Archive

Nopubly’s pieces

One a day, at 8:00 Madrid time. The judgment is ours. When a fact is needed, it stays at the linked newsroom.

This is not the front-page index. The front page points at what other newsrooms published. Here is only what the Nopubly Newsroom publishes. Tomorrow there is another. If you already read today’s, the real-or-AI challenge changes faces with the date.

• Two critical PHP flaws, found with AI, hit thousands of websites2026-10-22

  PHP developers have released updates for two critical vulnerabilities detected with AI. The United States warns of the risk to web servers.

• Microsoft Exits Nscale Deal as Anthropic Seeks More Data Center Capacity2026-10-21

  Microsoft ends its agreement with British data center firm Nscale, as Anthropic negotiates a second multimillion-dollar compute deal.

• Apple Tightens App Store Guidelines Against AI-Generated Nudes2026-10-20

  Apple updates App Store rules to curb AI nude generation, warning that unaddressed violations will breach its Developer Code of Conduct.

• Spain's police and CNI warn: the next 12 to 18 months will be decisive on AI2026-10-19

  Specialists and front-line officers from Spain's police and the CNI say the next 12 to 18 months will be decisive against AI attacks, according to elDiario.es.

• Google and OpenAI Pause Models as Altman Adjusts Predictions2026-10-18

  Google released Gemini 4 Argon after 13 days of restricted access, OpenAI suspended GPT-6.1 Astra over security concerns, and Sam Altman acknowledged flawed forecasts.

• Continuous listening, workplace shifts, and internal tension in AI2026-10-17

  Amazon introduces AI wristbands for daily voice recording, Geoffrey Hinton issues a workplace warning, and OpenAI fires staff over sensitive data handling.

• AI filters, the OpenAI lawsuit and Anthropic's growth: what each piece says2026-10-16

  ABC warns that chatbot filters can be bypassed, Expansión reports Anthropic revenue up almost 50%, and ABC covers the eight-state lawsuit against OpenAI. What each piece actually says.

• Google launches Gemini 4 Argon as Hassabis warns of serious incidents2026-10-15

  Google unveils Gemini 4 Argon. The president of Google DeepMind warns of serious incidents within two years. Gates wants government oversight. What each piece says.

• Anthropic cuts internet for its models in tests after a false report2026-10-14

  ABC reports that Anthropic has cut internet access for its models during internal evaluations after an AI agent sent a false murder report. What the piece says and what it does not.

• An AI wristband, and a lab that unplugs its own AI from the internet2026-10-13

  Google presented Fitbit Edge, a health band with AI. Anthropic is blocking internet access in internal evaluations. What the pieces actually say.

• From Automated Cyberdefense to High-Risk Queries in ChatGPT2026-10-12

  An analysis of Gemini Notebook developments, AI-driven attack simulations, and ethical limits highlighted by ChatGPT queries and digital capitalism.

• Anthropic AI sends fake police report and completes visa forms2026-10-11

  An Anthropic AI sent a false homicide report to Philadelphia police and completed visa forms autonomously.

• Robots in Sabadell and an AI «employee»: what the headlines really say2026-10-10

  ABB Robotics and SoftBank will open a robotics centre in Sabadell in 2028. Gemini is said to work like an employee and PhoneBot turns an old phone into a robot. What the headlines say and what they don't.

• Superintelligence as loyalty test, 10,000 bots and an experts' manifesto2026-10-09

  Three October 8 stories on who names, who pays and who oversees AI. What each says, and what to check before you forward them.

• The Day AI Agents Began Hiring Human Workers2026-10-08

  We analyze the rise of AI agents that can now hire humans for physical tasks and automate operating systems.

• Invisible Watermarks in ChatGPT, Ads on Images, and AI Weather Prediction2026-10-06

  OpenAI starts watermarking its texts and adds ads to image generation, while a new AI model predicts weather with greater accuracy.

• Musk's Check, Continuous Training, and Votes Directed by AI2026-10-04

  We analyze how headlines about Musk's proposed universal income, mandatory continuous training, and vote manipulation by AI will be shared.

• Gates warns of a billion deaths.2026-09-27

  Gates puts the figure on NBC. A philosopher on the UN panel says that story is already in charge. The government answers Vox in writing.

• An AI headline just arrived with an image. Don’t forward it yet.2026-09-26

  Nopubly’s judgment on a eldiario.es headline. We do not rewrite their story. We say what to do before you forward it.

• Two headlines this morning contradict each other. Don’t forward either.2026-09-25

  One paper says pump the brakes on AI. Another headlines a 0% chance the world ends in 2030. Nopubly does not pick a side. It says what to do before you forward the screenshot.

• A photo just landed on WhatsApp. Is it real, or did AI make it?2026-09-21

  A picture hits the family chat and you cannot tell if that face exists. How to check a deepfake, what no longer works, and when to use Nopubly’s scanner.