A powerful AI that is very good at finding the open doors on your computer just came back online worldwide. It hunts the weak spots. You want to see yours first. Here is what an open door really is, three things you can do today, and an hour with me to look together.
Save My Seat →Thursday, July 9, 2 PM ET. No charge, and it is not a pitch-fest.
The problem: Last month the government pulled a powerful AI off the market. It was too good at finding weak spots in software and breaking in. On July 1, it came back online worldwide. It hunts the open doors on your computer faster than any crook ever could.
The solution: You do not have to wait for anyone. Having security software is not the same as being secure. By the end of this, you will know what an open door really is, you will have three things you can do today, and you will see why the humble first move is always to look.
In this article
Picture a locksmith who can test every lock on your street in an afternoon. Not to help you. To find the one door that opens.
That is roughly what happened. Last month the US government pulled a powerful AI off the market. It was too good at finding weak spots in software and breaking in. On July 1, 2026, it came back online worldwide.
How good is it? The coalition that runs it, called Project Glasswing, has already turned up more than 10,000 serious flaws in critical software. Power. Water. Hospitals. The systems everything runs on. That is the tool that now hunts the open doors on your computer, at machine speed.
Here is the thing most people miss. The big AI makers built the tool first. Only now are they proposing a shared way to grade how dangerous these tools can get. They shipped the hunter, then started arguing over the rulebook. You do not have to wait for that argument to end.
I have been walking you through this AI for weeks, so let me connect the dots. This tool did not just find holes in sloppy code. It found them in BSD, some of the most carefully reviewed software on earth, the gold standard. One flaw in OpenBSD had sat open for twenty-seven years. Another, in FreeBSD, for seventeen. Nobody caught them in all that time. The AI did, in an afternoon.
That matters for your whole shop, even the Mac users. Apple built macOS on that very same BSD. For years folks believed "Macs don't get hacked." That was always a myth, and it is a dangerous one now. This AI does not care which logo is on the lid. Windows or Mac, every machine is full of programs nobody patches. If the gold standard had doors open for decades, the everyday apps on your machines have them too.
Let me make this plain, because the word "vulnerability" scares people off. An open door is simple. It is a way in that nobody shut.
• An app nobody patched. The PDF reader. The old accounting add-on. That program you forgot you had.
• A risky setting Windows leaves the wrong way. On when it should be off, or off when it should be on.
• A program Windows Update never touches, because the built-in updater only covers Windows itself.
• A weak spot in software that sat open for years, waiting for someone to notice.
Now look at your own security screen. It likely says "You're protected." Green check. Feels good. That same screen would say the exact same thing with three doors wide open, because it is only watching for known bad files trying to walk in. It never checks whether your own programs are up to date. Green does not mean checked.
The honest gap is simple. Having security software is not the same as being secure. Defender guards the front hall. The open doors on your computer are the side windows nobody looked at. A door you already shut is not a way in. The whole game is knowing which ones are still open.
You run a business with a handful of staff. You did not sign up to be a security expert. That is fair.
But the math just changed. Finding the open doors on your computer used to take a skilled crook real work. Now an AI tool can do it in an afternoon. When the cost of finding your weak spot drops that far, "we'll get to it" stops being safe.
And do not assume this skips the small shop. The AI does not check the size of your company before it knocks. It looks for the open door. A ten-person business has the same forgotten apps as a big one, sometimes more, because nobody there has the job of watching them.
You cannot stop the tool from existing. You can stop it from finding an open door on your machines. And to do that, you first have to see what you have.
This is where seeing and fixing part ways. Seeing what is open is one job. Fixing it is the next. Most owners try to skip straight to fixing and never look, so they patch the wrong thing and leave the real door open. The humble first move is always to look.
That is exactly what a Reveal-Scan is for. It looks at your computer and hands you a plain report card, A to F, of the open doors, the ones Windows and its built-in Defender never check. It shows you the risky settings and the unpatched, risky software. It does not magically fix everything. Seeing it is the first step, and I show you what to shut first. No card needed to run one.
You do not need a consultant to start. Three steps, and the first one takes minutes.
See the open doors on your computer for yourself, graded A to F. You can start one here: forwardtosafety.com/account/reveal-scan/signup. No card needed. It shows you what is open so you can decide what to shut first.
Walk your main computer. Write down every program that did not come from Windows itself. The PDF reader, the browser add-ons, the old tools. Then update each one or remove it. These are the doors the built-in updater never touches.
Join the Insider Session and I will run a Reveal-Scan on screen, read the report card out loud, and shut the first open door in plain English. You will know exactly what to do next on your own machines.
Do the first two today and you have made real ground. The third, watching a live scan and a real fix, is the part I walk you through myself.
My father fell for a phishing email. I had been in this field for decades, and it still got him. The scammers got remote access and hunted for his financial documents. My step-mother noticed something was wrong. I jumped in from across the country and stopped them before they reached his bank logins. We were lucky. We caught it in time.
I have spent more than 35 years at this, since 1991. FBI InfraGard. Not one client I manage has ever been hit by ransomware. And when someone does get in, it is almost always through an app nobody updated, a forgotten open door. So no lecture from me. Feeling behind on this is normal. Seeing your open doors is simple once someone shows you how, and I made that part easy.
That green check is watching for known bad files at the front door. It is useful, and it is not the whole story. It never asks whether your own programs are patched, and that is where the open doors on your computer actually are. Green does not mean checked.
An hour now saves a scramble later. And if your machines are already in good shape, I will tell you so, no scare tactics.
The hunter is loose, and the people who built it are still arguing over how to grade the danger. You do not have to wait for them. Spend one hour with me on Thursday, July 9 at 2:00 PM Eastern. I will run a Reveal-Scan live, read the report card, and shut the first open door on screen. You walk away knowing what is open on your own machines and what to fix first. No charge, and it is not a pitch-fest.
Save My Seat →Thursday, July 9, 2 PM ET. No charge, and it is not a pitch-fest.
Want plain-English security news every week? Sign up for Craig's Insider Notes at CraigPeterson.com.
— Craig
Join thousands of security professionals who receive Craig Peterson's Insider Show Notes and cybersecurity updates.
Join 10,000+ cybersecurity professionals