
Welcome Automaters, 👋
We need to talk about what "temporary" actually means to Google and other AI companies right now.
If you’re the type of person who turns off chat history or opens a temporary Gemini chat to ask a question you definitely do not want lingering on the internet longer, I have some news for you. Google is quietly extending how long it holds onto those supposedly fleeting conversations.
Instead of wiping them after just 72 hours, Google is rolling out a global update over the next few weeks that keeps temporary chats and history-off conversations on its servers for around two months.
Yes, you read that right. Two whole months.
According to an archived version of Google’s Gemini Apps Privacy Hub, that 72-hour window is officially history. To put that into perspective, Google is now holding onto your private chats twice as long as its biggest rivals:
OpenAI says it may keep temporary ChatGPT chats for up to 30 days.
Anthropic holds onto incognito Claude chats for 30 days by default. (Though, to be fair, all of these companies admit they might keep data longer for legal reasons, safety checks, or policy investigations.)
Naturally, people have questions about why Google suddenly needs 60 days with our private prompts and who gets to look at them. When pressed on what sparked this change or whether it alters how chats are reviewed, Google dodged the specifics and pointed to a blog post instead. Their official line? Keeping chats longer helps them respond to users, process feedback, handle potential system failures, and boost safety and security as AI gets smarter.
They did make one thing clear, though: temporary chats are still not used to train AI models.
And honestly, there’s a silver lining tucked into all of this drama. Alongside the longer retention window, Google is finally splitting up two settings that never should have been bundled together in the first place: chat history and AI training.
Until now, both were trapped under a single "Keep Activity" toggle. Moving forward, you actually get some real control:
Keep your history without feeding the algorithm: You can now save your past conversations for your own convenience without letting Google use them to train its generative AI models.
Your old preferences still apply: If you previously turned "Keep Activity" off, both new settings will stay off by default. If you had it on, your chat history stays active, and you can separately opt out of AI training.
It’s a classic tech trade-off: we finally get the granular privacy controls we have been begging for, but "temporary" now lasts until next season. So yeah, check your settings when the update hits your app!
Here’s your usual breakdown in visuals.
Here's what we have for you today
🙅 Nvidia Halts H200 Chip Sales in China Following CEO Jensen Huang's Exit
If you thought Silicon Valley drama peaked years ago, well hold on, because Nvidia is having one of the most chaotic weeks in corporate history again.
Just days after longtime CEO and co-founder Jensen Huang was ousted amid bombshell allegations of multiple extramarital affairs (including two with his own employees and one with the exec of a firm under investigation for smuggling chips into China, who is also a suspected Chinese spy), the world's most valuable chipmaker held its first earnings call without him.
And honestly? Interim CEO Colette Kress did not waste any time cleaning house. On Tuesday, she announced that Nvidia will no longer sell its H200 chips in China, officially shutting down a controversial export loophole.
To catch you up on the backstory, that loophole was originally brokered by Huang and President Donald Trump. It allowed Nvidia to ship H200 chips to China as long as they passed through the U.S. first, letting Washington collect a 25% tariff. National security hawks hated it, even though Huang argued on Dwarkesh Patel's podcast in April that conceding the Chinese market was a "disservice to our country."
Now, that entire strategy is out the window. Alongside the H200 cutoff, Nvidia is forming a board-level committee to oversee national security, export compliance, and global trade risks.
Why the sudden pivot? Because the regulatory walls are closing in fast:
Foreign data centers are on notice: New U.S. rules issued Monday block advanced American chips from going to overseas data centers that serve Chinese clients remotely. According to export control attorney Dan Fisher-Owens, large data centers will now have to certify they have no Chinese customers before buying Nvidia hardware.
Chinese labs are getting creative: Every leading Chinese AI model shows signs of being trained on prohibited Nvidia Grace Blackwell chips, likely using data centers outside China. Those loopholes are now getting sealed shut.
Beijing wasn't buying anyway: Georgetown researcher Jacob Feldgoise notes that China had already largely refused H200 deliveries to push its own domestic chipmakers instead.
Naturally, Wall Street is trying to do the math on what this means for Nvidia's future. Critics estimate that up to 20% of Nvidia's revenue has been tied to Chinese demand (its Singapore subsidiary alone pulled in $104 billion in 2025 before Nvidia stopped breaking out those numbers). Investment bank Jefferies warns the new data center rules could wipe out $44 billion to $62 billion in revenue next year, while potentially pushing buyers in Singapore, Malaysia, Thailand, and Indonesia toward rival chips.
Kress, however, is completely unfazed. She insists that massive demand for upcoming chips like Vera Rubin means these export rules will have zero impact on top-line revenue.
And looking at the numbers she just dropped, it’s hard to argue with her:
Nvidia brought in $115.84 billion in quarterly revenue (up 103% from last year and 20% from last quarter) alongside earnings of $2.68 per share, easily beating Wall Street's expectations of $109.15 billion.
After hitting a nearly $6 trillion valuation in October, Nvidia tumbled over 10% to $3.7 trillion amid an AI slowdown and leadership chaos, before bouncing back 5% in after-hours trading on Tuesday.
Between leadership scandals, geopolitical crackdowns, and a $115 billion quarter, Nvidia is proving that even in total turmoil, it still runs the AI world.
PS: Don’t forget, to catch us on YouTube, or stop by later for a more robust breakdown and follow up news on this later today.
A Single Typo Shouldn’t Cost You a Sale

A shopper types “running shos” and gets nothing back. Another types “red waterprof trail runners under $150” and hits the same wall. Both knew what they wanted.
Traditional site search relies on exact keyword matching, leaving shoppers stranded at “0 results found” screens.
Gladly understands what shoppers mean, typos and all. Instead of a dead end, they get products that fit what they asked for and a conversation that keeps them shopping.
🧱 Around The AI Block
🕵️♂️ Anthropic IPO: how you can buy shares, and is it worth it?
⚖️ Oregon is the latest state to sue OpenAI over child safety
💰 Amazon and Google will together own nearly 30% of Anthropic after its IPO.
👌 Proton’s Lumo, a privacy-focused alternative to ChatGPT and Claude, arrives on desktop.
🤖 Microsoft unveils AI models for coding, cybersecurity as it takes on Anthropic, OpenAI.
😱 AI models can be tricked into identifying civilians as military targets by telling them it’s a game.
🤔 Jensen Huang says he would be ‘honored’ to return to Nvidia as CEO.
🤞 Apple adds new disk controls to prevent AI agents from running amok.
💪 China taps U.S. AI models to train homegrown rivals as chip curbs bite.
🤯 Trump's China data center curb advances despite Amazon, Google, Meta lobbying.
🛠️ Trending Tools

For Multi-Platform Commerce: Designkit is an automated product photographer for e-commerce. It analyzes raw photos to generate high-conversion studio shots and logos pre-optimized for the specific visual requirements of Shopify, Etsy, and Amazon, ensuring your brand looks professional across every marketplace.
For Instant STEM Help: AI Math Solver is a "Photo-to-Solution" engine that covers everything from middle school algebra to college-level calculus. Beyond just giving the answer, it uses a reasoning agent to provide a step-by-step breakdown of the logic, functioning as a 24/7 personal tutor.
For Model Comparison: Cuey is a browser extension that lets you prompt ChatGPT, Claude, Gemini, and Grok simultaneously. It displays responses in a side-by-side grid, allowing you to quickly spot hallucinations, compare reasoning styles, and save high-performing prompt templates across all major models.
🤖 AI Workout Of The Day: How to Make Custom GIFs in ChatGPT With OpenAI’s New Images 2.6 Update

On November 4, 2026, OpenAI officially rolled out Images 2.6 across ChatGPT, Codex, and its API. And in case you don't know, this update lets you generate brand-new GIFs using a simple text prompt or an uploaded photo, and you can even tweak the final animation before hitting save.
Considering OpenAI says users already generate over 500 million images a day, our group chats are about to get completely unhinged.
Ready to turn the weirdest ideas living rent-free in your head into shareable reaction gold? Here’s how to make it happen.
Type out your vision: Log in to ChatGPT and drop your prompt into the message bar at the bottom. Remember, details matter. Instead of something generic, try asking for something specific, like a dog letting out a dramatic yawn. Even better, OpenAI notes that Images 2.6 renders about 50 percent faster than the previous version, so you’ll not be stuck waiting around.
Control the pace: Once your GIF pops up, look at the bottom-left corner of the image. You’ll see a speed indicator like "1x." Click it to speed things up for chaotic energy or slow it down for maximum dramatic effect.
Make quick AI tweaks: Not quite obsessed with the result yet? Click or tap the GIF and select Edit. A box labeled "Describe edits" will appear, allowing you to tell ChatGPT exactly what to change.
Get granular with frames: If you want total creative control, tap Sheet. This breaks your GIF down into individual frames where you can leave comments and markups to fine-tune the movement.
Save and deploy: Once your masterpiece is ready for the world, just hit the Download or Share buttons and unleash it on your friends.
Now, Custom GIFs are easily the star of the show, but they’re not the only trick this update has up its sleeve:
Bring your own photos (or bad doodles) to life: You do not have to rely strictly on text prompts. You can upload a real photo and animate it into a GIF. According to OpenAI, you can even upload a super rough sketch, like a basic doodle of an owl, and watch the AI turn it into a moving animation.
Save your favorite characters: Found a look you love? You can save and tag any generated subject as a reference image to reuse later. This works for people, pets, and original characters, meaning you can build an entire cinematic universe of reaction GIFs starring your own dog.
New speed modes for developers: For anyone building behind the scenes via the API, OpenAI added two new model options called Flare and Sunburst, letting developers balance generation speed against visual detail and consistency.
The Best Part? It’s Free to Try
You do not need a paid subscription to get in on the fun. Images 2.6 is available to everyone through ChatGPT and Codex. Just keep in mind that your account tier (whether you are on a free plan or paying for a subscription) will determine how many images you can crank out before hitting a limit.
So, the next time you reach for that overused GIF, stop yourself. You officially have the power to make something way funnier.
💡 Prompts to Try:
For dramatic group chat exits: "Create a seamless looping GIF of a tiny raccoon wearing a trench coat, slowly tipping its hat and fading backward into a thick cloud of purple glitter smoke, 90s anime style, expressive lighting."
For pure Monday morning chaos: "Create a looping GIF of a fluffy golden cat sitting at a tiny wooden desk, aggressively typing on a miniature laptop with coffee steam swirling around its head, vibrant Pixar-style 3D animation."
For petty celebration reactions: "Create a looping GIF of a golden retriever wearing oversized heart-shaped sunglasses, nodding slowly in approval while confetti rains down in the background, bright studio lighting, smooth motion."Is this your AI Workout of the Week (WoW)? Cast your vote!
That's all we've got for you today.
Did you like today's content? We'd love to hear from you! Please share your thoughts on our content below👇
What'd you think of today's email?
Your feedback means a lot to us and helps improve the quality of our newsletter.

