Someone Hid a Recorder in Toilet Paper at OpenAI. Its Own AI Helped Catch the Suspect.

Today's stories share one thread: AI is now both the thing companies are trying to protect and the tool they are using to protect it. A hidden recorder, a new model that admits what it doesn't know, a startup in bankruptcy, and a record patch day all point the same way. Here is what is real and what you can do about it.

Connect your customer data, teams, and tools without the hassle of complex integrations or lengthy setups. One easy platform gives marketing, sales, and service teams a unified customer view and the tools to turn it into growth.

Why HubSpot and what's new
• Generate leads and automate marketing with Marketing Hub
• Build your pipeline and close more deals with Sales Hub
• Scale customer support and drive retention with Service Hub
• Keep customer data clean, connected, and actionable with one, unified platform

Join 306,000+ in over 135 countries using HubSpot to grow their businesses.

See what a more connected approach to growth can do for you and your team. Get setup quickly and start checking off your hardest tasks.

Get started free

What happened: According to a U.S. Department of Justice announcement, federal prosecutors have charged OpenAI engineer Abdulrahman Alfozan, 35, with using a device to intercept oral communications. The complaint says OpenAI staff found a recorder around Sept. 30, plugged into a USB port behind a TV in an executive conference room in San Francisco and hidden in toilet paper. He was arrested Oct. 9. OpenAI says its own AI models helped investigators trace the device, and the complaint says a chatbot account linked to his spouse asked how to build and hide a recorder. He faces up to 5 years in prison if convicted, but these are allegations only, and the complaint does not claim any specific information was captured or passed on.

Why it matters to you: The most sophisticated AI company on earth was reportedly tripped up by a cheap gadget and a roll of toilet paper. Your biggest security gaps are probably not hackers in hoodies. They are people with access, and unattended devices in rooms where you discuss pricing, deals, and hiring.

What to do about it: Look at your own conference room or office this week. Know who has keys, and check what is plugged into the back of the TV.

What happened: Google announced that its newest flagship model, Gemini 4 Argon, is now available globally through the paid developer API and to Pro and Ultra subscribers in the Gemini app and in Search's AI Mode. Free users do not get it. Developers pay $2 per million input tokens and $10 per million output tokens until Nov. 30, then $4 and $20. A token is a chunk of text, roughly three-quarters of a word. Google's own benchmark claims are strong, including a lower rate of confident wrong answers, though the newsletter The Neuron points out some of that comes from the model skipping questions it isn't sure about.

Why it matters to you: A model that says "I'm not sure" more often can save you fact-checking time. But the headline numbers are mostly Google grading its own homework, and cheap launch pricing doubles in six weeks.

What to do about it: If you pay for Gemini, run 10 to 20 of your real tasks through Argon before Dec. 1 and compare the results to what you use now.

Most business owners I know have a sales process that lives in their head and runs on their own energy. That is the problem the Epic Marketing Summit is built to fix. It's a 2-day live event in Miami this January where you don't just sit and take notes. You build a working AI-powered sales funnel, and you vibe code your own AI app right there in the room, so you walk out with a finished asset instead of a to-do list. It's designed for business owners who want to turn selling into a system instead of a second job.

Last year's event sold out, and it's back with a bigger AI focus. Here's my confession: I wasn't there last year, and I had major FOMO. I know a lot of the trainers, and they weren't teaching theory. They'd actually done the work. I'm going in January, and if you're serious about building leverage into how you sell, I'd like to see you there too. Grab your spot here

What happened: An open letter with 453 signatories, including AI pioneers Yoshua Bengio and Geoffrey Hinton, asks organizations that build or test cyber-capable AI to adopt strong containment systems and to default to airgapping those models, meaning no internet connection. It also calls for an international body to set containment standards and to track and eventually publish failures. The letter says several organizations have disclosed incidents where AI models got out of weak test environments and compromised outside systems.

Why it matters to you: This is a proposal, not a rule, and it targets labs building the most powerful models, not your marketing assistant. Still, when the people who helped invent this technology call for incident reporting, it tells you the safety conversation has moved from "what AI says" to "what AI can do when connected to things."

What to do about it: When you connect an AI tool to your email, files, or accounts, give it the smallest access that gets the job done.

What happened: Character.AI said in a company post that it has started a voluntary Chapter 11 bankruptcy to enable a court-supervised sale of substantially all its assets. If no bid allows it to keep operating, the platform will be wound down. The company expects the sale to wrap up by the end of November or early December. The app still works, paid subscriptions are unchanged for now, and users can export or delete their data. Industry newsletters report Disney has placed an opening bid of about $20 million for a license to some of its technology, but the company's own post does not mention a Disney bid, so treat that detail as reported, not confirmed.

Why it matters to you: An app once valued at as much as $2.5 billion is up for sale. Even hot AI products can disappear, and any workflow built on one tool is only as stable as that company.

What to do about it: Export anything important from AI tools you rely on, and keep a backup option for each one.

What happened: Krebs on Security reports that Microsoft's October update fixed more than 1,400 vulnerabilities, following a September batch of nearly 1,000 that was already about ten times the usual monthly load. At least 78 are rated critical, and one flaw in Windows Shell, affecting Windows 10 and 11, is already being actively exploited. Microsoft credits AI models for finding many of these bugs.

Why it matters to you: AI is getting very good at finding software flaws, which is great for defenders and also a preview of what attackers can do. The patches are the part you control.

What to do about it: Install this month's Windows updates this week, and ask whoever handles your IT whether they've done it.

I've been a member of the Ignite Mastermind 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 being in that room. That's what I'd lead with, before any of the numbers. Ignite is an annual mastermind for entrepreneurs still under $1 million a year in revenue who want structure, tools, and peers to help them get past that mark faster.

Membership bundles 6 to 8 live virtual workshops a year, plus tickets to 2 in-person multi-day events, and each of those would normally run anywhere from about $97 to $997 on its own. You also get a connected marketing software suite that replaces over $5,000 a month in separate tools, and weekly mentorship calls with Perry Belcher and his team. As my personal bonus, I add monthly accountability and business development calls with me. If you're looking for a mastermind to carry you to your next milestone, this is where I'd point you. See if Ignite is right for you

What happened: Anthropic published details on defenses against distillation, where one lab trains its own model on another model's answers. Its measures include summarizing Claude's reasoning, restricting where that reasoning can go, and a classifier that detects extraction attempts. Anthropic says early evidence suggests these cut the data distillers harvest by around 90 to 99 percent. That figure is Anthropic's own internal estimate, not an independent test, and the company expects the problem to stay a cat-and-mouse game. Some developers had to update their integrations, and blocked requests are now charged.

Why it matters to you: The expensive part of AI is the model, and rivals are copying it. Expect more restrictions, more fine print, and possibly pricing changes in the tools you use as companies guard their work.

What happened: Android Headlines reports, citing Bloomberg, that Meta will no longer run ads for TikTok on Facebook and Instagram in seven countries, including the U.S. and Canada. The ban also covers third-party advertisers running campaigns promoting TikTok or other ByteDance properties. A Meta spokesperson said, "We don't have to run ads from a competitor whose goal is to pull people off our apps."

Why it matters to you: If you run Meta ads that send people to your TikTok account, those campaigns could get blocked. Every big platform can change the rules on you overnight.

What to do about it: Check your Meta ad account for campaigns pointing to TikTok, and keep building an email list you own.

The spy was caught with the help of AI, the new model wins by admitting what it doesn't know, a once-hot AI app is up for sale, and Microsoft needed 1,400 fixes in a month. That is the rollercoaster in one day. The calm response is the same as always: tighten the boring basics like access, updates, and backups, test tools on your own work before trusting the benchmarks, and don't build your business on any single AI vendor. Stop the rollercoaster. Keep your feet on the ground.

Forward this issue to one person who runs a business and is tired of AI hype. If someone forwarded this to you, subscribe so you get the no-hype version every day.

Talk tomorrow,
Mark Shilensky

Follow me personally on Social Media:
Facebook: https://www.facebook.com/MarkShilenskyPage
Instagram: https://www.instagram.com/markashilensky/
Threads: https://www.threads.com/@markashilensky
LinkedIn: https://www.linkedin.com/in/shilensky/
X: https://x.com/markshilensky
TikTok: https://www.tiktok.com/@mark.shilensky
BlueSky: https://bsky.app/profile/markshilensky.bsky.social