Using AI to reverse engineer hardware
blogs.gnome.orgBoth Apple and Nvidia are the most hostile and have the most locked-down heavy hardware and software and now Claude can reverse engineer and build a driver to get Nvidia GPUs working on Intel Macs [0] and it can unlock the ability to pair an Apple watch with Android phones. [1]
Now it makes even more sense to target reverse engineering CUDA, Nvidia firmware blobs and Apple software to finally make them run better on Linux without waiting for those vendors to fix their bugs or deliberately slowing their own hardware down.
No more excuses.
[0] https://github.com/nullmoth/nvidia-macos-driver
[1] https://www.androidauthority.com/apple-watch-companion-pair-...
Of course if it's cheaper, someone might, but the real barrier is really the whole first-party support, it's not the technical stuff.
I was hoping someone would do color printers inkjet and laser.
You can, however, use their equipment as a dumb modem and insert your own router downstream. It thankfully supports DMZ to expose your router directly to the internet without NAT. I use Ubiquiti.
If someone can do it I’d appreciate it, but it won’t be easy. Their whole business is spying on you.
Others like Samsung are unfortunately way worse, hard to secure.
I run a small pro audio company, we make guitar pedals and audio effects mostly. My current key objective is building up a processor similar to the Eventide H90, that would compete on having top-notch algorithms for reverb, delay, chorus etc. we already have single-effect pedals on the market that are very successful.
In the past 72 hours I have ONE-SHOT decompiled 3 of my main competitors products, including the Strymon Big Sky. I just... have their award winning algorithms now. Like, I could in theory just stick them in a box and resell them. It took one prompt and 3 hours for opus 5.5 to work that out.
The lexicon 224 then got added. While Claude was prepping It found someone else had already done the same thing 1 weeks earlier and dumped the result on GitHub, so it just stole that code and saved some time. It took all of 20 minutes to do the entire Ursa Major product range.
And yes, I've tested them all - and they all sound phenomenal. I specifically asked Claude to first build an emulator and execute the original firmware/ROM so I could be certain it wasn't just making "something similar" but actually executing the same code. Then I asked it to convert it into a self contained c++ library and compare the result to the emulator.
The problem is, everyone can now do this to my products as well. Basically, anyone with a Claude subscription can now rip of any best-in-class algorithm from any competitor. While I'm not going to do that, I know there are 30 Chinese companies that happily would.
The moat of hardware effects processors has just vanished.
Intellectual property incentivized many and laid many golden eggs, but why would someone toil to create a new algorithm to solve a problem if someone else is just going to steal it in hours?
But who knows, maybe imploding the software industry is a good thing. I may lose my house in the process, but I think we could all get back to touching a little more grass.
When that happens it shouldn't be forgotten that people (broadly speaking) chose this. AI proponents lit intellectual property on fire out of self-interest.
Regardless of whether or not we agree with it, adherence to our idea of IP law is not something the whole world agrees upon, even before LLMs.
Sometimes they're actually true.
Repeatedly tell people climate change is a serious problem for over half a century -> hand wringing.
Make solar panels cheap -> renewables become 99.2% of net new electricity generation in the US.
You need something that looks like the second one.
GP's argument is usually used in the sense of "if I hadn't done it, someone else would, so I might as well do it and the end result will be the same, except then I'll reap the rewards".
It's a crap justification for doing something unethical. But in many cases, even the unethical ones, it's also true, and I think in this case it is too. If the US had magically shut down every AI lab on day one, and banned all AI research, that wouldn't have stopped this train. Probably would have slowed it down, but not enough to really matter.
Also that age-old rationalization has generally always been true.
I think it goes deeper than that, in that the AI era will eventually force us to rethink how we value each other. It used to be economic scarcity: I provide a service that very few people can provide, therefore I make a lot of money (supply and demand.) Now we're moving to a place where most anybody can do anything, so our value system needs to evolve, as well. Instead of valuing scarcity, perhaps we should value personal creativity and self-expression. Kind of like paying money for a concert, or for someone's personal painting, even though you can listen online, or you can just have AI generate the same painting. It comes more from wanting to support people than to supply scarcity-based demand.
I think this idea makes many peoples' eyes roll, but IMO that's a testament to where we are today as a society. Most interactions get reduced to just business transactions without the humanity in it, and that's where our problems come from.
https://www.wsj.com/world/china-economy-gen-z-f9058941
https://www.economist.com/china/2019/11/14/homelessness-has-...
https://www.youtube.com/watch?v=Si2zT-wCs3c
We all know that "real" communist societies like North Korea are worse, much, much worse. In fact within China how bad this is varies enormously. If you're in a "rope village" somewhere deep inland, having the finest house in the entire city is arguably worse than homelessness in Beijing, but the party won't let you leave except as homeless, because you're allocated a "Household" (Hukou) and you aren't allowed to leave.
And, of course, half the problem is the very structure of the communist party itself: local officials hide the problem, because if they don't, they'll be purged or at the very least won't see promotions, ever. Which means that the police will be using violence to hide the problem rather than fix it.
AI can provide incredible wealth, or immeasurable misery, in both capitalist and communist societies. That is not what makes the difference, that is just how political parties like to frame themselves and so the argument why you should choose "left" or "right".
Capitalism and private home ownership can certainly provide housing, but if you go too far or not far enough in private ownership and "protecting capital" they won't. And communism can certainly fix homelessness (the Soviets did, for example), except nowadays they just won't.
AI will provide more tools for people with capital, which primarily will be the government. Whether that solves, say, homelessness is a choice, in either system.
I don't claim to have a solution for what comes after but it is certain that capitalism will collapse if we ever see significant job loss due to AI and robotics.
The Luddites said the same thing.
The only way AI and robots end up replacing most jobs is if they can do them for less money than people can. Which in turn means that those things become less expensive, since that's the thing causing suppliers to do it.
So there are really three possibilities here. First, we get robots that can e.g. build houses and housing scarcity goes down, so now you don't need to make much money to afford rent. Second, robots still can't build housing, so then you get a job building housing and have a job. Or third, robots still can't build housing but we have a bunch of rent-seeking incumbents that won't allow you to get a job doing that either, because they're trying to keep rents high or artificially limit the supply of trade labor.
It's the third one that really screws you and is the one that we need to prevent.
As long as nobody is stopping them from building houses for the people who do want them.
Let's say 50% of all jobs are lost. What would that kind of oversupply do to pay rates and working conditions in the remaining sectors?
Imagine if there were no intellectual protections for pharmaceuticals. The government would have to take over R&D.
You could just as easily argue "why are 200 companies burning through people and natural resources to come up with the exact same thing and then marketing it as different" as you can say "why would anyone create anything without the ability to making oodles of money by restricting imitation".
You are the one who needs to wake up from the assumption that the only way to live is to hoard your intellectual products instead of imagining a society where you can stop having to compete on that level.
You pay $1000 to a monopolist instead of $100 to a small business and then having the other $900 to spend on whatever you want, like patronizing 9 other small businesses and actually getting something more for your money. The $900 in monopoly rent goes to Wall St instead of Main St and is used to e.g. bid up housing costs so people can't afford to buy a house.
The incumbents lobby for restrictive zoning or professional licensing gatekeeping to keep those rents high, which means fewer jobs doing construction.
It doesn't help you for huge corporations to get more of your money.
On one hand, yes, IP = monopoly = rents, as you say.
But on the other hand, no IP means economy of scale is a lot more important, and that favors huge corporations.
Which effect is bigger? I don't know.
"IP" is the thing that favors economies of scale.
Suppose you're a manufacturing company and you want to make your own computers. You're great at making the chassis and screen and batteries and things like that, but to make a whole device you also need an operating system, a compiler, a web browser, a CPU, etc.
If you're free to use existing ones then you have a chance. If you have to develop every one of those things in-house because every existing one is the property of an incumbent who is using it as a moat then you have to be a megacorp or you're sunk.
Notice that this is what all of the smaller companies actually do. Every small consumer device OEM and some of the big ones are using some variant of Linux/Android or BSD rather than developing their own OS. NeXT, the scrappy company Apple bought and turned into Mac OS X when they were themselves the underdog, was based on BSD. RISC-V has shipped more than 30 billion cores and is taking a decent chunk of the embedded market from ARM, which is itself a company that licenses its designs to companies that compete with each other (and run Linux on their chips) rather than using exclusivity and mandatory vertical integration to try to monopolize the whole supply chain. Which is how the mobile market went to them rather than Intel and Microsoft. Who themselves took the desktop and server markets from even more vertically integrated mainframe companies.
The availability of composition is what allows you to be a small company that does one thing well. Which is the thing you can't do when someone else has a monopoly on something you don't make but need, and correspondingly isn't willing to sell it to someone who is trying to compete with them in a different part of the supply chain.
Moreover, the economies of scale you're talking about on the other side actually have diminishing returns. Suppose anybody who wants to make a product that typically sells for $100 has to buy a $10,000 machine in order to do it, but then can make an unlimited number of them. If you're so small that you only ship 100 units, the cost of that machine by itself is adding $100 to the cost of producing of each unit and you're pretty screwed. But if you're a medium sized company and you ship a million units, it's only adding one cent. The advantage of the medium company over the small one is dramatic. Whereas the megacorp that ships ten billion units shaves it down to 1/10000th cent, but then the advantage of the megacorp over the medium one is still less than one cent.
On top of that, it's often something like you need a $10,000 machine that can make a million units even if you only use it to make a hundred, and then the scale advantage of the medium company over the small one is all still there, but the advantage of the megacorp over the medium one doesn't even exist because then to make ten billion units they need ten thousand machines.
Which is even before we get to diseconomies of scale: The larger an entity is, the more efficiency you lose to bureaucracy, coordination costs, principal-agent problems, internal politics, etc.
There is a reason industries with low barriers to entry tend not to monopolize. Why is there no monopolist who grows all of the coffee or weaves all of the clothing? Shouldn't "economies of scale" allow a single huge company to out-compete all of these little ones in the absence of anyone having a way of excluding competitors?
Conversely, find a consolidated industry and you'll generally find some kind of government-backed monopoly or regulatory capture.
The same reason I have open sourced every line of code I have written in the last several years, even new inventions. Because it is good for society, and good for my reputation to show what I am capable of which has been very good for my career.
Better to be the one to open source it before someone else does.
The only business model left is convenience tax. Save people time and they give you money.
I feel that LLMs are destroying the reputation building benefits as well. It used to be that most people's published code was a result of their personal effort and thought process. This is no longer the case. Not only are many new contributions written by machine, but the very process of understanding and reasoning about code has been devalued.
Which is why LLMs are going to kill most of the software industry. Software products and services operate on this assumption - package some functionality into an opinionated UI that makes some kind of work marginally more convenient, slap a brand on top, and you now have a captive audience and the perfect upsale/ad delivery vector, i.e. the user interface.
LLMs flip that around, because suddenly, your UI is an obstacle sitting between the user and the problem they want to solve.
Throughout history, the golden goose has typically been owning capital, and having market reach (see: advertising, the silk road).
No matter how amazing your IP is, if you can't produce it at economically viable scales, and if you can't put it in front of people who will buy it, then you have nothing. So venture capital comes along, buys you out, and takes all your future profits in exchange for the "risk".
On the other side of the coin, there's open source software developers, PhD's publishing papers for free, and your local musicians that like to play at events just for funsies.
I mean, imagine the kind of world we'd live in if medicine were focused strictly on profitable moats instead of open research, and... oh. Right.
same reason I spent my entire career so far writing open-source code that anyone can copy ? (it's GPL, but imagine 100% companies actually caring about that). Ideas can't be owned.
as did the internet before it
> but it may just end up killing entire industries.
how many record companies have been put out of business by piracy?
> Neural Amp Modeler (is open source)
> JUCE on embedded, Rust DSP
> A board of guitar pedals with IDK one USB-A and a USB-C adapter with OSC and MIDI support; USB MIDI trololo pedals
There are multiple products like this on Amazon+Shopify w/ linked inventory now
It sounds like it will go faster if you use coding agents.
Anyone copying you will have to do the same.
Perhaps spend a month with your agent making it harder for other agents to reverse engineer?
It’s not ideal. It has serious drawbacks. It makes development and updates a PITA. But it does make it much harder to decompile your firmware.
The tricky part is figuring out the workflow, once that's in place, you can encrypt the protocol itself as much as you want, the output (audio, video, JSON, whatever) has to go somewhere, and that can be captured and replicated.
It’s not clear to me what else one can do to protect this kind of IP if you want or need to do the processing on-device.
Enough with patents, NDAs and trade secrets.
Also, I guess with agents it should be possible to explore new algorithms more efficiently. You may discover something cool! And yes, anyone can copy it, but you will still have first mover advantage. Also, in music tech brands matter a lot. People love getting stuff from small “artisan” companies.
Just be different. Just create a brand. Oh if only it were that easy!
Out of curiosity, did you just use vanilla Claude Code with Opus 5.5, or did you use this: https://github.com/morluto/rea?
If your firmware is open for someone to reverse engineer now, it has always been open. The real moat is firmware security, encryption, secure locked down hardware. Nobody, not even LLMs, can reverse your firmware if they can't obtain a cleartext copy of it.
You must use microcontrollers with security features, encrypted firmware, secure key stores. If you don't, you never had a moat in the first place. Your only security is the belief that it's too much effort to reverse your product. Security through obscurity is the moat that has vanished and everyone has been screaming for decades that this is not secure by any stretch of the imagination.
Take an audio stream, apply STT, interpret the result as an algorithm, run the algorithm, take its output, apply TTS, then send that audio stream to the output.
A moat isn’t a perfect defense. It’s a delay tactic that keeps the enemy from coming right up to your wall and climbing it or breaking it down. While you defend your moat, you whittle down the attackers so even when they do cross it they can be more easily defeated.
Moats and high walls became obsolete when canon fire became common.
Claude isn’t just a faster way to cross the moat, it’s gunpowder, and with it the rules have fundamentally changed.
Firmware security is useless if your product can be bought tested and reverse engineered through multiple trials. So long as you have something that can be understood and used by humans, you have something that can be reverse engineered.
Having the code accessible makes it easier, but just having both input and output is sufficient.
Moats and high walls, sure. But to this day, the absolute best way to stop any attacker is to dig a trench (most minus water) and to build low parapets, primarily of dirt, sand, barbed wire and ideally some anti tank stuff.
The moat never really went away.
LLMs only appear to do that, if they do it effectively is because you can.
if people like your stuff, you'll have a reputation because of what you made in the past, then you can offer to release your next project to the public domain if some crowdfunding goal is met.
Part of the issue is that people can feed the idea, descriptions, and pictures into the AI. Thus create knockoffs, for their crowdfund or as the basis for their similar software.
If the original person ever actually releases software, they can buy a copy and feed to the AI, to then beef up the features in the previously created similar software. The difference between the original and copycats can be few to just about none.
Unless the original creator is willing to fight it out in court, there is little they can do about the situation. Updates made in the original, can be "transferred" to the copycats. AI makes the speed and quality orders of magnitude higher.
Before the Singularity: build good products, offer them at a fair price and support them well, succeed at business.
After the Singularity: build good products, offer them at a fair price and support them well, succeed at business.
It sucks if you’re a rent seeker, but I’m enjoying this post-IP world. I don’t care who reads my source code anymore, code is disposable. It’s expertise that matters now.
To add to this, the Second Industrial Revolution started in 1870 and the 40 hour workweek became common starting around 1920. That's basically a lifetime.
For a short while, it's great, but in the long run, it makes it difficult to financially invest in building software with either time or money.
Uhh I think you're missing a tiny 3 trillion dollar detail. I understand though, something that small and inconsequential is easy to overlook.
$10 a month to unlock the world 6 months before the same is available for free. $200 a month if you're in a hurry.
The fee is flat, and all the value produced stays with the user.
It's one of the best deals on anything software industry produced ever. And they can't walk back on it, because open weight models act as a rachet.
Oh yes, this new world is terrible for rent seekers. It's like the kum-bah-yah free software world Stallman might have imagined. Information wants to be free and all. Can't stop the signal. Very counterculture. Like the Adobe clone apps.
Except for the closed SOTA model weights, out of reach hardware, and the resulting $200 subscription rent now paid readily both by anybody trying to create something new, and anybody who wants to rip it off. That particular flavor of rent seeking is booming.
I myself am a "rent-seeker" in an alternative sense of "I seek a way to pay for my own continued shelter via my creative output" and it definitely seems like it's gonna suck for me.
SOTA only provides agents that are easier to manage. Such agents will likely also be made available in open models, as it is primarily a post-training issue.
Even if it never comes, the best open models are already capable of performing these tasks.
I know that it's much easier to get JLC to make a board and then flash some stolen firmware on it than it is to build a tube amplifier, but musicians are weird creatures.
All I know is, I would rather be in the before times.
I don't care either way but depending on where you live this might get you in trouble
That sucks that your business and mental well-being are threatened by AI. It's a scary time for everyone. I hope you figure it out.
Compute is the only true moat. Ownership of high performance computers will be the deciding factor in the future. Haves and have-nots will be determined by this.
It was a fun project because it predates DSP chips by a couple years (1978). So there's an 8080 host driving some DSP boards made out of 7400 series chips.
Bio of the designer with a photo of him prototyping it: https://www.davidgriesinger.com/biographies/
I haven't written it up yet, that part is still hard :)
I paid someone on Upwork to trace the schematics into KiCad. Had LLM agents turn those into Verilog models. Lots of little mistakes had to be found and fixed at this point, it took months...
It seems like agents are a lot better at transcribing and grokking schematics now compared to 6 months ago, so it might be much easier now.
Agree that consequences for the future seem depressing...
I have written probably a dozen software synthesizers over the last 20 years, for fun, never profit. I would have preferred those to be hardware products. But I lack all of the above. As do most.
I have worked in several embedded companies. The software that runs on them was often considered to be the least important part of the expense chain despite involving the most amount of total work hours.
The only winner (kind of), is the AI companies, with "token addicts" and subscription fees. But even that is in doubt, because of the data center debt[1][2] many are in and artificially low prices (to create addicts).
Unless laws are put up quickly to protect IP, which is doubtful (as courts and politicians are mostly clueless about technology), this has become the new weird reality.
[1] https://archive.ph/wWY2g (Investors dump data centre bonds as AI borrowers face ‘Darwin moment’)
[2] https://www.youtube.com/watch?v=KChHY1kDOFQ (AI Bubble | Ed Zitron)
Protect IP from what? The fundamentals of the situation have not changed. The only difference now is that the models are doing the work.
Replace 'models' with 'humans', and you get bog-standard RE. The models just accelerated the timeline where this scenario would've happened.
The solution you propose already exists in its primitive form: It's called Section 1201 of the DMCA. What you're asking for extends that protection to the whole program, not just the DRM, and thus killing all of reverse engineering altogether.
------
Copyright as it exists is no different than real estate: Landowners extracting rent from properties they own, in exchange for promises on routine maintenance & upgrades.
What we have now is a machine that can look at a building & remake it to at least 90+% similarity, given enough iterations.
And we're discussing about banning the machine.
No matter which laws are used in the attempt to protect IP, it requires lawyers and going to court, and the little guys will usually have neither the money or time to fight.
The moat a physical pro audio product is its interface and its branding. And I know that's tough to hear especially if you're an engineer but its true. People buy Strymon pedals because they know exactly what they're getting and they know its going to be in a format where if you want to dial in a tone without getting into the weeds of menus you can (but if you need to advanced features you can also do that). If people wanted the cheapest way to get those sounds, they could always hook up to a live VST host rig. But you get Strymon because you know that brand name and you know you're going to get a box that you can twiddle the dials on the floor and yes you are also getting that as close to perfect algorithm as well. The sound is part of Strymon's branding but it isn't all of their branding.
Look at it this way, pretty much every original circuit design in the analog guitar pedal space has been copied, duplicated, sold for pennies on the dollar, modified, whatever. But those companies still exist and do well because their moat has never been the circuit, its the branding.
Anyway, I hope you find some peace in the next few days. Its hard to argue that your moat hasn't shrunk and maybe you're in the process of building up that customer base. You gotta think about the copycats and it sucks, but you always had to do that. Its just harder now.
And in live music, this is rather critical; some musicians get it instantly, some need a single gig (even in a rehearsal) with their computer setup going dark or blocked or whatever to get it.
Or: "Can you turn this tablet I got from AliExpress into an iPad?"
It's going to be interesting to watch the big companies struggle with this.
The printer 'debugging' was one of the funniest and most WTF situations I've been in lately. I used Claude for the drivers and was essentially used by it as a meat sock to diagnose what's happening. Was the text on each line progressively less visible, ie did grayscale work? How about quality between page 1 and 2, the first one should be lower DPI. Ok now unload the paper so I can see the error and add some handling for it. Good, now load paper again we'll check the duplex.
If you are reasonably familiar with computers, it really is as simple as pointing a strong model at a device and burning some tokens. Feels like a fever dream, especially since I had very little idea about anything that has to do with hardware.
Wish I could release the apps I've "made", but worry about DMCA
I wish someone would Fluxer [1] GitHub for self hosting.
It is a bit of a pain to make an anonymous GitHub account. That's literally what this whole identity is (I'm making a Nitter style frontend for a meditation app and I'd rather not take any chances). At this point just use a residential proxy like the Telegram people.
[1] Fluxer is an excellent most promising privacy focused Discord clone.
The UI is almost an exact copy of Discord, which, despite what people will say, is actually very good. More privacy focused alternatives should rip the UI of their competitor if it's good so that users get the benefits of more privacy while knowing exactly how the service works.
That sounds interesting. What's the usual approach, install the app in some emulator or something, and ask your agent to reverse engineer it? Or is the setup for the agent to work simpler than that?
Also managed to bring back compatibility old printer with Linux and even unsupported one working only over bluetooth!
Before this era, you could imagine how it could work but would have rarely taken the time to do this grunt work. Now you ask, wait for a few minutes, do so meat proxy integration testing, iterate a few times and here you go.
Brings some hope for sustainable computing!
One of the best things you can do in these situations is to actually audit the code that was contributed against the code you have. If you have more than was contributed, you could potentially still make a contribution of what's missing, against the structure of the existing contribution.
Soldered UART pins on it, plugged my USB adapter into my computer, then asked an agent to reverse engineer the firmware blob. Took DeepSeek v4.1 flash around an hour to get a full OpenWRT build-and-flash pipeline up and running, using an isolated podman build container and everything.
The outdated OEM Kernel made it easy to identify the SoC hardware. drivers and proprietary Wi-Fi and Bluetooth driver mods (not really legal btw, because they "conveniently forgot" to publish GPL'd source code). Driver mods were successfully reversed through analyzing the .so and .ko files with frida in another podman container.
Overall took me less than half a day to do all this, and now I have OpenWRT running on that device, compiled from source with the upstream modern kernel.
To me, observing all that, and how quickly it found encryption keys, signing mechanisms, encodings, binary formats, instruction sets, disassembling it etc by just looking at the hexdumps ... this would have taken me weeks if not months.
And now comes the kicker, all that cost me 0.86$
AI and reverse engineering is like tap water now. There's no reason to not consider using it, because it's so accessible. There's just absolutely no point in proprietary hardware and software anymore.
My current project after this: I've soldered the UART pins to a proprietary Xiaomi smartphone I had laying around, and I'm currently letting it reverse engineer the baseband firmware. Let's see where this ends up ...
In a few hours yesterday, I built a macOS menu bar app [0] that ramps up the volume when the USB output is triggered. This is not reverse engineering but an effective workaround.
I am tempted to try to reverse engineer the speakers and see if I could fix them myself. But 1) I have a feeling they have security features enabled and 2) I would rather not brick this rather expensive pair of speakers.
Probably it does come from the manufacturer not caring and just releasing the bare minimum, whereas hobby developers can, and want to, spend more time on getting things perfect.
As an example, to show off my LLM-reverse-engineered-hardware project: a user-space driver and software for an old FireWire to USB box, the manufacturer drivers for which were last updated in 2006 and don't really work anymore on modern OSes (and never worked on Mac). Firewire these days in general is also incredibly buggy (needing legacy drivers for PCIe cards, software that wasn't updated since 2006, etc).
With a modern app, even something written in a few days, is already so much more pleasant to use and less buggy, than _all_ the old options that people were using before.
The hardware is quite unique because it's the only widely sold device(-family) that had a DV (FireWire/1394) connector for connecting mid-2000s camcorders, allowing you to transfer video via USB on computers that didn't (don't) have a FireWire port. And internally it works by packing the digital DV video stream into USB, without converting it, so you get the raw, full-quality video data as it was on the tape, the same you'd get it with a "real" firewire setup.
I guess the downside is that if people find out that this old, unique hardware is useful now, it will become expensive and hard to find.
It's here in case someone ever finds it useful: https://github.com/JonasCz/MarvinCapture