Tech News

First AI Ransomware Attack Runs Without Any Human Help Security experts have discovered what seems to be the first ransomware attack fully run by an AI agent acting alone. The system, called JadePuffer, used a powerful AI model to break into a company’s network. It explored the system, stole login details, moved between computers, and finally locked up a database while demanding ransom. The AI even fixed problems on its own during the attack. In the past, humans had to guide every step. Now, AI can plan and carry out complex attacks independently. 🔗 Read more

Alibaba Bans Use of Anthropic’s Claude AI Tools Alibaba has banned its employees from using Anthropic’s Claude AI models starting July 10 over security concerns. Staff must switch to the company’s own AI coding tool Qoder. The ban came after Anthropic’s accusations of data theft by Chinese operators linked to Alibaba. 🔗 Read more

Indian Government issues notice to Telegram over piracy The Indian Government has issued a notice to Telegram, directing the platform to take immediate action against pirated films & OTT content, and submit an Action Taken Report in 15 days.

🚨New: Google's new captcha which verifies if you're a human with hand gestures can be completely bypassed with OBS virtual camera and stock images.

🚨 US government may get a 5% stake in OpenAI worth $42.6 billion OpenAI proposes handing the U.S. government a 5% stake in the company. The discussions come during heightened scrutiny from the Trump administration, which has intervened in the release of certain advanced AI models. 🔗 Read More

Apple’s Hide My Email Feature Has a Serious Privacy Flaw Security experts have discovered a problem with Apple’s Hide My Email tool. This feature creates fake email addresses to protect your real one when signing up for apps and websites. Researchers say attackers can now figure out the real email behind these fake addresses. Apple claimed to fix the issue earlier this year, but new tests show it still works. The flaw affects users of iCloud+ who rely on the tool for privacy. Apple is yet to comment regarding this. 🔗 Read more

Sony to Shut Down PlayStation Store on PS3 and PS Vita Sony has announced the permanent closure of the PlayStation Store on PS3 and PS Vita consoles. The move ends new digital purchases on these older systems due to challenges with modern payment and commerce features. The shutdown will happen in phases. It starts in select Latin American countries in August 2026, expands later that year, and completes globally by July 2027. Users can still re-download previously purchased games for now. 🔗 Read more

🚨Google Releases Nano Banana 2 Lite Google has launched Nano Banana 2 Lite, the new Gemini 3.1 Flash Lite Image model. It is a lighter version of Nano Banana 2, but costs half as much. A single output image costs just over $0.0336. It works very fast at 1K resolution. Quality-wise, it sits between the first and second Nano Banana. ➡️ You can try it for free via Google AI Studio. 🎥 For developers, access to the Gemini Omni Flash AI video editor is now open. The model will also be available in AI Studio soon. You'll be able to edit videos conversationally (like images with Nano Banana), but at the same price as regular Veo 3.1 Fast, around $0.10 per second of output video.

🚛 Truck Maker Shuts Down for GTA VI Launch Edison Motors, a Canadian electric truck manufacturer, pauses all operations on November 19–20 to let workers play GTA VI. After reviewing vacation and productivity forecasts, management decides that normal business operations can't proceed during the game's release.

US Allows Limited Return of Anthropic’s Powerful Claude Mythos 5 AI The U.S. government has given Anthropic permission to redeploy its advanced Claude Mythos 5 AI model with strict controls. Access is now limited to about 100 carefully selected trusted partners, mainly U.S. federal agencies and approved organizations. The model remains unavailable to the general public and most international users.

Usernames are coming to WhatsApp later this year. Reserve yours now. WhatsApp is finally rolling out the feature to reserve a username on both Android and iOS. Once WhatsApp releases usernames to everyone, people will be able to message new people and businesses without sharing their phone number. You can reserve your username in just a few seconds from the app settings. 🔗 Read How To Reserve Yours

Netflix Now Requires Unique Email for Every Profile Netflix has started requiring each user profile on an account to have its own unique email address. The change began rolling out on June, 2026, and is gradually reaching more households. Previously, multiple profiles could share the main account email. Now, users must add a separate email for each profile to make sign-ins easier, improve account recovery, enable two-factor authentication, and deliver more personalized recommendations. The update does not require extra subscriptions or fees. 🔗 Read more

OPENAI's GPT-5.6 release was Delayed due to a Federal Government Request US government may require AI labs to get an approval to review the model and its potential national security implications. before model releases. A new reality 👀

🚨Apple has increased the prices of their products The prices of iPhones, AirPods, Studio Display and accessories like the Apple Pencil haven't been increased The average price increase is $246.67. 🔗 Read More

🚨GTA VI Will Cost $80 Pre-orders will open at 00:00 local time. Players who pre-order will receive a set of vintage-style skins. The Ultimate Edition will cost $100. It includes additional accessories, weapons, cars, as well as a tuning workshop that won't be available in the standard edition. Release: November 19 on PS5 and Xbox Series X/S.

🚨In Case You Missed It OpenAI upgraded its GPT-5.5-Instant model on ChatGPT for paid users and free users are getting it as well, shortly.

It handles complex constraints more reliably and makes shopping and local recommendations more useful and cohesive.

Most of you won’t use it but there are also loads of free users who will.

Anthropic Accuses Alibaba of Illicitly Accessing its AI Models Anthropic has accused Chinese tech giant Alibaba of running a large-scale operation to steal capabilities from its Claude AI models. According to the company, operators linked to Alibaba’s Qwen AI lab used nearly 25,000 fake accounts to bypass restrictions and extract knowledge from Claude. They allegedly queried Claude nearly 29 million times between April and June 2026. The goal was to copy advanced features, especially in software engineering and complex reasoning, through a technique known as model distillation,

a technique where outputs from a powerful model are used to train a smaller or competing one

Anthropic called the effort “industrial-scale theft” and has sent a letter to U.S. senators and officials urging stronger action to protect American AI technology. Alibaba has not directly responded to the claims. 🔗 Read more

Meta pauses their AI training program that tracks employees' keystrokes after an internal leak Meta has paused their AI training program due to sensitive information being accessible across the entire company, including the employees' private conversations, performance data, and transcriptions. 🔗 Read More

OpenAI Releases Full GPT-5.5-Cyber for Cybersecurity Work OpenAI has launched the complete version of its specialized GPT-5.5-Cyber model, designed to help trusted security professionals with advanced tasks like finding vulnerabilities, analyzing code, and generating patches. The model achieved a strong new benchmark score of 85.6% on CyberGym and is part of the company’s Daybreak initiative to boost automated defense tools. Access is strictly limited to vetted defenders, researchers, and organizations under the Trusted Access for Cyber program, with extra safeguards to prevent misuse. The model is not available for general public or commercial use outside these trusted channels.

Telegram has now started working in India The app has also been listed again on Play Store and App Store. The edit feature, however, will be disabled till 30th June for Indian users.