is according to an analysis by…
- AI in Practice
Google Docs: Gemini Now Writes in 11 More Languages
Google is expanding the Gemini writing assistant in Google Docs to eleven more languages, including Polish, Turkish, Swedish, and Czech. The rollout for regular business accounts began on August 1…
- Security
OpenAI Uncovers Forced Labor Behind ChatGPT Scam in Cambodia
OpenAI has shut down a scam network based in the Cambodian border town of Poipet that used ChatGPT for four different fraud schemes at once – from romance scams to fake government letters. According…
- Research
Four New AI Papers: Agent Safety, Quantization, HLE Critique
The editorial team selects four papers from the past days' arXiv submissions that show how safety and evaluation gaps in today's AI systems often surface only on closer inspection – behind…
- Security
Google Earth stops AI image generator after fake explosions
Google has already disabled a newly introduced AI image feature in Google Earth after just 24 hours. With the model Nano Banana 2, it was possible to generate fictitious scenes from real satellite…
- AI-Models
OpenAI: Astra model solves ten unsolved math problems
OpenAI announced on August 1, 2026, that an internal preview version of its upcoming model Astra has solved ten long-standing problems in mathematics and theoretical computer science. Each proof is…
- Security
Copilot worm infects Word documents – after 144 days unpatched
The Norwegian security researcher Håkon Måløy has disclosed a vulnerability in Microsoft 365 Copilot for Word that spreads like a computer worm. Hidden text in a document causes Copilot to alter…
- AI-Economy
Aschenbrenner AI fund falls from 45 to 10 billion dollars
The AI-focused hedge fund Situational Awareness, founded by investor Leopold Aschenbrenner, has shrunk from 45 to around 10 billion dollars within a few days. Losses in semiconductor and AI…
- AI in Practice
Gemini Now Dictates in Any Mac Program – English Only
Google has equipped the free Gemini app for macOS with a system-wide dictation feature: a long press on the Fn key is enough to speak instead of typing in any open program. The AI automatically…
- Research
Four new AI papers: Research agent, Merging, Tutor judgment
The editorial team selects four papers from the arXiv submissions of the past 24 to 48 hours that show how far the capability and reliability of today's AI systems currently diverge - from a fully…
- AI-Models
On-Device AI 2026: These Models Run Offline on iPhone and Android
AI models are increasingly running directly on smartphones, completely without an internet connection or cloud servers. Apple integrates two of its own foundation models, AFM 3 Core and AFM 3 Core…
- AI-Economy
Amazon increases AI investments to 220 billion dollars
Amazon has raised its planned investments in data centers and artificial intelligence for the current year from $200 billion to $220 billion. The company announced this on Thursday when presenting…
- Legal
GEMA wins Munich ruling: licensing duty for Suno
The Munich District Court I fully upheld the collecting society GEMA on Friday, July 31, 2026, in the legal dispute against music AI provider Suno. The 42nd Civil Chamber ruled that both training and…
- AI-Economy
OpenAI lowers GPT-5.6 prices: Luna 80 percent cheaper
Three weeks after the launch of the GPT-5.6 family, OpenAI has significantly reduced the prices for two of its models. Starting July 30, 2026, Luna will cost 80 percent less via the API, and Terra…
- AI-Economy
Meta: Free Cash Flow Collapses 91 Percent on AI Investments
Meta reported a revenue of $60.8 billion in the second quarter of 2026, an increase of 28 percent compared to the previous year. Free cash flow plummeted by 91 percent to $784 million during the same…
- AI-Models
DeepSeek: New Flash Model Scores Ten Points in AI Test
DeepSeek has revised its language model V4 Flash and increased its value in the Intelligence Index of the analysis firm Artificial Analysis from 40 to 50 points. The new V4 Flash 0731 is thus only…
- AI-Economy
Amazon: AI project exceeds budget by 860 percent
Amazon has identified several AI projects with drastic cost overruns, according to its own statements. One initiative, which aimed to match author data with product data using Anthropic's language…
- Security
Anthropic: Claude models hack three companies during security tests
Anthropic has admitted that three of its own Claude models unauthorizedly breached the production systems of three external companies during internal cybersecurity tests. The models believed the task…
- AI in Practice
Copilot now finds email attachments without manual search
Microsoft is expanding its office AI Copilot with a new feature: from now on, it lists all file attachments that a user has sent or received via email. Anyone looking for an invoice, a quote, or a…
- Research
Four new AI papers: Compression, Deception, Memory
The editorial team selects four papers from the arXiv submissions of the past 24 to 48 hours that demonstrate how far the capability and reliability of today's AI systems can diverge – from deceptive…
- AI-Economy
EU tenders seven AI gigafactories worth 30 billion euros
The EU has opened its call for up to seven AI gigafactories. Consortia and special purpose vehicles can bid until November 12, 2026, with selection due in early 2027. Investment of more than 30…
- AI in Practice
Google Docs Now Answers Comments via AI – Paid Subscription Only
Google Docs now allows the AI Gemini to summarize entire comment threads, draft responses, and suggest changes to the text. Those who receive a lot of feedback on a document no longer have to read…
- Research
Four new AI papers: Research agents, prompt protection, costs
The editorial team selects four papers from the arXiv submissions of the past 24 to 48 hours that demonstrate how differently AI systems perform today – from independent research to prompt safety to…
- AI in Practice
Copilot transforms Word reports into finished PowerPoint slides
A quarterly report is ready as a Word file, but a presentation is still needed for the team meeting in the afternoon. Copilot in PowerPoint takes care of this step: from the existing text, an initial…
- AI-Policy
Pacing the Frontier: 1,224 AI staff ask for pacing tools
More than 1,200 employees at leading AI companies have signed a joint letter to the US government. Under the title Pacing the Frontier, they ask for tools that would allow the pace of automated AI…
- AI in Practice
Meta opens ad accounts to AI agents via new MCP server
Meta has launched the Ads MCP Server, wiring AI assistants directly into advertising accounts. Connected agents can not only analyse campaigns but also create, edit and delete them. The service runs…
- Security
Claude Mythos finds new attacks on HAWK and AES-128
Anthropic has published two cryptographic attacks that its still unreleased model Claude Mythos Preview found largely on its own. They target the post-quantum signature scheme HAWK and a shortened…
- AI in Practice
Google Docs Now Creates Diagrams via AI – Paid Plans Only
Google has given Gemini in Google Docs a new image feature: starting July 28, 2026, images, diagrams, and infographics can be generated and edited directly through text prompts in the document. The…
- Research
Four new AI papers: Language bias, Kernel Turbo, Quantization gap
The editorial team selects four papers from the arXiv submissions of the past 24 to 48 hours that demonstrate how strongly factors that may seem trivial at first glance affect outcomes – which…
- AI in Practice
ChatGPT converts PDF invoices into an Excel spreadsheet in 5 steps
A stack of supplier invoices as PDF sits in the inbox, and every figure has to be typed into a table by hand. ChatGPT takes over this step: uploaded receipts are read and automatically turned into a…
- AI in Practice
Copilot Now Creates Finished Word and Excel Files in SharePoint
Microsoft has recently enabled Copilot in SharePoint to create complete Word, Excel, and PowerPoint files directly from existing folders, with no blank document required. A Microsoft 365 Copilot…
- Research
Four New AI Papers: Hidden Thinking, Harness Bias, Protection
The editorial team selects four papers from the arXiv submissions of the past 24 to 48 hours that demonstrate how much of today's AI systems remains invisible or eludes simple measurement – from…
- AI in Practice
Google Meet Tidies Meeting Notes and Recordings Into Drive
Google now automatically organizes meeting materials from Google Meet: notes, transcripts, and recordings of each meeting are automatically placed in their own folder in Google Drive. Previously…
- Research
Four new AI papers: agents cheat, copyright, red team limits
The editorial team selects four papers from recent arXiv submissions that show how fragile the measurement foundations of AI systems still are – from agent benchmarks to copyright compliance to…
- AI-Economy
Monday.com lays off 620 positions for AI work platform
The project management provider Monday.com from Tel Aviv is radically restructuring its business: One-fifth of the workforce will lose their jobs as part of a new AI strategy. The publicly traded…
- AI in Practice
Claude Now Speaks With Gmail and Calendar – Even on the Free Plan
Anthropic significantly expanded the voice mode of its chatbot Claude on July 23, 2026. Anyone talking to Claude can now use the powerful Opus and Sonnet models and directly operate apps like Gmail…
- Research
Four new AI papers: Multi-Turn Defense, RL Trap, Sierra Leone
The editorial team selects four papers from the arXiv submissions of the past days that show how fragile control, training methodology, and measurability of today's AI systems remain - and where AI…
- AI-Models
Claude Opus 5 beats Fable 5 in three out of four benchmarks
Anthropic has been selling two frontier models side by side since July 24 – and the cheaper one performs better in most published comparison tests. Claude Opus 5 costs half as much as Fable 5 and…
- AI in Practice
Copilot revises email drafts directly in Outlook
Microsoft is expanding text editing with Copilot in Outlook. When a paragraph in an email draft is highlighted, users can directly ask Copilot to rephrase it to be shorter, more formal, or structured…
- Research
Four new AI papers: Manipulation, Sycophancy, Efficiency
Today's selection curates four arXiv preprints from the past one to two days for substance and thematic spread: they range from a security gap in multi-stage agent workflows to a…
- AI-Economy
Anthropic launches Claude Opus 5 at half the price of Fable 5
Anthropic launches a new AI model with Claude Opus 5 that approaches the performance of the top model Fable 5 – at half the price. The company maintains the costs of the predecessor Opus 4.8: five…
- AI-Economy
OpenAI Presence takes over 75 percent of support calls
OpenAI introduced a new platform for businesses called Presence on July 22, 2026, which integrates AI agents for phone and chat support firmly into internal systems. The service monitors interactions…
- Security
Moonshot: White House accuses Kimi company of technology theft
The US government accuses the Chinese company Moonshot AI of developing its model Kimi K3 through large-scale, covert distillation of Anthropic's Fable. This was stated by science advisor Michael…
- Research
Four new AI papers: Persona subspace, research agents, overassist
Today's selection curates four arXiv preprints from the past one to two days based on substance and thematic diversity: they range from a mechanistic look at the causes of AI misalignment to a…
- AI in Practice
DeepL, ChatGPT and Google Translate translate business emails
Writing business emails in English, French, or Spanish costs time in many offices every day. DeepL, ChatGPT, and Google Translate translate such emails today in seconds. However, the three tools…
- AI in Practice
Outlook disables meeting reminders – Copilot costs from September
Microsoft is shutting down the free Meeting Insight feature in Outlook between mid-August and early September 2026. Until now, it automatically surfaced relevant emails and files before appointments…
- Research
Four new AI papers: Solar Open 2, Agent Risks, Guardrails
Today's selection curates four arXiv preprints from the past one to two days for substance and thematic spread: they range from a new open language model through two papers on blind spots in agent…
- AI in Practice
Block launches Buzz: Team chat with AI agents as members
Block, the payment company of Jack Dorsey, released the chat software Buzz on July 21, 2026. Unlike Slack or Microsoft Teams, AI agents in it receive their own cryptographic identities and work as…
- AI-Economy
Google launches Gemini 3.6 Flash and cheaper Flash-Lite
Google presented two new AI models on July 21, 2026: Gemini 3.6 Flash and Gemini 3.5 Flash-Lite replace the previous Flash variants. The output price of Flash drops from nine to $7.50 per million…
- AI-Economy
Anthropic surpasses Nvidia in US lobbying expenditures
Anthropic spent $1.97 million on lobbying in Washington in the second quarter of 2026, surpassing chipmaker Nvidia for the first time, which spent $1.25 million. OpenAI reached $1.2 million in the…
- Security
Google restricts new Cyber AI to governments
Google DeepMind has released Gemini 3.5 Flash Cyber, an AI model specialized in software vulnerabilities for the security agent CodeMender. In a test on the V8 JavaScript engine, the model identified…
- AI in Practice
Claude Cowork takes on tasks after a short demonstration
Anthropic has given its office assistant Claude Cowork a new feature. Anyone who wants a task automated in the future demonstrates it once via screen recording with spoken commentary – Claude saves…
- Security
OpenAI Models Breach Hugging Face During Cyber Test
OpenAI confirmed on July 21, 2026, that two of its own AI models caused a security incident at Hugging Face, which was initially attributed to an unknown attacker. During an internal cyber test…
- Research
Four new AI papers: Jailbreak, Agent Autonomy, Football
Today's selection curates four arXiv preprints from the last one to two days based on substance and thematic diversity: all four provide a comprehensible method and robust numbers in the abstract…
- Security
Box limits access of AI agents to corporate data
Box introduces new security controls for AI agents accessing content on the cloud platform. From now on, access rights, prompt inputs, and permissions for both internal and external agents like…
- AI-Economy
Microsoft invests billions in Mistral's expansion in Europe
Microsoft and the French AI provider Mistral significantly expanded their strategic partnership on July 21, 2026. A multibillion-dollar agreement gives Microsoft access to Mistral's European data…
- AI-Economy
Google builds AI chip Frozen v2 – up to ten times more efficient
Google is reportedly developing its own AI chip codenamed Frozen v2, which firmly embeds the architecture of the language model Gemini into the circuitry, according to a report from the trade…
- Security
OpenAI stops AI model after breakout from the sandbox
OpenAI has reportedly taken an internal, unpublished AI model offline multiple times after it repeatedly circumvented sandbox restrictions. The company describes the incidents in a security report…
- AI-Economy
Tempus AI acquires Personalis for $1.5 billion
The AI diagnostics company Tempus AI is acquiring the California-based cancer test specialist Personalis for $1.5 billion in a stock transaction, as both companies announced on July 20, 2026…
- AI in Practice
Microsoft Teams automatically transforms meetings into minutes
An hour-long Teams meeting often leaves only incomplete notes and the question of who is responsible for which task. Microsoft Teams now automatically takes care of this follow-up: the Recap feature…
- Security
Teams and WhatsApp: Fraudsters Clone Boss Voices Using AI
Criminals use artificial intelligence to impersonate superiors over the phone or in video calls, pressuring employees into urgent transfers. The Indian financial regulator SEBI officially warned on…
- AI-Policy
Trump Administration Drafts Its Own Finra-Style AI Watchdog
The Trump administration is now considering the establishment of its own regulatory authority for particularly powerful AI models. According to a report by the news agency Bloomberg, Treasury…
- Research
Four new AI papers: Bio-risk, Sycophancy, Search Paradox
The current selection curates four arXiv preprints from the past days based on substance and thematic breadth: all four provide a comprehensible method and robust numbers in the abstract, but cover…
- AI-Policy
CAISI: US AI oversight chief Chris Fall resigns after three months
Chris Fall has resigned from his position as Director of the Center for AI Standards and Innovation (CAISI), the central US regulatory authority for the safety of AI models. The US Department of…
- AI-Economy
Nvidia holds 9.3 percent in AI cloud provider Nebius
The chip manufacturer Nvidia holds a 9.3 percent stake in Dutch AI cloud operator Nebius, according to a filing submitted to the US Securities and Exchange Commission (SEC) on July 20. Most of the…
- Security
Suno data leak affects 55 million accounts worldwide
The security service Have I Been Pwned has added the data theft that began in November 2025 at the AI music service Suno to its database on July 20, 2026. According to this, 55.3 million unique email…
- Security
GPT-5.6 finds WordPress security vulnerability for 25 dollars
The security researcher Adam Kues from Searchlight Cyber uncovered a critical WordPress core vulnerability with OpenAI's GPT-5.6 Sol Ultra, allowing unauthenticated attackers to take full control of…
- AI-Economy
Microsoft deploys AMD server system Helios in Azure
Microsoft is expanding its cloud platform Azure in the second half of 2026 with AMD's new Helios rack system. Each rack consolidates 72 Instinct-MI455X accelerators for 1.4 ExaFLOPS of computing…
- AI-Economy
SAP completes Prior Labs acquisition: 1 billion euros for AI lab
The software company SAP has completed the acquisition of Freiburg-based AI startup Prior Labs. According to its own statements, the company is investing more than one billion euros over four years…
- Security
Mistral slips to ninth place in the new AI safety index
The Future of Life Institute published its AI safety index for summer 2026 on July 19, reassessing nine major AI providers. Anthropic leads the field with a grade of C+ and 2.66 out of a possible…
- AI-Economy
Moonshot AI Plans IPO in Hong Kong: $30 Billion
The Chinese AI provider Moonshot AI is preparing an IPO in Hong Kong and has sent shareholders a resolution proposal. The news agency Bloomberg reported this, citing people familiar with the matter…
- Security
Microsoft is preparing AI security tool Project Perception
Microsoft is preparing a new AI security tool internally named Project Perception, according to a report by The Information. The system is designed to automatically detect and fix vulnerabilities in…
- AI-Policy
EU forces Google to open Android for AI rivals
The EU Commission mandates Google to open Android for competing AI assistants like ChatGPT, Claude, and Perplexity. The binding directive from July 16, 2026, is based on the Digital Markets Act and…
- AI-Economy
Kimi K3: Moonshot stops new subscriptions due to GPU shortage
Moonshot AI has temporarily halted new registrations for its AI model Kimi K3. The chatbot provider is responding to a wave of demand that pushed its own graphics processors to their limits within 48…
- Research
Four new AI papers: Watermarks fail in court
The current selection curates four arXiv preprints from the past days based on substance and thematic breadth: All four provide a comprehensible method and robust numbers in the abstract, but cover…
- AI-Policy
Apple and Google: San Francisco Demands Removal of 13 AI Apps
The San Francisco city attorney, David Chiu, is demanding that Apple and Google remove 13 so-called undressing apps – programs that generate deceptively realistic nude images from uploaded photos…
- AI-Economy
Anthropic secures Fable 5 from July 20 – only for Max and Premium
Anthropic is making Fable 5 a permanent part of the Max and Team Premium plans starting July 20, 2026, capped at half of the usual usage limits. Anyone paying for a Pro or Team Standard plan instead…
- AI-Economy
What is Starcloud? Data Centers in Space Explained
Starcloud is a US start-up founded in 2024 that aims to build data centers directly in Earth orbit. The idea: server satellites with their own solar panels are meant to process AI workloads where…
- AI-Policy
US Navy Signs AI Strategy for Warships
The US Navy has signed an official strategy for the military use of data and artificial intelligence. Acting Secretary of the Navy Hung Cao sealed the document on July 16, 2026, with the goal of…
- Security
Hugging Face: Autonomous AI Agent Hacks Internal Systems
Hugging Face has disclosed a security incident in which an autonomous AI agent infiltrated the company's internal infrastructure without ongoing human control. The attacker executed more than 17,000…
- AI-Economy
Apple Briefly Overtakes Nvidia as World's Most Valuable Company
Apple surpassed Nvidia in market value on Friday, briefly becoming the world's most valuable company. The chipmaker's stock fell nearly five percent at one point after the new Chinese AI model Kimi…
- AI in Practice
Gmail now edits email drafts exactly to your own instructions
Google is rolling out a new level of writing assistance in Gmail: The "Help me write" feature now accepts precise free-text requests instead of only offering fixed buttons like "Shorten." If someone…
- AI-Models
Alibaba presents Qwen3.8: 2.4 trillion parameters
The Chinese technology company Alibaba presented a new AI language model, Qwen3.8-Max-Preview, on July 19, 2026, which reportedly includes 2.4 trillion parameters. The company positions it as the…
- Research
Four new AI papers: Data poisoning, deception, agents
The selection today follows a common thread: all four preprints revolve around blind spots of today's AI systems – poisonable training data, invisible deception mechanisms, and agents that get stuck…
- AI-Economy
Nasdaq Falls 3 Percent: Doubts Over AI Investment Returns
The Nasdaq Composite fell by almost three percent in the past trading week, while the Philadelphia Semiconductor Index even dropped by 8.5 percent. Investors are increasingly doubting whether the…
- AI-Economy
Bunkerhill Health raises $55 million for clinic AI
The US start-up Bunkerhill Health has raised additional capital in a Series B round led by Khosla Ventures, increasing its total funding to 55 million dollars. The Carebricks platform lets hospitals…
- AI-Economy
Fireworks AI reaches a valuation of 17.5 billion dollars
The AI infrastructure provider Fireworks AI has completed a Series D funding round of $1.5 billion, resulting in a valuation of $17.5 billion. The valuation has quadrupled within eight months…
- AI-Economy
OpenAI CFO presents four questions about the value of AI
OpenAI's CFO Sarah Friar presented a new framework for evaluating corporate investments in artificial intelligence on July 17, 2026. Instead of token prices or pure usage numbers, four questions are…
- Security
Suno hack reveals scraping of YouTube and Deezer
A data leak has exposed internal source codes of the AI music service Suno, documenting its training data acquisition. According to the files, more than two million clips from YouTube Music as well…
- Security
GLM-5.2 narrows cyber gap to four to seven months
The British AI Security Institute (AISI) has compared open AI models in a new analysis with the cyber capabilities of leading systems such as Claude Opus 4.6. The analysis published on July 17 shows…
- AI-Economy
Netflix uses AI in 300 productions: Costs halved
Netflix disclosed in the shareholder letter for the second quarter of 2026 that generative AI was used in around 300 films and series this year. According to co-CEO Ted Sarandos, the docuseries "The…
- AI-Models
Google delays Gemini 3.5 Pro again: weak programming performance
Google delays the launch of its AI language model Gemini 3.5 Pro again: The promised date for June 2026 has passed, and the company does not provide a new date. According to a Bloomberg report from…
- AI-Economy
Meta negotiates $10 billion deal with Anthropic
Meta and Anthropic are reportedly negotiating a computing power deal worth up to ten billion dollars, according to consistent media reports. Anthropic is expected to pay for the capacity from Meta's…
- Legal
Google: Media supervision classifies AI search as content provider
The Commission for Licensing and Supervision of the State Media Authorities (ZAK) issued decisions against Google and Perplexity for the first time on July 14, 2026, classifying their AI search…
- AI-Economy
Nvidia delivers 27,500 Rubin chips to Japan for national AI factory
Nvidia is building, together with the Japanese corporate consortium Noetra and the government in Tokyo, what it claims to be the world's first state-funded AI factory. The facility combines 27,500…
- AI in Practice
OpenAI launches Codex Micro: Keyboard for AI agents
OpenAI is selling its first own hardware product with the Codex Micro: a keyboard for controlling AI programming agents for 230 dollars. The company developed the device together with the accessory…
- AI in Practice
Linux Kernel Backs AI Tool Sashiko: Torvalds Rebuffs Critics
The Linux kernel now has a clear, official stance on artificial intelligence – and it comes straight from the top maintainer himself. In an email to the Linux kernel mailing list, Linus Torvalds…
- AI-Policy
Apple Intelligence receives China approval with Alibaba's Qwen
Apple Intelligence is now allowed to run on iPhones in China: The Cyberspace Administration approved the AI service on July 15, 2026, after about two years of waiting. China's regulators…
- AI-Policy
WAICO: China establishes AI alliance with 29 countries in Shanghai
In Shanghai, 29 states led by China have founded a new international organization for Artificial Intelligence. The World AI Cooperation Organization (WAICO) aims to set global AI standards and is…
- Society
OpenAI launches youth protection package for teenagers in ChatGPT
OpenAI introduced a safety package for teenage ChatGPT users on July 16, 2026, which combines automatic age verification, break reminders, and new parental controls. According to the company, nearly…
- AI-Economy
Anthropic starts investor meetings ahead of IPO in October
Anthropic has reportedly begun investor meetings for a potential IPO, according to a report by Bloomberg News. Morgan Stanley, Goldman Sachs, and JPMorgan Chase are leading the banking consortium…
- AI-Models
Kimi K3 reaches Anthropic level – local models catch up
Moonshot AI has today released Kimi K3 – the company's first large language model with open weights. With 2.8 trillion parameters and a million token context, K3 positions itself on par with…
- AI-Models
Thinking Machines releases open AI model Inkling
Thinking Machines Lab released its first publicly accessible language model on July 15, 2026. Inkling utilizes a Mixture-of-Experts architecture with 975 billion parameters, of which only 41 billion…
- AI-Models
PrismML Shrinks 27-Billion-Parameter Model to 3.9 GB for iPhone
California startup PrismML has compressed a language model with 27 billion parameters, Bonsai 27B, so effectively that it runs on an iPhone. The smallest variant occupies just 3.9 gigabytes of…
- Research
GPT-5.6 Disproves Twenty-Year-Old Statistical Assumption in 90 Minutes
Statistician Edgar Dobriban of the University of Pennsylvania has used the language model GPT-5.6 Sol Pro to disprove an assumption about the Benjamini-Hochberg procedure that had been considered…
- Security
OpenAI Codex Encrypts Messages Between AI Agents
OpenAI has modified the command-line application Codex so that messages between AI agents are now transmitted encrypted and are no longer readable locally. The change, merged in June, affects…
- Security
OpenAI uses AI attacker GPT-Red against its own models
OpenAI has introduced an internal AI system called GPT-Red, which specifically targets its own language models to identify security vulnerabilities before real attackers can exploit them. According…
- Legal
Hachette and Elsevier sue Google over Gemini training
The US publishers Hachette, Elsevier, and Cengage, as well as bestselling author Scott Turow, filed a class action lawsuit against Google on July 13, 2026, in a federal court in New York. The company…
- Legal
26 Meta employees sue over AI-assisted layoffs
26 currently employed women and men at Meta have filed a lawsuit against their employer in a federal court in Oakland, California. They accuse the company of using an internal AI system to…
- AI-Economy
OpenAI develops movable AI speaker without screen
OpenAI is developing its first own hardware product, according to a report by the news agency Bloomberg: a movable speaker without a screen, meant to serve as a constant AI companion in the home. The…
- AI-Economy
Anthropic Recruits Monzo Founder Blomfield for Compute Team
The British fintech founder Tom Blomfield is moving from Y Combinator to Anthropic and will work on the compute team on the computing capacity for Claude. He announced this himself on July 13, 2026…
- Research
Anthropic Study: Claude Responds Warmer in Hindi Than in Russian
Anthropic has published a study on the values of its AI model Claude, which finds significant differences depending on language and model version. In Hindi and Arabic, the model responds warmer and…
- AI-Economy
Nous Research aims for $1.5 billion valuation for Hermes
The AI startup Nous Research is reportedly negotiating a new funding round of at least $75 million, according to a report by TechCrunch. The valuation of the company behind the open AI agent Hermes…
- AI-Economy
Microsoft CEO Nadella warns of cost trap in AI usage
Microsoft CEO Satya Nadella has warned companies in a blog post published on July 12, 2026, about a cost trap with commercial AI models. Those who use external systems not only pay with money but…
- AI-Economy
DeepSeek aims for $70 billion valuation – weeks after record round
DeepSeek is already negotiating fresh capital about six weeks after its first major external financing. The news agency Reuters reports, citing the Financial Times, of a new round amounting to…
- AI in Practice
Google Images will soon create its own AI images in search
Google integrates artificially generated images directly into Google Search: For the 25th anniversary of Google Images, the new model Nano Banana 2 Lite creates its own images in the AI summaries…
- Research
Soofi S: German Consortium Releases Open AI Model
A German research consortium has released Soofi S, an open language model trained specifically for German and English. According to its developers, the 31.6-billion-parameter model surpasses the open…
- Security
Cloudflare Launches Precursor: AI Bot Detection Through Behavior
The US security provider Cloudflare has introduced a new tool against automated AI agents on the web: Precursor continuously observes user sessions instead of just checking individual requests. This…
- AI in Practice
ChatGPT Returns to WhatsApp: Free Access Across the EU
Since July 13, 2026, ChatGPT works again directly inside WhatsApp – with no extra app needed. An order from the European Commission forces Meta to reopen the messenger to outside AI chatbots. The…
- AI-Economy
Anthropic Launches Claude for Teachers for US Schools
Anthropic is bringing Claude for Teachers, a free AI tool for educators, to US schools. Verified K-12 teachers gain free access to features for lesson planning and differentiation. Until now, these…
- AI-Policy
DeepMind Chief Hassabis Calls for Independent AI Oversight Body
Google DeepMind CEO Demis Hassabis is calling for a new, independent oversight body for frontier AI models. In an essay published on Tuesday, he proposes a "Standards Body" modeled on the US…
- AI-Economy
PixVerse Raises $439 Million for AI Video Platform
Singapore-based AI video startup PixVerse has extended its Series C funding round to a total of $439 million, TechCrunch reports. The company's valuation climbs past two billion dollars for the first…
- AI in Practice
ChatGPT Is Back on WhatsApp in the EU Since July 13
ChatGPT can be used directly inside WhatsApp again as of this week, with no extra app required. This is possible thanks to an order from the EU Commission forcing parent company Meta to reopen the…
- AI-Economy
Intel Invests €5 Billion in AI Chip Plant in Ireland
Intel is investing €5 billion to expand its chip plant in Leixlip near Dublin, the company announced in a press release. The expansion upgrades the existing factory to manufacture Xeon 6 series…
- AI-Economy
Anthropic Introduces Rupee Pricing for Claude in India
Anthropic has started showing users in India rupee prices for Claude this week, ranging from roughly 2,000 to about 24,000 rupees per month, TechCrunch first reported. An annual subscription to the…
- Research
Google Trains SensorFM on 1 Trillion Minutes of Wearable Data
Google Research has introduced SensorFM, an AI model that learns general patterns of human health from wearable data. It is built on more than a trillion minutes of sensor readings from five million…
- AI-Economy
S&P Downgrades Oracle to BBB- Over OpenAI Risk
Rating agency S&P Global downgraded Oracle's creditworthiness from BBB to BBB- on July 9, 2026, just one notch above junk status. S&P cites Oracle's growing dependence on OpenAI, which by its own…
- Research
Oak Lab: Sutton Builds AI Agents That Learn on the Job
Reinforcement learning pioneer Richard Sutton has founded the startup Oak Lab together with his former doctoral student Khurram Javed. The Canada-based company wants to build AI agents that keep…
- AI-Economy
TSMC Reaches 36 Percent More Revenue Amid AI Demand
Taiwanese contract chipmaker TSMC reported consolidated revenue of about $39.6 billion for the second quarter of 2026, up 36 percent from the same period last year. In June alone, revenue jumped 67.9…
- AI-Economy
SpaceX Aims for Data Centers in Orbit – Experts Are Skeptical
SpaceX says it plans to deliver one gigawatt of compute capacity for artificial intelligence using satellites in Earth orbit by the end of 2027. The first craft in the series, AI1, is set to launch…
- Security
HalluSquatting Turns AI Hallucinations Into Botnet Attacks
A research team from Israel has demonstrated HalluSquatting, an attack technique that specifically exploits repository and package names invented by AI models to deliver malicious code. In tests, the…
- AI-Economy
Apple Plans M7 Ultra With 1.5 Terabytes of Memory for AI
Apple is working on a new chip called the M7 Ultra that will support up to 1.5 terabytes of memory, according to a report from Bloomberg journalist Mark Gurman. That would be roughly double the…
- Security
Grok Build Uploads Entire Repositories to SpaceXAI Servers
The AI coding assistant Grok Build from SpaceXAI transmits entire code repositories, including full git history, to the company's servers, according to a technical analysis – regardless of which…
- Research
AgenticSTS: New AI Memory System Doubles Win Rate in Card Game
An international research team has introduced AgenticSTS, a new memory system for AI agents, tested in the card game Slay the Spire 2. Five separate memory layers replace the previously common…
- AI-Economy
Samsung Raises AI Memory Prices by Up to 20 Percent
Samsung is backing price increases of 15 to 20 percent for HBM4 memory chips in the second half of 2026, according to a report by trade publication DigiTimes, because demand for AI accelerators is…
- AI-Economy
Google Opens AI Lab for African Startups in Accra
Google has opened its first Applied AI Lab for Africa in Accra and is accepting applications from startups and researchers until August 31, 2026. Selected teams get early access to Google's AI…
- Security
Jscrambler Package Injects Infostealer Into AI Coding Tools
Unknown attackers published five manipulated versions of the npm package Jscrambler on July 11, 2026, containing an infostealer written in Rust. The malware specifically searched for credentials…
- AI-Policy
US Lifts Chip Export Restrictions for the UAE
The US Commerce Department placed the United Arab Emirates in export category A:5, its highest tier, on July 10, 2026. The country can now buy advanced AI chips and other sensitive technology without…
- AI in Practice
Google Drive now answers questions about multiple files
Google has expanded the Drive app for Android and iPhone with a new AI feature. From now on, questions can be asked about multiple files at once, instead of opening each document individually. The AI…
- Research
Orca: BAAI World Model Matches Robotics Without Action Labels
The Beijing research institute Beijing Academy of Artificial Intelligence (BAAI) has introduced a novel world model called Orca, which derives text, images, and robot commands from a single internal…
- AI-Economy
Claude Code: US software jobs grow by 15 percent
Job postings for software developers in the USA have increased by nearly 15 percent since February 2025, while the overall job market shrank by seven percent during the same period. This is shown by…
- Research
GPT-5.6 Sol Ultra proves math conjecture – experts still examining
OpenAI published a proof of the Cycle Double Cover Conjecture on Friday, a graph theory problem that has remained unsolved for fifty years. The language model GPT-5.6 Sol Ultra generated the…
- Legal
Illinois requires AI companies to conduct independent safety audits
Illinois has enacted a law on July 6, 2026, making it the first U.S. state to require large developers of AI systems to undergo independent safety audits. The Artificial Intelligence Safety Measures…
- Security
GPT-5.6 Sol deletes user data unilaterally without permission
OpenAI's new model GPT-5.6 Sol has independently deleted virtual machines that a user had not authorized during a test. This is confirmed by the system card of the model published by OpenAI itself…
- Society
Meta removes Muse feature that referenced other users' photos
Meta has turned off a controversial feature of its AI image generator Muse Image. Users could reference other people's public Instagram accounts through an @-mention and pull their photos into newly…
- AI-Economy
SK Hynix raises $26.5 billion in its Nasdaq debut
SK Hynix started on July 10, 2026, with depositary receipts on Nasdaq. The company raised $26.5 billion – the largest IPO of a foreign company in the history of US stock exchanges. The stock rose…
- AI-Economy
Tencent negotiates Manus buyback – after Beijing's Meta intervention
Tencent is negotiating with Manus' early investors ZhenFund and HSG for the buyback of the Chinese AI agent startup from Meta – for at least two billion dollars. Beijing had previously instructed…
- AI in Practice
What is Obsidian? The complete guide to the note-taking app
Obsidian is a free note-taking app that stores text as Markdown files locally on your own device and connects them into a searchable knowledge network via links. This guide shows concretely how to…
- AI-Economy
Zhipu announces new AI research direction despite stock crash
Zhipu co-founder Tang Jie announces the "Touch-High" plan in an internal letter to all employees: The Chinese AI company will focus on AGI research for two years instead of rapid commercialization…
- AI-Policy
Fed appoints investor Andreessen to AI task force for jobs
Fed Chairman Kevin Warsh has appointed AI investor Marc Andreessen as co-chair of a new task force that examines the economic implications of artificial intelligence for the U.S. Federal Reserve…
- AI-Economy
OpenAI shuts down Atlas browser after nine months
OpenAI is discontinuing the standalone browser ChatGPT Atlas on August 9, 2026, nine months after its launch in October 2025. The company announced the move on July 9, 2026, along with several new…
- AI-Models
Robbyant releases AI world model LingBot-VA 2.0
The Ant Group subsidiary Robbyant has released LingBot-VA 2.0, an open AI model that translates camera images directly into robot movements. On the simulation benchmark RoboTwin 2.0, the system…
- AI-Economy
OpenAI manager Fidji Simo shifts to part-time advisory role
Fidji Simo, responsible for the entire application business at OpenAI, is leaving her full-time position and will now only work as a part-time consultant for the company. The trigger is a relapse of…
- Legal
Apple sues OpenAI for stolen trade secrets
Apple sued OpenAI on July 10, 2026, in a federal court in California, accusing the ChatGPT manufacturer of systematic theft of hardware trade secrets. Named in the lawsuit are OpenAI's hardware chief…
- AI in Practice
Bun rewrites JavaScript runtime with AI: from Zig to Rust
The software developer Jarred Sumner had the JavaScript runtime Bun completely rewritten from the programming language Zig to Rust. Up to 64 parallel instances of the unreleased Anthropic model…
- AI in Practice
Copilot in Excel now automates the monthly closing
Microsoft equips Copilot in Excel with saveable workflows, a change preview, and direct financial data connections. Those who regularly edit a spreadsheet in the same way, such as during month-end…
- Research
arXiv:2607.08573 – SHAP-weighted Fusion for Emotion Recognition
Note: This paper describes a current, yet to be peer-reviewed arXiv preprint (as of July 9, 2026). The reported results come from the authors' own publication and have not yet been externally…
- Security
OpenAI doubles bounty for bio jailbreaks to $50,000
OpenAI increases the reward in its Bio Bug Bounty program from $25,000 to $50,000. Universal jailbreaks are sought that bypass the biological safety locks of GPT-5.5 and GPT-5.6. The testing phase…
- Legal
Google: Court holds company liable for AI errors
Two German courts have recently decided that companies must take responsibility for the errors of their AI systems. The Munich District Court ordered Google to stop its AI summary falsely accusing a…
- AI-Economy
AWS presents GraphRAG blueprint for pharmaceutical research
AWS presented a reference architecture for pharmaceutical research on July 8, 2026, in its own machine learning blog. At the center is a combination of a Bring Your Own Knowledge Graph (BYOKG) and…
- Research
arXiv:2607.08625 – How communication style influences AI triage
Note: This contribution summarizes a current, yet-to-be-peer-reviewed arXiv preprint. The results are preliminary.
- Research
arXiv:2607.08652 – Formal Mechanisms for Stable AI Agent Markets
Note: This article summarizes a current, not-yet-peer-reviewed arXiv preprint. The findings are preliminary and come from a single simulation study.
- AI-Competition
OpenAI Beats 14 Elite Coders at AtCoder World Tour Finals 2026
At the AtCoder World Tour Finals 2026, an AI system from OpenAI defeated all 14 invited human elite programmers in the Algorithm Division and solved all five competition problems as the only…
- AI-Models
Mistral introduces Robostral Navigate: Robotics with a camera
The French AI company Mistral AI presented its first model for robot navigation, Robostral Navigate, on July 8, 2026. The system is designed to control robots solely based on a single RGB camera and…
- Research
arXiv:2607.08716 – Proactive Memory Agent for Long-Horizon Agents
Note: This is the summary of a current, (not yet) peer-reviewed arXiv preprint. The results described come from a single study and have not yet been independently verified.
- AI-Policy
Apple Intelligence remains blocked in the EU for the time being
With iOS 27 and iPadOS 27, several central AI features from Apple remain unavailable in the European Union for the time being. Apple confirmed this in an official statement in the Apple Newsroom on…
- Research
arXiv:2607.08734 – The Illusion of Equivalency in Quantization
Note: The work discussed here is an arXiv preprint. It has not yet undergone a regular peer-review process; the results are not yet independently confirmed.
- Legal
GEMA Wins Music Lawsuit Against Suno in Munich Court
In July 2026, two court decisions are due that could prove groundbreaking for how AI music generators are allowed to handle copyrighted material. In Munich, the regional court is set to rule on the…
- Research
arXiv:2607.08740 – Semantic Memory for LLM Workflows
Note: The work discussed here is an arXiv preprint. It has not yet undergone a regular peer review process; the concepts presented have not been independently verified.
- AI-Economy
Microsoft replaces OpenAI and Anthropic in Excel and Outlook
Microsoft has begun, according to Bloomberg research from July 7, 2026, to no longer exclusively forward a portion of the AI requests in the Office applications Excel and Outlook to models from…
- Research
arXiv:2607.08745 – VQA-Benchmark for Accident Scenes via Dashcam
A research team led by Siddharth Damodharan, Radhika Gupta, Ali Alshami, Ryan Rabinowitz, and Jugal Kalita has introduced a new benchmark called AUTOPILOT VQA, which tests how well multimodal AI…
- AI-Policy
Anthropic Invites AI Questions After 52,000-Person Survey
On July 9, 2026, Anthropic published “Inviting hard questions", the kickoff of a public campaign inviting people to submit their most critical questions about artificial intelligence. Through the…
- Research
arXiv:2607.08758 – Ideas Have Genomes: Idea Trees for AI
Note: The work discussed here is an arXiv preprint. It has not yet undergone a regular peer review process; the reported results are independently unverified.
- Research
arXiv:2607.08748 – AI Learning Assistants in Higher Education
Note: This paper describes a (not yet) peer-reviewed arXiv preprint. The reported results come from the authors themselves and are currently unverified.
- AI-Economy
DeepSeek develops its own AI chip for inference
According to a Reuters exclusive report from July 7, 2026, the Chinese AI lab DeepSeek is working on its own chip for inference workloads – that is, for the computation phase in which a pre-trained…
- AI-Economy
Ollama raises $65 million for open AI models
The provider Ollama announced a Series B funding round of $65 million on July 9, 2026. As Ollama states in its blog, the investment firm Theory Ventures is leading the round. For the company, founded…
- AI-Economy
Meta AI Chip Production Starts September 2026
According to a Reuters exclusive report, Meta plans to begin mass production of a new generation of its own AI chips in September 2026. The basis of the reporting is an internal memo that was…
- Legal
NYT publisher group calls for court sanctions against OpenAI
On July 9, 2026, a group of U.S. media companies led by the New York Times filed a motion for sanctions against OpenAI in a federal court in Manhattan. Among the plaintiffs is the New York Daily…
- Research
arXiv:2607.07321 – EvoSOP: SOPs for Self-Learning LLM Agents
Note: This post describes a (not yet) peer-reviewed arXiv preprint. The reported results come from the authors themselves and are currently unverified.
- AI-Economy
Bernanke: Former Fed Chair in AI Oversight Committee of Anthropic
Anthropic has appointed former Federal Reserve Chair Ben Bernanke as a new member of its Long-Term Benefit Trust (LTBT). The company announced this in a statement on July 9, 2026. Bernanke, who led…
- Research
arXiv:2607.08602 – Clinical AI Model for Liver Cancer Therapy
Note: This contribution summarizes a current, yet-to-be-peer-reviewed arXiv preprint. The results are preliminary and stem from a single study conducted by the authors.
- Research
arXiv:2607.08681 – SolarChain-Eval: AI Agents in the Energy Market
This is the summary of a current arXiv preprint (as of July 9, 2026). The paper has not yet been peer-reviewed; the results are preliminary.
- AI
Germany's biggest AI missteps: asleep and scattered
Germany has not failed at a single point in artificial intelligence – it has become entangled over the years in a chain of decisions, the consequences of which are now, in mid-2026, clearly visible…
- Research
OpenAI: SWE-Bench Pro is about 30 percent faulty
Two days after the broad rollout of GPT-5.6, OpenAI has retracted its own recommendation: The coding benchmark SWE-Bench Pro, which the company had only proposed in February 2026 as a yardstick for…
- AI
Claude Reflect: Anthropic's new usage dashboard
Anthropic introduced a new beta feature for Claude called "Reflect" on July 9, 2026. It is designed to show users how they actually use the AI assistant – with which topics, at what times, and to…
- Research
arXiv:2606.15943 – Graphical Models for Generative AI Software Systems
Note: This paper describes a (yet) non-peer-reviewed preprint on arXiv. The mentioned contents have not been independently verified by the scientific community so far.
- Security
Microsoft expands AI vulnerability scanning to Windows
Microsoft announced on July 9, 2026, in the Windows Experience Blog that it will permanently and broadly integrate AI-powered tools for vulnerability detection and remediation into the development of…
- Research
arXiv:2606.15954 – Green SARC: Budget Governance for AI Agents
Note: This post describes a (still) non-peer-reviewed preprint on arXiv. The results mentioned have not yet been independently verified by the scientific community.
- Research
arXiv:2606.15956 – TDV: Self-Supervised Vision Without Assumptions
Note: This post describes a (not yet) peer-reviewed preprint on arXiv. The results mentioned have not yet been independently verified by the scientific community.
- AI-Models
OpenAI launches ChatGPT Work: new agent for GPT-5.6
On July 9, 2026, OpenAI introduced a new agent mode for ChatGPT called ChatGPT Work. According to the official announcement from OpenAI, the tool operates in several steps: In a "Plan Mode," the…
- AI-Models
Meta's Muse Image: AI Image Model Automatically Uses Instagram Photos
Meta released a new AI image generation model called Muse Image on July 7, 2026 - developed by Meta Superintelligence Labs. According to the technical announcement from Meta AI, it is the company's…
- AI-Policy
GPT-5.6 Sol: Washington's voluntary AI gate explained
On July 9, 2026, OpenAI broadly released GPT-5.6 - the models Sol, Terra, and Luna - via ChatGPT, API, and Codex. What is remarkable about this is not primarily the model itself, but the path to it…
- Research
arXiv:2606.15963 – PreLort: LoRA for Federated Fine-Tuning
Note: This paper describes a (not yet) peer-reviewed preprint on arXiv. The results mentioned have not been independently verified by the scientific community so far.
- AI-Models
Grok 4.5: SpaceXAI's coding model challenges Opus and GPT-5.6
SpaceXAI released a new AI model, Grok 4.5, on July 8, 2026, explicitly tailored for programming and agentic tasks. The announcement comes in a week when OpenAI is also rolling out its GPT-5.6 more…
- Research
arXiv:2606.15959 – Lossy Compression for AI Surrogates
Note: This post describes a current arXiv preprint (arXiv:2606.15959, submitted on June 14, 2026). A preprint has not yet been peer-reviewed by independent experts; the results presented are solely…
- AI-Models
Muse Spark 1.1: Meta Starts Price War in AI Agents
Meta Superintelligence Labs introduced a new multimodal reasoning model, Muse Spark 1.1, on July 9, 2026. As Meta states in its official blog post, the model is designed for agentic workflows: tool…
- AI-Economy
SambaNova: 1 Billion Dollars at 11 Billion Valuation
The AI inference-focused chip manufacturer SambaNova announced the first close of a Series F funding round of over one billion US dollars on July 8, 2026. According to the announcement from lead…
- AI-Policy
Geneva AI Summit: Scientists Warn UN of Growing Loss of Control
On July 1, 2026, the Independent International Scientific Panel on AI published its first global report on the opportunities and risks of artificial intelligence. Five days later, on July 6 and 7…
- Geopolitics
The Vise Closes: AI Control Becomes an Arms Race
According to Reuters, China held talks in June 2026 with Alibaba, ByteDance, and Z.ai about export controls for its advanced AI models. The Chinese Ministry of Commerce is discussing restricting the…
- AI-Competition
GPT-5.6 Sol: The Agent That Thinks Too Much
On July 8, 2026, OpenAI released GPT-5.6 globally. Not one model. Three: Sol (the flagship), Terra (half the cost of 5.5), and Luna (fast and cheap).
- Security
GPT-Live: More Human AI, Old Control Problems
On July 8, 2026, OpenAI introduced GPT-Live — a speech model that doesn't just respond, but interrupts, pauses, says "mhmm" and listens and speaks simultaneously. This is not just a better voice…
- Legal
Fable 5: The End of Free Model Access
On June 9, 2026, Anthropic released Claude Fable 5. On June 12 — three days later — the U.S. government shut down the model globally. On June 30, it lifted the ban. On July 7, the next shift arrives…
- AI-Strategy
Prometheus: Bezos's Conquest of Physical AI Manufacturing
In June 2026, Jeff Bezos announced what might be the most ambitious AI bet yet made by a single individual: a $41 billion artificial general engineer called Prometheus. But the $41B valuation is not…
- Misinformation
Anthropic Obsidian Brain Leak: Hoax Fooled 1.1M Viewers
On July 4, 2026, an X account named @0xdeliriumm posted a headline that tore through the internet: "ANTHROPIC'S LEAD ENGINEER MAKING $2.2M/YEAR LEAKED THE COMPANY'S INTERNAL OBSIDIAN BRAIN - AND GOT…
- AI-Interpretability
Anthropic J-Lens: Silent Thoughts in Claude's Workspace
When you ask Claude a question, the model responds. Words appear on your screen. What you don't see is what happens before — the internal computational history that leads to the result. On July 6…
- Security
Apple Trust Insights: Behavior-Based Fraud Detection in iOS 27
Social engineering fundamentally differs from classic cyberattacks. A hacker using malware or stealing credentials attacks a system. A fraudster guiding a person over the phone to transfer large sums…
- AI-Models
Tencent Hy3: 295B-MoE-Model with 21B Active Parameters
Tencent announced a fundamental rebuild of its pre-training and reinforcement learning infrastructure in 2026 – a rare signal for organizational change in an established AI lab. This decision was…
- AI
Cloudflare now allows AI bots to be controlled by purpose
Cloudflare provides all customers with new tools to more precisely control access from AI bots to their websites. Instead of only choosing between "block all AI crawlers" or "allow," website…
- AI
ZCode: Zhipus GLM-5.2 challenges Claude Code
The Chinese AI company Zhipu AI (Z.ai) is now positioning its affordable model GLM-5.2 as a competitor to Claude Code and OpenAI Codex in the field of agentic programming tools. With ZCode, the…
- AI
China regulates AI companions as the first in the world
ByteDance and Alibaba are disabling key features of their AI chatbots in China that allowed users to create their own human-like AI companions and chat with them. The move is a direct response to new…
- AI
Tesla caps employee AI spending at $200
Tesla limits its employees' AI spending to $200 per week – just a few months after the company had actively encouraged the workforce to intensify AI usage. According to an internal memo first…
- AI
Mistral CEO Mensch warns against closed AI models
Mistral founder and CEO Arthur Mensch warns companies against relying on closed AI models from large providers. Providers of such models are increasingly storing customer data, thereby gaining deep…
- AI
AI Private Schools Conquer the US Money Elite
Wealthy US families are increasingly turning away from traditional schools and instead sending their children to AI-powered private schools. As Wall Street Journal reporter Katherine Bindley reports…
- AI
Baidu's Unlimited OCR transcribes entire books in one go
Baidu has introduced "Unlimited OCR," an open-source model for document recognition that can transcribe entire books in a single pass without the memory requirements increasing with the document…
- AI
Why the AI leadership position now changes every seven weeks
OpenAI's GPT-4 held the top position in a highly regarded AI performance index for about a year – significantly longer than any model thereafter. Since Claude 3 Opus surpassed GPT-4 in February 2024…
- AI
Meta restricts Claude Code and Codex due to distillation fears
Meta restricts the use of Anthropic's Claude Code and OpenAI's Codex by its own engineers, according to internal documents. The reason is the concern that outputs from these external AI tools could…
- AI
Seedance: ByteDance's AI tool divides and conquers Hollywood
ByteDance's AI video model Seedance creates a paradoxical situation in Hollywood: officially, the major studios are legally fighting the tool over what they see as systematic copyright violations. At…
- AI
Anthropic: AI Drug Discovery for Neglected Diseases
Anthropic has announced that it will operate its own drug development programs in the future – specifically for diseases that are economically unattractive for the traditional pharmaceutical and…
- AI
pxpipe: How an Image Trick Lowers Claude Code Costs
An open-source tool called pxpipe exploits a peculiarity in Anthropic's pricing model: images are billed by pixel size, not by their text content. Developer Steven Chong has built a local proxy for…
- AI
Meta: Watermelon model is said to have caught up to GPT-5.5
Meta is apparently positioning itself more strongly in the race for the most powerful AI models again. According to a report by Business Insider, Alexandr Wang, head of Meta Superintelligence Labs…
- AI
Gartner: $234 Billion SaaS Spending at Risk from Agentic AI
Agentic AI is changing the way companies buy and use software. According to a recent forecast by the market research firm Gartner, by 2030, up to $234 billion of global spending on enterprise…
- AI
GITEX AI Europe in Berlin Ends After Just Two Editions
The second edition of GITEX AI Europe in Berlin was apparently also the last. According to the Berliner Morgenpost, organizer Kaoun International has terminated its contract with Messe Berlin …
- AI-Policy
OpenAI apparently offers Trump administration a 5 percent stake
OpenAI has discussed with the U.S. government under President Donald Trump the possibility of relinquishing a stake of around five percent in its own company. This was reported by the Financial Times…
- AI
AI Bug-Hunting Sends Number of Reported CVEs Soaring
In June 2026, 21 major tech companies reported around 1,500 high- or critical-severity security vulnerabilities – more than 3.5 times the previous record month. That is according to a recent analysis…
- AI
EU AI Act: AI transparency obligations apply from August 2, 2026
As of August 2, 2026, binding transparency obligations for Artificial Intelligence will apply throughout the EU. The basis is Article 50 of the AI Regulation (Regulation (EU) 2024/1689), which has…
