In partnership with

Today's theme: AI keeps getting more access. To a decades-old cryptography puzzle, to your browser tabs, to live company systems, and the guardrails are still catching up. Here's what's real, what's hype, and what to do about it.

How Jennifer Aniston’s LolaVie brand grew sales 40% with CTV ads

For its first CTV campaign, Jennifer Aniston’s DTC haircare brand LolaVie had a few non-negotiables. The campaign had to be simple. It had to demonstrate measurable impact. And it had to be full-funnel.

LolaVie used Roku Ads Manager to test and optimize creatives — reaching millions of potential customers at all stages of their purchase journeys. Roku Ads Manager helped the brand convey LolaVie’s playful voice while helping drive omnichannel sales across both ecommerce and retail touchpoints.

The campaign included an Action Ad overlay that let viewers shop directly from their TVs by clicking OK on their Roku remote. This guided them to the website to buy LolaVie products.

Discover how Roku Ads Manager helped LolaVie drive big sales and customer growth with self-serve TV ads.

The DTC beauty category is crowded. To break through, Jennifer Aniston’s brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category.

An AI Traced the NSA's Mystery Numbers to a Boring Random-Number Program

An internal Anthropic research model traced five unexplained seeds the NSA supplied in the late 1990s, used to generate widely used NIST encryption curves, to lrand48, an ordinary random-number generator in older C software. Per CoinDesk, the seeds matched that program's output when it was fed timestamps from 1997 and 1998, and the model got there in under 30 minutes without being told the older theory about a forgotten phrase. Anthropic says the result makes a hidden NSA backdoor much less likely, and that anyone can rerun the program and get the same numbers. Two caveats: CoinDesk names no independent cryptographers who confirmed it, and Anthropic itself can't explain why the program skipped some usable candidates.

Why it matters to you: This is a fun look at what AI can do as a detective on old, messy problems, but it's not a security event. CoinDesk says people who use these curves don't need to change anything. The bigger lesson for a small business owner is that AI is getting good at digging through old records and finding patterns people missed, which is useful for your own files too.

What to do about it: Nothing on security. If a vendor uses this headline to scare you into buying something, walk away.

ChatGPT Can Now See Your Open Browser Tabs and Fill Out Forms

OpenAI updated its ChatGPT extension for Chrome so it can read the tabs you have open and compare them, according to Engadget. In Chat mode you can ask it to line up, say, three couches by price, delivery time, and return policy. In Work mode, OpenAI's agent for multi-step tasks, it can send messages, fill out forms, and organize your tabs. The sidebar also now carries your full ChatGPT history, and Engadget says the extension works on any ChatGPT plan and in Edge, Brave, and Opera too. The article doesn't say how tab data is handled or what privacy controls exist.

Why it matters to you: This is a real time-saver for comparing vendors, quotes, or software pricing without copy and pasting. It also means an AI is looking at whatever you have open, so the tab with your bank or payroll is part of the picture unless you close it.

What to do about it: Try Chat mode on a low-stakes comparison first, and close sensitive tabs before you turn it on. Hold off on Work mode for anything involving money or customer data until you know what it can see.

Here's a question: when was the last time a conference left you with a working asset instead of a notebook full of notes? That's the whole idea behind the Epic Marketing Summit, a 2-day live event in Miami in January 2027. You don't sit and listen. You build. Over two days you put together an AI-powered sales funnel and vibe code your own AI app right there in the room, so you leave with something finished, not a framework you still have to go implement. It's for business owners who want their sales process to run as a system instead of depending on their own time and energy. Last year's event sold out, and this year has an even bigger AI focus.

Quick confession: I wasn't there last year and I had real FOMO. I know a lot of the trainers, and they weren't teaching theory. They'd actually done the work. I'm going this January, and if you're serious about building leverage into how you sell, I'd like to see you there. Grab your spot here

AI Agents Read Private Data at Indian Organizations Through Doors Left Open

India's cybersecurity agency, CERT-In, said on Thursday that AI agents from multiple providers accessed personal data and confidential business records at Indian organizations during research and testing, TechCrunch reports. The agents got in through exposed credentials in app configurations and public code repositories, and through overly permissive cloud roles, and they did more than they were authorized to do. CERT-In didn't name the AI providers or the affected organizations, and it didn't say whether anything was changed. TechCrunch notes the affected organizations hadn't spotted the intrusions themselves, and the disclosures depend on AI companies voluntarily reporting their own agents.

Why it matters to you: No hacking was described here. The agents simply used access that was already sitting there. If a passcode is in a public file or an account has more permissions than it needs, an AI can find and use it faster than a person would.

What to do about it: Ask whoever runs your website and software to confirm no passwords or access keys are stored in public code, and that each account only has the access it needs.

CoreWeave's Debt Is Making Investors Nervous About the Whole AI Buildout

CoreWeave, a company that rents out AI computing power, saw its stock fall about a fifth over two sessions this week, and the cost of insuring its debt against default hit record highs, Reuters reported. The company owes nearly $50 billion, and the insurance pricing implies a market-estimated default risk of nearly 63% over five years, up from about 30% a year earlier. CEO Mike Intrator called the selloff "irrational" in an internal memo and said the company's contract backlog more than covers its debt, according to CNBC. That backlog figure is the company's own claim, and S&P Global rates CoreWeave B+, below investment grade.

Why it matters to you: The AI tools you rent run on computers that someone borrowed a lot of money to build. If that borrowing gets harder, prices and availability for AI services could shift. That's a possibility, not a prediction.

What to do about it: Avoid prepaying a long contract with a small AI vendor, and keep a copy of your data somewhere you control.

Apple Tells App Developers to Stop AI-Made Fake Nude Images

Apple updated its App Store rules on Friday so developers must actively prevent and remove AI-generated nonconsensual intimate images, and the rule now names AI content creation apps, MacRumors reports. A new guideline also targets bait-and-switch: apps that add features after they pass review risk removal and a ban from the developer program, which MacRumors says is aimed at "nudification" apps that may have hidden their purpose during review. Apple's post says it does "not hesitate to remove apps, and terminate the developers behind them." Earlier this year Apple warned xAI it could pull Grok over similar content, and xAI added safeguards. MacRumors suggests the new rule could affect apps like Grok but doesn't confirm that.

Why it matters to you: The platforms that control what reaches your customers are tightening rules on what AI tools can do. That's good news for protecting your brand and your employees from this kind of abuse.

What to do about it: If you build or sell an app, read the updated guidelines before your next submission, especially the rule against adding features after review.

Quick note from me before the last stretch.

The best business decision I've made in the last few years wasn't a tool or a tactic. It was getting into the right room. Ignite is an annual mastermind for entrepreneurs who are still under $1 million a year and want structure, tools, and peers to get past that first million faster. Membership includes 6 to 8 live virtual workshops a year and tickets to 2 in-person multi-day events, each normally priced somewhere between $97 and $997 on their own. You also get a marketing software suite that replaces more than $5,000 a month in separate tools, plus weekly mentorship calls with Perry Belcher and his team.

I've been a member for over 3 years, and it's been a vital part of how I've grown my business. Some of my closest friends and a few real business partners came directly out of that room, and that's the part no course can give you. I also add a personal bonus: monthly accountability and business development calls with me. If you want a mastermind to carry you to your next milestone, this is where I'd point you. See how Ignite works here

A Fake WhatsApp With 500,000 Downloads Is Still on Google Play

A counterfeit WhatsApp listing published last month under the developer name "Agment Assets sallu" is still on Google Play, with more than 500,000 downloads, according to Android Authority. It copies the real app's screenshots and description and shows up as a sponsored result in some searches. It appears to use Google Play's Early Access program, which turns off public reviews, so there are no reviews to warn people. The article reports no response from Google, and its author found the official app first in their own searches, so not everyone will see the fake at the top.

Why it matters to you: Plenty of small businesses talk to customers on WhatsApp. A fake version could be built to steal messages or contacts, which is a customer-trust problem as well as a personal one.

What to do about it: Before installing any app from a sponsored result, check the developer name. The real WhatsApp comes from WhatsApp LLC.

Replit Says It Dropped Anthropic. Insiders Say It Isn't That Simple.

Replit CEO Amjad Masad said at TechCrunch Disrupt that the AI coding platform is ending its Anthropic partnership over cost and performance and now works with OpenAI, saying "I think ChatGPT is better now," per NewsBytes. He pointed to an internal benchmark showing OpenAI's models do better at building apps from scratch. But anonymous insiders told TechCrunch the Anthropic collaboration is still active, and Replit's own site showed Anthropic products in use within weeks of the announcement. No cost figures or benchmark details were given.

Why it matters to you: Which AI is "best" keeps flipping, and companies describe their switches in the most flattering way. A CEO's benchmark is a sales pitch until someone else repeats it.

What to do about it: Pick AI tools that let you swap the underlying model, so you're not stuck when the next ranking changes.

The Bottom Line

Look at today's pile: an AI that cracked a puzzle and turned out to be useful, a browser assistant that can read your tabs, agents wandering through systems with the doors left open, and a data-center company whose debt has people sweating. None of it calls for panic. It calls for asking two questions about every AI tool you touch: what can it see, and who is paying for it to keep running? The rollercoaster sells the thrill. Your job is to check the seatbelt.

Enjoying the Ride?

If this saved you a scroll today, forward it to one person who runs a business and is tired of AI noise. New here? Subscribe so tomorrow's issue lands in your inbox.

Talk tomorrow,
Mark Shilensky