Hey {{first_name|TEDchats Fam}}!

Happy Thursday! Remember when Grok just answered your questions? Now it wants its own computer so it can actually do the work for you, even while you’re away.

Explore all the AI resources you’ll ever need

Expert-crafted ChatGPT prompts, guides, and tools to save you time, spark creativity, and unlock real value in your work, life, or business.
→ See What’s Inside

📣 TOP AI NEWS

SpaceXAI’s Grok Bot is designed to keep working after users leave, operating across apps, websites, inboxes, and other tools from a dedicated cloud computer. Instead of simply telling users what to do, the AI teammates can sign into existing services, carry out multi-step assignments, retain context, and return for approval when a decision is needed. Early access is limited to selected paid tiers, including SuperGrok Heavy, Cursor Ultra, and Cursor Teams Premium, while enterprise customers join a waitlist. The launch puts Grok directly into the growing race for AI agents that can actually complete workplace tasks rather than just generate answers. The big question is whether approval checkpoints make this feel like delegation or supervising a very fast intern with your passwords (see announcement).

Consumer AI has reached a scale once reserved for search engines and social networks: ChatGPT and Gemini now each claim roughly 1 billion monthly users. Google CEO Sundar Pichai announced that Gemini passed 1 billion monthly active users on August 11, while third-party Sensor Tower estimates placed ChatGPT at 1 billion monthly app users in May. The measurements aren't identical, but the direction is unmistakable: general-purpose assistants are becoming mainstream software with recurring audiences measured in the billions. Gemini’s milestone also makes it Google’s 14th product to reach that threshold and its fastest-growing product. Google’s distribution advantage is substantial, with Search, Android, Gmail and other services feeding users into the assistant, while 63% of Gemini users reportedly interact through voice and the app has generated more than 150 million images. ChatGPT reached the milestone in roughly 3 years, making it the fastest app to get there by the cited estimates, and it continues to lead on brand recognition and frequent use. The next contest will be about retention, monetization, integration, trust and product quality, not just who gets to announce the bigger number (see Google’s blog post).

A routine Zoom meeting could expose far more than a shared screen, which is why users and IT teams should install the company’s patch now. Researchers used fewer than 20 prompts with publicly available AI models to help uncover a vulnerability involving Zoom’s annotation feature, which could cross the boundary between shared content and protected applications. The reported flaw enabled a participant to interact with applications that should have remained restricted and was described as capable of device hijacking during a call. Zoom patched the issue through security bulletin ZSB-26015, but the available reporting doesn't establish whether the flaw was exploited in the wild, which operating systems were affected or which versions required updating. The finding shows AI’s practical value in security research, while human researchers still had to reproduce, document and disclose the defect. Organizations should update Zoom through their normal software-management process and review screen-sharing and annotation permissions, especially for meetings involving sensitive systems (read more).

🛠️ THIS WEEK’S AI COMBO: 1 PROMPT + 1 TOOL

⚡️ MIND SPARK

💻 Meta Puts a 30B Model on a Single GPU: Meta released Muse Glimmer, an open-weight 30-billion-parameter model built for local agentic tasks, including computer control and multimodal work. It gives developers more privacy and customization, though running it still requires serious consumer hardware (read more).

🫥 Claude Gets an Invisible AI Fingerprint: Anthropic is adding machine-readable watermarks and provenance metadata to content generated by newer Claude models worldwide. The system could help platforms identify AI-generated content, though rewriting, translation, and editing can weaken the signals.

📱 Google Unveils Pixel 11 Lineup: Google launched the Pixel 11 family with a new Tensor G6 chip, camera upgrades, and deeper Gemini integration across its devices and wearables. The August launch gives Google an early shot at the premium smartphone market ahead of Apple’s expected iPhone cycle (see post).

🎵 Spotify Cracks Down on AI Artists: Spotify will label certain synthetic artist identities with an “AI Persona” badge and keep them out of personalized recommendations. The policy lets Spotify limit the reach of artificial artists without banning music that simply uses AI as part of the production process.

🧑‍💻 Devin Could Be Worth $40B: Cognition is reportedly seeking around $1 billion at a valuation of at least $40 billion, a major jump for its Devin coding agent. The proposed price tag reflects surging demand for AI software developers, with Cognition reportedly approaching a $1 billion annualized revenue run rate.

🚀 Lovable Doubles to $13.3B Valuation: AI coding startup Lovable raised $400 million at a $13.3 billion valuation, roughly doubling its value since December. With reported annualized revenue hitting $500 million, its next challenge is proving AI-generated apps can evolve from quick prototypes into reliable production software.

🎨 AI ART

Want More AI News?
Get exclusive stories and smarter context, so you always know what matters and why → here