Good morning. It’s Thursday, February 5. Today we are covering:
Russian troops on front lines lose access to Starlink internet terminals, Ukraine says
Google’s Gemini app has surpassed 750M monthly active users
GitHub adds Claude and Codex AI coding agents
Newsletter platform Substack notifies users of data breach
Britain to work with Microsoft to build deepfake detection system
Let’s dive in
Russian troops on front lines lose access to Starlink internet terminals, Ukraine says
By Vaughn Cockayne via Washington Times
Ukraine says Russian front-line units lost access to Starlink, after Kyiv and SpaceX coordinated steps to disrupt Moscow’s battlefield communications while keeping Ukrainian service online.
Ukrainian officials describe a “white list” approach that verifies and prioritizes registered Ukrainian Starlink terminals, enabling targeted deactivation of unauthorized Russian-used terminals rather than a blanket blackout.
Ukrainian sources claim the shutdown caused an immediate command-and-control breakdown for Russian forces, pausing some assault operations, though some Ukrainian terminals were briefly impacted if they weren’t properly registered.
𝕏: 1/ Russian forces in Ukraine are experiencing a devastating loss of connectivity as the Starlink terminals they rely upon are systematically shut off. With Russia’s own Starlink alternative years away from implementation, Russian warbloggers say the army is in crisis. - ChrisO_wiki (@ChrisO_wiki)
Google’s Gemini app has surpassed 750M monthly active users
By Lauren Forristal via TechCrunch
Google’s Gemini app has crossed 750 million monthly active users, up from 650 million last quarter, underscoring how quickly the chatbot is scaling via both free and subscription usage.
The figure puts Gemini ahead of Meta AI’s ~500 million monthly users, but still behind ChatGPT’s estimated ~810 million MAUs in late 2025, keeping the competitive race tight at the top end of consumer AI.
Google credits the momentum to the launch of Gemini 3 and broader AI demand, noting its first-party models now process 10+ billion tokens per minute via direct API use, alongside record usage in Search.
𝕏: Very proud to see the progress across so many areas. Great to see @GeminiApp hit 750M monthly active users, and the use of our Gemini models across many products and Cloud surfaces is strong: 10B tokens/minute is 166M tokens/sec (TPUs!), or ~1750 tokens/person on earth per day. - Jeff Dean (@JeffDean)
The best way to reach new readers is through word of mouth. If you click THIS LINK in your inbox, it’ll create an easy-to-send pre-written email you can just fire off to some friends.
GitHub adds Claude and Codex AI coding agents
By Tom Warren via The Verge
GitHub is rolling out Anthropic’s Claude and OpenAI’s Codex as built-in AI coding agents across GitHub, GitHub Mobile, and VS Code, starting in a public preview for Copilot Pro Plus and Copilot Enterprise subscribers.
The integration is part of Agent HQ, letting developers pick between Copilot, Claude, Codex, or custom agents to work on tasks, including assigning agents to issues and pull requests, with each run consuming a premium request.
GitHub says the goal is less context switching and more side-by-side comparison of how different agents solve the same problem, and plans to expand access and add more agents from partners like Google, Cognition, and xAI.
𝕏: GitHub is adding Anthropic’s Claude and OpenAI’s Codex AI coding agents to GitHub, GitHub Mobile, and Visual Studio Code. Microsoft-owned GitHub continues to embrace OpenAI and Anthropic AI advances. - Tom Warren (@tomwarren)
Newsletter platform Substack notifies users of data breach
By Sergiu Gatlan via BleepingComputer
Substack says attackers accessed limited user data in October 2025, including email addresses, phone numbers, and other internal metadata, but not passwords, credit card numbers, or other financial information.
CEO Chris Best told users the company only detected evidence of the issue on February 3, 2026, claims the underlying flaw has been fixed, and urged extra caution around phishing attempts using the stolen details.
A threat actor allegedly posted a database on BreachForums with 697,313 records, though Substack has not disclosed how many users were affected or exactly how access was obtained.
𝕏: Substack just emailed me saying my email and phone number have been leaked in a data breach - Shannon Liao (@Shannon_Liao)
Britain to work with Microsoft to build deepfake detection system
Reuters
The UK government says it will work with Microsoft and researchers to build a system that can detect and respond to deepfakes online, amid growing concern about AI-manipulated media.
Technology minister Liz Kendall warned deepfakes are increasingly being used by criminals to scam the public, exploit women and girls, and erode trust in what people “see and hear.”
Government figures cited alongside the initiative show the scale of the problem: deepfakes shared online are projected to jump from about 500,000 (2023) to roughly 8 million (2025).
Trending in AI
Thanks for reading to the bottom and soaking in our Newslit Daily fueled with highlights for your morning.
I hope you found it interesting and, needless to say, if you have any questions or feedback let me know by hitting reply.
Take care and see you tomorrow!
How was today’s email?
