Modern Creator
Slop Architect · YouTube

People Are Using AI to Hack Everything They Own

A tour of the REA movement: people pointing coding agents at their TVs, mirrors, printers and air fryers, what they found inside, and who might stop them.

Posted
today
Duration
Format
Essay
sarcastic
Views
358.6K
8.6K likes
Big Idea

The argument in one line.

Coding agents have made reverse engineering your own devices trivial, so the real power over what you own is shifting from manufacturers to the AI companies that decide what their models will help with.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…
  • You own smart devices with bad apps, forced ads or dead companion software and wonder whether an AI agent could fix them for you.
  • You build with Claude Code or similar agents and want concrete examples of what people ship when the target is hardware instead of a SaaS app.
  • You follow right-to-repair, privacy and device ownership debates and want the current state of the AI angle.
SKIP IF…
  • You want a technical walkthrough of how to root a specific device; this video shows results and stories, not step-by-step instructions.
TL;DR

The full version, fast.

A GitHub project called Reverse Engineer Anything went from under 10,000 to nearly 400,000 stars in three weeks, and it lets a coding agent inspect firmware and app traffic without source code. People used it, and plain agent prompts, to strip ads from TVs, turn a $50 photo frame and a discarded Lululemon mirror into dashboards, voice-control an air fryer, and rebuild a dead Mac printer driver. Digging also exposed always-on voice and face code in a Samsung TV. The main pushback comes from AI safety filters and account bans, plus real risks like fake malware copies and bricked devices. The creator argues companies should document their APIs instead of shipping slop.

Free for members

Chat with this breakdown — free.

Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.

Create a free account →
Chapters

Where the time goes.

00:00 – 00:40

01 · AI ripped the ads out of this TV

Cold open with TV, photo frame and mirror hacks, and the roadmap for the video.

00:40 – 01:07

02 · 400,000 GitHub stars, and nobody knows who made it

REA's explosive growth and its anonymous author.

01:07 – 02:31

03 · Every smart device keeps two secrets

Firmware and the app protocol, and how REA exposes both.

02:31 – 03:24

04 · Why hack your own stuff?

Ownership, enshittification and the lamp, treadmill and printer examples.

03:24 – 04:34

05 · Jailbroken TVs

The Xiaomi TV that showed an ad before Netflix, Samsung jailbreaks, and a TV that mutes ads.

04:34 – 05:31

06 · Wes Bos's $50 photo frame

A cheap frame becomes a family dashboard that replaces a $300 product.

05:31 – 06:23

07 · Lululemon's $500 million mirror

A written-off product becomes a particle-based AI assistant.

06:23 – 07:01

08 · Kindles with no ads

Agent-driven Kindle jailbreaks and family dashboards.

07:01 – 08:33

09 · SlopHouse: the app Claude built me

The creator's own air purifier and baby monitor app, and why it is not a security hole.

08:33 – 09:41

10 · An AC, an air fryer and a dead treadmill

Voice-controlled air fryer, a laptop AC button, and a treadmill fixed over Bluetooth.

09:41 – 10:13

11 · Printers: a new Mac driver in 2 prompts

REA pointed at an abandoned Brother driver produces a native Mac one.

10:13 – 10:45

12 · What's hiding inside a Samsung TV

Always-on voice, face detection and sound classification code found after rooting.

10:45 – 11:22

13 · ACR: how your TV knows what you're watching

Automatic Content Recognition and what disabling it did.

11:22 – 13:21

14 · Who's pushing back?

TV makers stay silent while AI safety filters, an OpenAI account ban and Kimi's data removal become the friction.

13:21 – 14:01

15 · The catch

Fake REA copies with malware, bricked TVs and no support line.

14:01 – 14:16

16 · Is this legal?

Copyright bypass rules and the exemptions that already cover smart TVs and repair.

14:16 – 14:40

17 · Is this all just slop?

The creator argues a self-built app beats shipped slop.

14:40 – 15:51

18 · Two paths for the companies

Lock down harder or document the API, and the call for comments.

Atomic Insights

Lines worth screenshotting.

  • REA went from under 10,000 to nearly 400,000 GitHub stars in about three weeks, putting it in the top 10 most starred projects ever.
  • Most of the hacks in the REA wave never used REA itself; once people saw it was possible they just asked their agents to do it.
  • Every smart device keeps two secrets: the firmware running inside it and the protocol it uses to talk to its app.
  • A third-party app that controls your device is not a security hole, because it calls the same cloud endpoints as the official app and still needs your login.
  • A $50 Amazon photo frame rooted with a coding agent replaces a smart family calendar that retails around $300 plus a subscription.
  • A Lululemon-owned Mirror that cost over $1,000 new now sells for a couple hundred dollars secondhand, which makes it a cheap hacking target.
  • A man fixed his dead treadmill keypad by finding a hidden second Bluetooth connection, after four days of AI safety refusals.
  • A Brother laser printer abandoned by its last Mac driver in 2019 got a native replacement written with two prompts.
  • Rooting a 2025 Samsung TV exposed code for an always-on voice service, face detection and sound classification of noises like a barking dog.
  • Automatic Content Recognition lets a TV identify what is playing using screenshots and audio clips, and turning it off reportedly made one TV run faster.
  • When one AI refuses a reverse engineering request, users simply switch to another model such as DeepSeek or Codex.
  • To an AI safety filter, rooting your own TV and breaking into someone else's look very similar, which is why refusals are so frequent.
  • Fake copies of REA with malware inside appeared quickly, so a similar repo name does not mean the download is official.
  • Locking devices down harder only starts a cat and mouse game where attackers burn more tokens; documenting the API is the cheaper path for companies.
  • If a company ships slop software now, users can take it apart and rebuild it better themselves.
Takeaway

Who decides what your devices can do

WHAT TO LEARN

Coding agents have turned device reverse engineering from an expert skill into a prompt, which moves control from manufacturers to whoever runs the models.

  • Smart devices keep two secrets, their firmware and their app protocol, and an agent can often map the protocol without any source code.
  • A third-party app that calls the same cloud endpoints still has to log in with your account, so it does not make the device less secure.
  • Cheap or abandoned hardware, like a $50 photo frame or a discarded fitness mirror, is the best target because the downside of bricking it is small.
  • Digging inside your own devices can reveal surprises such as always-on voice, face and sound detection code, and tracking features you can turn off.
  • AI safety filters cannot easily tell rooting your own TV from attacking someone else's, so expect refusals and be ready to try another model.
  • Risks are real: fake copies of popular tools carry malware, updates can undo your changes, and an unchecked device on your home Wi-Fi is a possible way in.
  • Companies can either lock down harder and lose a cat and mouse game, or document their APIs and let owners build what they want.
Glossary

Terms worth knowing.

REA (Reverse Engineer Anything)
A GitHub project that lets an AI agent inspect an app or device without source code, explain how a feature works, and help build a version of it.
Firmware
The machine code running inside a device such as a TV or air fryer. It cannot be read directly, which is why it is treated as a secret.
Jailbreak / root
Gaining deeper control over a device than the manufacturer intended, such as removing preinstalled apps or installing your own software.
ACR (Automatic Content Recognition)
A TV feature that identifies what is playing by sampling screenshots and audio clips, mainly used to target ads.
API endpoint
A web address a company's server exposes for actions like turning a fan on or fetching a camera feed. Official apps and third-party apps can both call them.
Enshittification
Cory Doctorow's term for products getting steadily worse for users as companies squeeze more value out of them.
Resources

Things they pointed at.

00:20toolReverse Engineer Anything (REA)
04:34channelWes Bos photo frame video
07:01productSlopHouse
14:40channel404 Media
Quotables

Lines you could clip.

00:20
“Pretty cool to be able to plug in the frame and just tell your clanker to hack it.”
A reply quoted in the video that sums up the movement in one line.→ TikTok hook↗ Tweet quote
13:30
“There's no option to ship slop software anymore. Because if you do, people are just going to take it apart and make it better themselves.”
Clean thesis with no setup needed.→ newsletter pull-quote↗ Tweet quote
10:00
“The power was never in the repo. It was in the AI models. And whoever runs these models decides what they'll help with.”
Reframes the whole story in three sentences.→ IG reel cold open↗ Tweet quote
The Script

Word for word.

Read-along

Don't just watch it. Burn it in.

See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.

This TV was wall -to -wall with ads until AI ripped them all out. This $50 photo frame was just for pictures of grandma until AI hacked it into a custom family dashboard. And Lululemon spent $500 million on this smart mirror and eventually gave up on it.
Then AI turned it into somebody's personal assistant. And all this started with a GitHub project called Reverse Engineer Anything, or REA for short. In this video, we'll cover how REA works, why you'd want to hack your own stuff, the coolest stuff people are building, what they found hiding inside, and who's trying to shut it all down.
And stick around because later on, I'll show you a little project I've been working on myself. All right, so at the start of October, aka Sloptober, REA had under 10 ,000 stars on GitHub. Three weeks later, it's nearing 400 ,000 stars.
And that puts it in the top 10 most starred projects of all time on GitHub, right behind OpenClaw. REA did it in less than a month, and nobody really knows who made it. The account's called Morluto.
No real name, no company. The bio just says member of technically staff. Interesting.
Concerning, but interesting. Real quick on how this stuff works. Every smart thing you own, your TV, your printer, your air fryer, keep at least two secrets.
The first is the code running inside on a device that's called firmware. You can't read it. It's just machine code that your device runs.
The second is how it talks to its app. Over Wi -Fi or Bluetooth, your phone and your smart stuff speak their own price.
Then it tells you how the whole thing works and it shows you where it found each piece. Their pitch is pretty blunt. See a feature in an app that you want in your own product?
Ask your agent to investigate it with REA. It can inspect the app without the source code, explain how the feature works, and build a version for you. It's being called the new Wild West of AI.
Funny thing is, a lot of what you're about to see didn't even use REA. Once people saw it was possible, they just started asking their agents to do it. REA was just the spark for what I call the REA movement.
All right, guys, before we go any further, between a full -time job and taking care of my one -year -old, these videos get made in whatever free time I have left. So here's our gentleman's agreement. If you're enjoying this so far, hit subscribe, drop a like, and leave a comment for the algorithm.
And in return, I'll keep bringing you grade A USDA certified slop. Okay, so why would anybody want to take apart their own TV? Cory Doctorow, the guy who coined the word in shitification, put it simply.
You bought the TV, it's yours. But it doesn't really feel that way anymore, does it? One commenter under one of the big REA videos got almost 3 ,000 likes.
It said, companies will own nothing and be happy. Imagine this, you buy a lamp, a literal light bulb with a stand, and the only way to turn it on is through some trash app that shows you ads. And somebody at the lamp company probably got promoted for this.
Or the keypad on your treadmill dies, and now it's a really expensive coat rack. Or your printer is working perfectly fine, but nobody bothered to make it work with your new computer. For years, the answer to all this was you can't.
You don't know enough. That's for engineers to do. And then a whole bunch of regular people figured out you can just do things.
Okay, let's start with the TV, because the TV is where this whole thing blew up. One guy's parents had a Xiaomi TV, and it was showing a sponsored ad before you could even get to Netflix on a TV they paid for. That's a non -consensual ad.
Somebody should be in jail for that. So he gave Claude Code the model number, it found the TV on the Wi -Fi, ripped out a bunch of Xiaomi apps, and wrote a whole new home screen. In his words, quote, looks like an Apple TV now, lol.
That post did over 8 million views, and then everybody wanted one. He turned it into a free project, and within a couple of days, people all over the world were sending in photos of their own TVs, and it works on hundreds of different models now. Then people started getting creative.
So this guy jailbroke his Samsung TV over Wi -Fi, ripped out the ads, and also gave it that Apple TV style. Then he built a little app for his wife that shows a cheesy photo montage on the TV. Fellas, this is the bar now.
Flowers aren't enough anymore. You have to jailbreak the television. And then this one's my favorite.
Somebody set up their TV to take a screenshot of whatever's playing and send it to an agent on their computer. If it's an ad, the TV mutes itself and throws a card over the screen until the ad is over. But it's not just TVs.
It's anything with a screen or anything that comes with an app. Wes Boss, a web developer with a huge following, used a coding agent to root a cheap $50 Amazon photo frame and turn it into a little home display. It's got Wi -Fi and a touchscreen, and it runs off USB.
He even made a whole video about it on his YouTube. So now it's got the family calendar, the chores, the grocery list, and when piano lessons are. Companies sell this exact thing as a smart family calendar for about 300 bucks.
plus a yearly subscription for extra features. And now you have tons of people hacking these things. This guy replied with his own.
His has lights, door locks, the garage, and he summed up the whole movement in one line. Quote, pretty cool to be able to plug in the frame and just tell your clanker to hack it. That's where we're at, folks.
Just tell your clanker to hack it. And side note, I looked up the word clanker. Apparently it's a slur for AI agents.
So be careful how you use it. Okay, now Lululemon. Back in 2020, Lululemon bought a company called Mirror for $500 million.
It's literally a mirror with a screen inside it. You stand in front of it and do workout classes. And these things cost well over $1 ,000 new.
So Lulu bought it at the peak of the pandemic when everybody was working out at home. Now I'm no MBA, but maybe a thousand dollar mirror wasn't the smartest. Then the home workout boom faded.
The mirrors didn't sell like they expected and they wrote off almost all that money and stopped selling the thing. So now there's a bunch of these on Facebook marketplace for a couple hundred bucks, which makes them the perfect thing to hack. One guy turned his mirror into an AI assistant made out of glowing particles.
It shows the time, his music, the cars charging in his garage. and it even listens. And this person said they paid $100 for theirs, and it was, quote, the best project they ever did.
And then there's Kindle. This guy plugged his Kindle into his computer, let Claude Code jailbreak it, and had it strip out all of Amazon's stuff. And somebody else turned an old Kindle into a dashboard for the whole family.
It runs for weeks on one charge, and it doesn't show you a single ad, which ironically, if you want the ads removed, Amazon will charge you an extra 20 bucks. Now, people have been jailbreaking Kindles for years, and it's not just Kindles. Open source communities have spent hundreds of hours bringing the abandoned smart gadgets back to life by hand.
The difference is that you don't need to know how anymore. You just hand the task to your agent, and that's kind of the whole point of this video. Okay, now let's talk about the terrible apps, and I'll start with mine.
I recently bought two air purifiers from two different brands, one of them for my room and one of them for my kid's room. Each air purifier has its own app, and they're both complete dog shit. Now, a normal person would just use the buttons on the air purifier, but remember my one -year -old?
This guy is an absolute madman. He can't keep his hands off it. So the air purifier has to stay locked in child mode, which means the only way I can control it is through an app.
So I got sick of the apps and decided to take matters into my own hands. I spent about 30 seconds writing a prompt and a few minutes later, Claud had built me a Mac OS app called Slophouse. That lets me control both air purifiers straight from my computer.
I even added a video feed from the baby monitor. So everything's all in one place and I can pop the video out and pin it to my desktop. Now I bet some of you are already thinking it.
Does this mean the air purifier and the baby monitor aren't secure? Can any perv just hack in and watch my kids sleep? Well, no.
The app didn't actually break anything. The official apps don't talk to the device by magic.
They send the request to the company's servers through an API. They have a set of endpoints for things like turn the fan on or give me the camera feed. Cloud reverse engineered those endpoints, what each request looks like, where it goes, and what comes back.
Slophouse just calls the same ones directly. And it still has to authenticate first. It logs in with my account and password, and it gets the same kind of access token the official app gets.
And if you don't have my login, you can't get in. Okay, next example.
This guy did the same thing with his air conditioner so he'd never have to look at the remote again. Now it's just a little button at the top of his laptop. His caption, quote, no first world problems remain unsolved in 2026.
And then he put it on his AR glasses because why not? And this next one has to be the most ridiculous. This guy got tired of pushing the buttons on his air fryer, so he used REA and got his shit voice enabled.
Now he just says 190 for 30 minutes and it starts cooking. Probably onion rings. Still has to put them in himself though, so there's room for improvement.
And sometimes it's not the crappy app. Sometimes the thing just breaks and nobody's coming to fix it. The keypad on this guy's treadmill has been dead for over a year.
So he went digging, found a second Bluetooth connection hidden inside the thing, and now he can run the treadmill directly from his phone. The guy had one complaint about the AI part, quote, after four days of cybersecurity denials from both Claude and Chad GPT, meaning for four days, both AIs refused to help him because their safety checks flagged it as hacking.
So yeah, apparently cardio is a security threat now. And then there's the final boss of technology. printers.
This guy had a laser printer that worked perfectly fine, but Brother's last Mac driver for it came out in 2019, and it was only built for the old Intel Macs. Then he says the latest Mac OS stopped running it, and just like that, a perfectly good printer was dead. Nothing wrong with it, his Mac just wouldn't talk to it anymore.
So he pointed REA at Brother's old driver to figure out exactly what it does, and had Claude build a new one that runs natively on his Mac. He says it only took two prompts. Okay, so when you open something up, you'd better be ready for what you find.
A guy who goes by Log4J owns a brand new Samsung TV, and he kept getting what he called, quote, weirdly targeted ads. So he got root access to his own TV, and when he dug through the software, he found code for an always -on voice service, code for detecting faces, and sound detection that classifies noises in your living room into labels like a dog barking or a baby crying.
So riddle me this, why does your TV need to know that your dog And we already know what TVs do with what you watch. They use it to sell ads.
So how does your TV know what you're watching in the first place? Well, there's actually a name for it. It's called ACR, Automatic Content Recognition.
It's described as the TV identifying what's playing on your screen using screenshots and audio clips. Another Samsung owner went through his TV, turned that off. He also turned off the feature that keeps track of what's plugged in the HDMI ports.
Ironically, he says the TV runs a lot faster now. And this is all crazy because for decades, what your stuff actually does was secret by default. Now, anybody with these tools can take a look.
So who's trying to stop this? Well, you'd think it'd be the TV companies, but as far as I can find, Samsung, LG, Xiaomi, Amazon, none of them have said a word. If there's any pushback, it seems to be coming from the AI side.
So remember, most of this stuff didn't actually need REA. The power was never in the repo. It was in the AI models.
And whoever runs these models decides what they'll help with. Some people think the models have started saying no more often. Around October 21st, a bunch of users on Reddit claimed Claude started cutting off reverse engineering requests halfway through, hitting a safety check in the middle of a sentence.
Nobody's proven it, and as of filming, Anthropic hasn't said anything about it. And people aren't waiting around. One guy said Claude refused to root his own LG TV, so he just did it with DeepSeek instead.
And the Samsung owner who turned off the tracking, well, Claude refused him too, so he used Codex. If one AI says no, people just use another. And the guy who found the always -on voice code in the Samsung TV, well, a couple days after he posted what he found, he shared an email from OpenAI deactivating his account for, quote, cyber abuse.
He says he never used it to exploit anything, just reverse engineering. And OpenAI hasn't said what triggered it. And then Moonshot, the Chinese company behind the AI model Kimi, said it deliberately took hacking -related data out of its model's training.
But to be fair to the AI companies, I don't think anybody's colluding with the TV makers here. I use Claude every day, and I can tell you its safety checks trip up pretty easily. To an AI, rooting your own TV and breaking into somebody else's look very similar.
And it's not that simple. The same skill that opens up your TV can also find holes in anyone else's software. Sam Altman recently said, quote, I no longer think we're going to be able to just patch all the bugs.
So they've got real reasons to be careful. Who decides what you're allowed to do with the stuff you own? It used to be the companies that made it.
Now it might be a handful of AI companies. And there's a catch on our side too. REA got so popular that people started making fake copies of it with malware inside.
REA's own page now warns that a repo with a similar name, quote, does not establish that a download is official. And you've got people who've bricked their TVs letting an agent delete apps the TV actually needed. And the next software update can undo all of it anyway.
You're essentially running code nobody has checked on a device that's sitting on your home Wi -Fi. If there's a hole in it, that's a way into your home network. And when it breaks, there's no support line to call.
You are the tech support now, which is unfortunate. I was hoping to speak to someone who knew what they were doing. Meanwhile, the law is trying to catch up.
In the U .S., copyright law can make it illegal to bypass the locks on the software inside your own stuff. However, there are exceptions, and funny enough, smart TVs already have one. So does fixing your own stuff.
A popular tech news site called 404 Media said we're about to get, quote, an onslaught of semi -functional software and hardware slop. From my own experience, I'll say this. The Slophouse app I made for myself is way better than the shitty apps these companies ship with their products.
So is that slop? Well, not really. I'd say it's an upgrade.
We were drowning in slop long before AI showed up. Here's the way I see it. Companies have two paths.
Path number one, they lock their stuff down even harder. More encryption, more scrambled code, and more of it moved onto their servers. It just turns into a game of cat and mouse.
People are still going to get in. They're just going to burn more token. doing it.
Then there's path two, which is the one I'm hoping for. The companies actually lean into it. The same way they ship an app with their product, they'll also document their API so you could talk directly with it.
So imagine this, you buy a new TV and you want to run your own software on it. The manufacturer says, cool, here's exactly how to do it. And you want your AI agent talking to your air fryer?
They say, here's the end points. Go cook pizza rolls by clapping your hands. And this is what I keep coming back to.
There's no option to ship slop software anymore. Because if you do, people are just going to take it apart and make it better themselves. In the old world, the company decided what your stuff was allowed to do.
In the new world, people decide for themselves. And I don't think we're going back. So I want to hear from you guys.
Are you part of the REA movement? What have you built so far? What are you planning to take apart?
Drop it in the comments. I read them all. Gentlemen, see you in the next one.
Stay sloppy.
The Hook

The bait, then the rug-pull.

A $50 photo frame, an ad-choked TV and a $500 million fitness mirror all got a second life once regular people started handing their devices to coding agents. This video follows that wave and the questions it raises about who decides what you can do with your own stuff.

Frameworks

Named ideas worth stealing.

14:40list

Two paths for companies

  1. Lock it down harder with more encryption and server-side logic
  2. Lean in and document the API for owners and their agents

Locking down only starts a cat and mouse game, while documenting the API accepts that users will build their own software anyway.

Steal forAny product team deciding how to treat third-party tinkerers.
01:17list

Two secrets of every smart device

  1. The firmware running inside the device
  2. How the device talks to its app

Reverse engineering targets either the code in the device or the protocol between device and app.

Steal forExplaining reverse engineering to a non-technical audience.
CTA Breakdown

How they asked for the click.

VERBAL ASK
15:30link
“Are you part of the REA movement? What have you built so far? Drop it in the comments.”

A comment prompt with a casual sign-off; an earlier mid-roll asks for subscribe, like and comment framed as a gentleman's agreement.

Storyboard

Visual structure at a glance.

open
hookopen00:00
Xiaomi TV tweet
valueXiaomi TV tweet03:48
Lululemon mirror
valueLululemon mirror05:31
SlopHouse
valueSlopHouse07:33
Samsung findings
valueSamsung findings10:24
Two paths
ctaTwo paths14:40
Frame Gallery

Visual moments.

One-click upgrade to your Google

Get more breakdowns in your search results

Add Modern Creator as a preferred source and Google shows you more of our breakdowns in Search, Top Stories, and AI Overviews. It only changes what you see, and you can undo it in your Google settings anytime.

Add to Preferred SourcesOpens your Google source preferences with us pre-loaded. Tick the box and you're done.
Watch next

More from this channel + related breakdowns.