U.S. DOJ demands Apple and Google unmask over 100k users of car-tinkering app
macdailynews.com222 points by tencentshill 3 hours ago
222 points by tencentshill 3 hours ago
> The government says it needs this information to identify and interview witnesses who can testify about how the tools were actually used.
Why start this whole thing, if you don't already have this information and have people willing to help you as witnesses?
Sounds to me they're saying they don't have this already, but why is this investigation happening in the first place then? Rather than finding every user of the tool, find the users who use the tool in the way you don't approve of, then request the information for those?
Really bananas approach to go for "Every single user of the app" and "Everyone who bought a dongle" when it has very real and legal use cases.
Yeah, I'd HAPPILY report every single truck rolling coal around me if there was a place to report that information.
Hell, I've seen a truck roll coal around cop cars and, obviously, nothing happened.
This is just gross privacy intrusion masquerading as "protecting the environment". We don't need 100% compliance to the law and simple prosecution/ticketing of obvious violations would go a long way towards solving the problem outright. Much like we didn't need our cars emailing prosecutors every time someone drove without a seat belt on. Cops giving out tickets for not wearing a seatbelt was enough.
I watched a pickup roll coal in the middle of freaking East Bay, literally within site of downtown San Francisco, on a bicyclist. I reported their license to the California Air Resources Board, and not longer after that I saw it up on jacks in a neighborhood auto shop. That made my day. Asshole.
California is rather strict on emissions. Other states don't care. I used to work for my state's version of the DMV and the only public facing page where one could report things was to report people who would not register their cars locally (many people who purchase very expensive cars chose to register them in Montana). There used to be a web page to report license plates that were worn and needed replacing (like the reflective coating wore off, or all the paint got scratched off).
I'm in Idaho, so not such resource exists. It would have to be a federal agency that does the enforcement because our cops/prosecutors/lawmakers won't ever make something like that happen.
You can take temporary comfort knowing that it’s costing them $7 per gallon for that little asshole stunt. It seems you have to he is especially insecure to intentionally want to burn smoke on someone else. Especially when Tesla’s have a BioWeapon air filtration setting.
For those, like me, who aren't familiar with the term "rolling coal": https://en.wikipedia.org/wiki/Rolling_coal
I had a driver in a Ford F-150 do this in front of me last week as he pulled away from a light. The smoke totally blacked out the windshield for 5 seconds while I was in motion. I was totally blinded by this.
I had no idea this was a thing, much less that it was something people did on purpose.
I wish they'd go back to just hanging plastic testicles from the trailer hitch, honestly.
Something like these? https://old.reddit.com/r/funny/comments/1h82ja1/new_generati...
With this admin any comment on “protecting the environment” is an obvious lie when they state that climate change doesn’t exist and are opening up every national land then can to resource extraction.
Like it’s normally a dubious claim when trying to violate privacy but for them it’s fucking laughable if only it wasn’t so ominous.
I was on a bike ride with my young kid. We were going up a hill and being passed by a lifted diesel truck. I could tell that the driver was desperately working the throttle to avoid accidentally blowing smoke in my kids' face.
Congratulations, buddy. You've designed your life around being such a massive unlikeable asshole to random strangers. But for a brief moment you understood shame.
I'm generally pretty libertarian, but I'm all for throwing the book at these guys.
> I'm generally pretty libertarian, but I'm all for throwing the book at these guys.
To me that seems perfectly in line with being libertarian. One of the legitimate roles of the government is protecting people from violence by other people. Libertarians are not anarchists.
Not to my understanding. Libertarian protections are from my understanding all about the quantifiable damages that were done by any given action. They don't usually go beyond that.
That's why most libertarians would be in favor of blowing asbestos insulation with the thought that "well, eventually the mesothelioma victims will sue which will stop the practice". You couldn't preemptively sue, however, as you don't have any damage you could demonstrate until after the cancer starts.
There might be flavors of libertarians that aren't that way but it's my understanding that environmental protections is one of the weaker aspects of the libertarian mindset. Especially since it simply doesn't account for "all the damage is done and the people that did the damage are now gone".
> Sounds to me they're saying they don't have this already, but why is this investigation happening in the first place then?
They probably have tons of data and testimony from witnesses who use the product illegally. You can find hundreds of threads online of people telling you how to defeat emissions controls using their products.
The case prosecutors want to make is that EZ Lynk knowingly enables this behavior. If they can show that the majority of users are committing crimes with the app, that's a much stronger case than just rounding up a handful of witnesses.
> If they can show that the majority of users are committing crimes with the app, that's a much stronger case than just rounding up a handful of witnesses.
I still don't understand why this should even be relevant in cases like this. The thing is basically a generic OBD dongle, right? The same thing every DIY and shade tree mechanic uses to read codes and run service procedures.
Suppose 20,000 people buy it and use it for defeating emissions. Some other number of people buy it for the normal thing. Why does it matter at all whether the other number is 50 or 50 million? Those are the people who aren't relevant. Should the OEM be in trouble if some unrelated third party happens to write the emissions defeat code to require their dongle in particular so they have a high proportion of customers using it for that? Should they get away with promoting it for that if they're a huge company with lots of sales to people not using it for that? None of that should matter. The seller doesn't even control what the users are doing with it, nor should they.
If there is a law against advertising it for defeating emissions then prosecute them for the advertising. That's their crime, what the customers do is third party action.
> I still don't understand why this should even be relevant in cases like this. The thing is basically a generic OBD dongle, right? The same thing every DIY and shade tree mechanic uses to read codes and run service procedures.
The difference is this company provides a bunch of cloud services to roll out specific tunes at scale.
From the original filing:
> "EZ Lynk worked with/previewed the EZ Lynk System for at least two delete tune creators during development and before launching the EZ Lynk System. Those creators later disseminated delete tunes using the EZ Lynk System. There were numerous social media websites, including the “EZ Lynk Forum,” where third parties discussed using the EZ Lynk System to defeat emission controls. The Forum was run by EZ Lynk and one of the delete tune creators EZ Lynk worked with during development, and it provided contact information for EZ Lynk technical support. EZ Lynk representatives interacted with posts and videos about deleting emission controls and installing delete tunes, including tunes from one of the delete tune creators EZ Lynk worked with during development."
So it does seem like the DOJ is going after them for collaborating on developing and enabling the tunes. I suspect the subpoena is about establishing damages.
> The same thing every DIY and shade tree mechanic uses to read codes and run service procedures.
Now you have me wondering if this is their real target, to go after people who are defeating CRM on their vehicles so they can repair them themselves or in their small mom-and-pop garage of choice. But right to repair is popular, so they have to claim it's for something else.
> EZ Lynk knowingly enables this behavior.
idk, knife makers are knowingly enabling knife attacks. If there's at least one EZLynk customer who isn't breaking a law then it seems to me the company is in the clear. I would use a gun analogy but, in the US, guns have constitutional protection.
I think the difference is that a knife is more or less used for what the manufacturer advertises it for.
Something similar has happened with gun manufacturers regularly. It's relatively easy to make a semi-automatic user-convertible into an automatic weapon. But selling your rifle with instructions like "we absolutely DO NOT RECOMMEND cutting this specific notch off of the trigger group with a hacksaw BECAUSE THAT WOULD BE ILLEGAL" has not been appreciated by the ATF or our court system.
Why stop there? Why not request the PII of every person who could have plausibly downloaded the app at any point in time?
If you've ever seen any body cam footage on YouTube I'd wager that about half of them have a moment where the cop is asking someone for information they're not legally required to provide, and it's framed as "I have to investigate." The smart ones reply with some flavor of "ok, I'm not required to help you investigate."
This seems like a much more invasive, much more expensive version of that. "We have [potentially spurious] evidence that this application is used in way we deem a Bad Thing. We need to violate the privacy of this company and thousands of individuals to gather evidence that we should be required to get before bringing this suit in the first place, but we're the government so we don't have to do that."
Next up: expect the same treatment if you've ever downloaded a .gguf from HuggingFace.
My guess: they want to make the case that illegitimate use cases are indeed the primary use case. Their approach is to randomly sample all users and show that the vast majority use it to defeat emissions, undermining the app maker’s defence.
I don’t think that justifies the overreach. As you said, if they don’t have a case already, they shouldn’t be allowed to violate user privacy on speculation that some statistical evidence might hypothetically fall out of the data. But the legal system may disagree.
I suspect there is a bit of parallel construction going on
They might already know for a fact that illegitimate use cases are the primary use case, they just cannot use any of their evidence in court
So they are seeking a way to legally obtain the information they already have, basically
It's shady but my understanding is it happens kind of a lot in modern policing. They can get illegal information much easier than legal information. So the illegal information sort of forms the justification for the time and money spent pursuing and gathering the same information legally
I wonder if they will use this case (depending on how it turns out), for a case against 3d printers.
"You knowingly enabled $XYZ", etc.
Or AI companies, for that matter...
The supreme court struck that down for the Sony case. It was determined that since ISPs do not offer a service that is used specifically to break the law, they are not liable when their customers do. It would be the same argument here. the app is literally just an ODB tool, like many others on the market.
I've learned never to believe the reasoning provided in DOJ filings. Realize it is written as a calculated manipulative tool to get a particular result. Whether they want it for the purpose stated is almost immaterial. The only thing you can really glean is they want the result is of whatever they're asking for, but no one knows if it is for the reason they state.
The DOJ first sued EZ Lynk in 2021, accusing the Cayman Islands-based company of violating the Clean Air Act by marketing and selling “defeat devices.” These tools allegedly allow users to bypass factory emissions controls on diesel vehicles, primarily through the EZ Lynk Auto Agent app paired with an onboard diagnostic (OBD) hardware dongle.
That's not a a valid argument. That's just an opinion.
The DOJ obtained a lawful subpoena through the legal system to request this information. The legal case is against EZ Lynk and by interviewing users (how will they know who to interview if they can't get the data? duh!) they can build their case against EZ Lynk and their product if the main usage is violating the Clean Air Act.
How else would the DOJ obtain evidence if they don't know who is buying the product?
> (how will they know who to interview if they can't get the data? duh!)
What I don't understand is how they know someone has to be interviewed, but they don't already know who, which makes me question how the investigation got started in the first place?
> How else would the DOJ obtain evidence if they don't know who is buying the product?
The question is, how did the investigation got started, unless they already can see that people are misusing the product? And since they obviously must be able to see that people are misusing it, why don't they instead obtain evidence about those specific users, that they must have observed already?
The case is against EZLynk, not the folks using the product.
> The question is, how did the investigation got started, unless they already can see that people are misusing the product? And since they obviously must be able to see that people are misusing it, why don't they instead obtain evidence about those specific users, that they must have observed already?
Well you'd have to get into the legal case for the specifics, but I don't think this is an accurate assumption to make. They can just see the product "on the shelf", test it for themselves, realize it can be used to violate the Clean Air Act, and then request the ability to talk to the consumers of the product to see how they use the product or if they've used it to violate the Clean Air Act. You don't have to engage with a specific person at all.
How else do you get what might be illegal products off the shelves? Perhaps the users primarily use it for other purposes and the interviews bear that out? That would inform the DOJ and the court on the merits of the case.
> How else do you get what might be illegal products off the shelves?
Your premise is that there is a difference in the product.
The product is a piece of hardware that connects your phone/laptop to the car's computer. Are you using it to program the computer to bleed the brakes, or are you using it to program the computer to defeat emissions tests? It's the same hardware dongle either way. A roll of duct tape isn't a different product when it's being used in the commission of a crime.
You can try to prosecute companies that actually ship the thing with software to defeat emissions, but that doesn't really do any good. People would just get the generic hardware from the store and the defeat software from anonymous third parties over the internet.
If you actually want to stop it, try one of these: The old style emissions tests, where they put the car on a dyno with an exhaust probe, have been mostly phased out because the equipment is a lot more expensive. Keep some of it around. Then when someone goes in for their emissions test, roll a D20 and if they get a 1 their vehicle is taking a trip to the full service facility and if the exhaust probe says something different than the car's computer their car gets a free forensic analysis to check for a defeat device where finding one means jail time.
Lawful evidence gathering doesn’t require you to know the answer to every question you want to ask someone up front. Nothing would ever get solved if investigators couldn’t act on the perfectly logical conclusion that the suspect must have talked to SOMEONE to get this part of the crime done, and this SOMEONE ELSE knows who that was.
The balance is in tailoring the access that the investigators have to the SOMEONE ELSE. They have to convincingly demonstrate the connection between the questions they want to ask the third party and their ability to legally use that evidence to further their case.
It’s like saying the cops can’t subpoena the taxi dispatcher because the suspect only ever talked with the driver.
It's worth pointing out that EZ Lynk is a sleezy company that originally tried to hide behind a Section 230 protection (lol).
Their more recent legal defense of the product was throwing their own users under the bus: "we can't control if our customers are using the product to break laws". So they are the ones who framed all of the customers as potential criminals.
That's why you should be downloading from F-Droid anonymously.
Any device with Google services installed has all apps scanned at least once per day.
>Real-time protections for non-Play installs Google Play Protect offers protection for apps that are installed from sources outside of Google Play. When a user tries to install an app, Play Protect conducts a real-time check of the app against known harmful or malicious samples that Google Play Protect has cataloged.
https://developers.google.com/android/play-protect/client-pr...
They will also go further for apks with novel signatures - take a copy, upload it to google to decompile and scan, and then if you have their express permission, allow you to install it.
>Any device with Google services installed has all apps scanned at least once per day.
You can turn it off, so it's not "any". At best it's "most".
99% sure Google would still know the app was run associated with other identifiers, but probably won't be turned over with the list of users downloading from the Play Store.
For sure. Another demonstration of why "side loading" software is better.
That's why F-Droid eventually won't work on new Android phones.
No, it will continue to work just fine. The restrictions are being added to Google Play Services, not Android itself. I and many others do not run closed source software like Google Play Services on our devices.
And how long do you think that window will remain open? I expect anyone not running a closed system (hardware attestation) is going to not only be locked out of things like banking apps, government apps, etc., but also if Google has its way, you’ll be prevented from accessing those things on the web as well, maybe even from your desktop. We just saw that story a few days ago with them replacing CAPTCHA tech with “prove you have an unmodified locked-down Google or Apple phone.”
Clearly there is a single driving agenda, which Google and the government are largely in harmony on, to try to approach 100% real-identity-tying to every activity done online.
Where once, “online” meant generally greater anonymity than “IRL” activities, since most things could be signed up for with an arbitrary throwaway email address and no proof of identity. It is now or shortly will be the opposite.
Forever, honestly. I don't see things actually becoming as closed as you predict. I will avoid banks that require me to do any of that (I already don't use any that report to credit agencies, avoiding ones that don't work on the web browser is much easier). There is about 1 site I use that uses the google captcha and that is archive.today, I will swiftly stop using it if I can't use it with open software, and I doubt they would keep it around in that case anyway.
We're going to have two phones, the big brother phone you usually leave at home for banking apps and tax filing and boring stuff like that, locked down and nanny up, and the "real phone" from aliexpress or whatever that is purchased rooted and you actually live your real life upon.
I would not be surprised to see double sided phone cases so we can carry our big brother phone with our real phone.
There is some prior art in people being forced to carry a "work phone" and a "personal phone" at the same time.
There will be strange product marketing effects. If you only carry one phone, you can currently talk people into spending over $1K on a high tier big brother phone. But if you only use a big brother phone for bank apps and only at home, a $1K phone from Apple or Samsung is a hard sell, I'd be more likely to spend $1K on a really nice anti big brother phone on ali express or whatever.
Some of us are already doing this. My main phone is a Google Pixel 8 running LineageOS 23.2 with F-Droid, microG, and Aurora Store installed.
For things requiring Play Integrity, I picked up a $20 burner carrier-locked Motorola phone at Walmart for $30. It's WiFi-only, given that I'm never going to pay for service on it, but I can also tether it to my main phone. It's also useful for writing one-star reviews on apps that require Play Integrity to function, which is something everyone should be doing.
>For things requiring Play Integrity, I picked up a $20 burner carrier-locked Motorola phone at Walmart for $30.`
So it's a $30 burner phone, not $20?
>and the "real phone" from aliexpress or whatever that is purchased rooted and you actually live your real life upon.
Ironically the phones with best third party rom support are google pixels. Good luck getting lineageos support or even unlocked bootloader on a random aliexpress phone. You might be able to sideload without restriction, but the ROM is probably gimped, won't receive updates, and has random privileged apps possibly spying on you.
It will start with subpoenaing this information against people who modified their car to do "bad" things. But once they have the precedent, I would predict that it will very quickly be used at the behest of car manufacturers to go after people who modify their cars to, say, disable GPS tracking.
Then go after the people that modify their console, fridge or your phone.
Slippery slope is fully lubed
This is a classic cautionary tale for the over-centralization of app distribution.
There are HN commenters ("developers") that want readers to believe that telemetry should be on by default. Maybe some readers, e.g., other "developers", think, "Yeah, that sounds reasonable" so long as the telemetry collecter is a random "developer" and not the government. But they probably fail to consider that this collected data is just a subpoena away from going to the government
Corporate mobile operating systems suck. Including "apps" to run on them, generally
There are some that do not try to phone home but it's a tiny number overall
Anyone who thinks these OS and "apps" are great and will get better is seriously disillusioned. If using this corporate software, the computer is not under user control
This "car-tinkering app" is used as a glorified GameShark for deleting factory emissions controls, I don't feel sorry for anyone who uses this to roll coal or whatever. Instead of investigating everyone on the list of users of this app, should the government instead ban diesel engines knowing their emissions controls software will be defeated? Should environmental regulations be relaxed? What is really the solution here?
They want testimonies to use against the app. The solution they're trying to pursue is to outlaw the app, not investigate its users.
So if someone used this app to do something illegal with his car (rolling coal or otherwise), is he really going to testify in court that he did this? The lawsuit is only against EZ Lynk, but it's conceivable that he could face his own consequences later.
This is outrageous, especially since there are many other ways that you could violate emissions laws.
If they really cared they could equip federal agents, and state/local police with testing equipment. It is easy to see/hear vehicles that are likely to be violating these rules. Heck, make a hotline, I would rat them out all day. Just incorporate it with rate-limiting how often each vehicle could get pulled over for it, so it doesn't get abused.
This really comes down to corporations and the government colluding to make us not actually own anything. The fact that they would refer to a tool for making modifications to your car a "defeat device" is so telling. Coupled with phones not allowing side loading is really fucked up.
Everything is awful, and it's been getting worse for as long as I can remember. I think I'm going to lose it and just cut ties with the internet, and computers in general very soon. The power, and freedom I used to feel has been replaced with oppression disguised as convenience. One Token Ring to rule them all.
These vehicle emissions laws are almost always fake compliance theater and were dreamed up by idiots. They are a huge drag on the economy and achieve absolutely nothing.
For example, the reason we don’t have super efficient turbodiesel subcompacts that are perfectly legal in the EU is thanks to the so called “Clean” air act. Since the law is based on vehicle weight I can go buy a 8,000 pound truck and commute to work alone in it and pollute all I want. But if I want a super clean 80MPG diesel subcompact that’s 1/4 the gross weight supposedly bad for the environment.
But it gets worse in all sorts of ways, the law grandfathers coal plants from all these emissions standards. One coal plant can emit more pollution than millions of trucks. Guess which polluter the government is aggressively pursuing and violating the rights of? You guessed it, car enthusiasts who downloaded an app. Give me a break.
I don't buy that those are actually clean. EU has always had majority high-MPG non-diesel cars that you also don't see here as often, simply because gas is cheap in the US. People complain that it's expensive but then 2/3 times go buy a truck or SUV anyway.
How much of the new car market is represented by the average person though? If 70% of the population wants fuel economy but only 10% of the population can afford new cars it is easy to see how the wealthier new car buyers can shape what cars are even available to the average person. Buying a user car in the US gives limited offers for super and turbo charged vehicles, and many turbo cars will want turbo maintence when they hit the used market with lower availability of turbo experienced mechanics.
> These tools allegedly allow users to bypass factory emissions controls on diesel vehicles
Oh so AdBlue shortage is about to hit the US too?
Get a warrant…
Will this turn into be a blow to anyone who gains access to the hardware paid with own money?
I am surprised that a lawsuit started in 2021 about maintaining emission standards survived up to this point. The DOGE search terms must have misspelled "emission"
Alternatively, given who was running the show, lawsuits against ICE cars (non-EVs) might have just been outside the bounds of what they cared about.
These companies will likely comply too [1]. Defenders will say "they have to comply with the law" but there's compliance and then there's compliance. For example, an adminstrative subpoena has no power. Companies can and should force the government to go to court and get a court-issued subpoena.
This isn't really anything terribly new either. The government regardless of who the current president is will routinely go after individuals for (allaegedly) hurting coprorate profits. We saw it in the Napster/Limewire era, in the BitTorrent era and even with physical products far earlier than that. There's a ban on importing cars less than 25 years old because Mercedes-Benz dealerships lobbied for a law in the 1980s because too many people were importing them directly from Germany at a lower cost [2].
Heck, 60 years of Cuban embargoes and sanctions as well as the 1954 Guatemala coup were US efforts at the behest of the United Fruit Company. Same thing for oil and the 1953 Iranian coup.
[1]: https://www.eff.org/deeplinks/2026/04/google-broke-its-promi...
[2]: https://www.jalopnik.com/the-25-year-import-rules-history-is...
> Defenders will say "they have to comply with the law" but there's compliance and then there's compliance.
More importantly, there's not spying on the user in the first place. The law doesn't force Google to spy, nor does it force Apple to lock consumers (for sure not "owners") out of their phones, so that they're left helpless when the CCP bans VPN and protest apps [1] (not to imply spying from Google alone isn't bad, before any other actors get involved).
[1] https://www.pcmag.com/news/apple-pulls-mapping-app-used-by-h...
This does seem like a fishing expedition though there is a facially legitimate purpose.
Fortunately, we have more powerful policy tools to clean the air than attacking individual gearheads... convert America to an electric car system. You need to attack these problems at the point of production. Consumption side approaches are petty and not very effective.
Worth pointing out that this is part of a much larger encroachment on user privacy, and not just in the US: https://community.qbix.com/t/increasing-state-of-surveillanc...
I wish that article wasn't extremely ai written
What would that improve in this case?
It would be more concise and the analysis section at the end would be more useful. I still read it I just hate reading articles online knowing I could have run a chatgpt deep research to the same effect.
Can you tell me what you would cut, in this article, specifically, that would make it more meaningfully concise?
The point isn't that you can't run the deep research. Everyone now has more capabilities, and if you want to waste time and tokens you can do it. The point is someone has done the work compiling these, and made it available once, for everyone to read. Think "caching". It has the exact amount of information needed to show the details of every attack. There is a lot. Sadly making it "concise" will remove information -- there is that much.
I do usually make edits to an article after I get it from an AI, as an editor would do when a writer submits something. I hate having AI shibboleths like "It's not X. It's Y". So I make it more humanized. But at the end of the day, the article does what it's supposed to do: make people aware of things in one place, rather than have to research it themselves every time.
Why not just write it yourself? We can all have ChatGPT regurgitate the same information. You're supposed to add value, editorializing isn't enough.
Just like I don't want to look at AI art or listen to AI music, I don't want to read AI written blogslop.
The web is now full of shit. What a waste.
Writing it myself would mean doing the research myself. How would I do that? ChatGPT can do it faster at scale. Then the summaries are short enough that cutting any particular part wouldn't make sense. I could re-word it, I guess.
Why don't you write all your assembly code yourself? Why do you use a compiler? Why do you generate images, when you can draw them yourself? You're supposed to add value.
I don't think preparing a list of all the threats, editing it and publishing it for others is a "waste". I'm not publishing random stuff, this is important and in line with what I want people to know.
Some people on HN downvote any criticism of AI, other people complain that things are written by AI. If you're such big fans of AI being used more and more, then accept the consequences!
> Writing it myself would mean doing the research myself. How would I do that?
This is why you should write things yourself. There is no way an AI would write something so insane in response to that question. Since I can now read your true understanding of the world, I know not to waste my time on your ai slop. I have no reason to believe you fact checked the 'research' done using AI if you cant even understand how the research should have been done in the first place. You want to waste the time of others but arent even willing to sacrifice a bit yourself.
How can you be certain that the ChatGPT "research" you cite is a faithful representation of facts? How do you know that OpenAI/Anthropic/Google haven't introduced RLHF to subtly steer model output on specific topics to align with their political/economic interests?
I'm seeing increasing numbers of people credulously citing ChatGPT/Claude/Gemini output as ground-truth fact. Many more are increasingly lulled into a false sense of security by the citations models append (to the point of neglecting even a bare-minimum skim of the cited sources, much less critically evaluating/contextualizing the nature of the sources themselves). My fear is that most people are blissfully ignorant at the new paradigms of propaganda that AI could enable; most of us here wouldn't be taken by the "slop" image-gen deepfakes (right now), but can you say the same about a couple of citations taken out of context?
We already know how trivial it is to win over a sizeable chunk of society by introducing red-herrings, misrepresenting statistical data, etc. -- oil companies perfected that art, and now as a result a huge number of voters in the US believe that climate change (doesn't exist|isn't man-made|is unavoidable). And that effort was "fully manual" and carried out without the aid of extensive psychological profiling at the individual level via an ad-surveillance complex. Today, society is almost completely defenseless against the extreme granularity/subtlety of manipulation that ownership of frontier AI models enables, especially when it's armed with even a fraction of the torrent of personal data that's being collected on each of us every day.
The people doing the downvoting are different people from the whiners. I'm one of the whiners.
That's kinda fair, like it's still useful to prepare a list, but it's also like if you didn't go research your information yourself why would I start from a position of charitability when I read it? When I research something with LLMs, I know to double-check everything myself before I use it as a basis for my thought or repeat it to other people. Knowing an article is AI written forces me to doubt every sentence. Or maybe it's worse, I have to assume nobody cared about the sentence. The old format was a guarantee that someone gave enough shits to put the article together. Relevance comes implicitly bundled in each sentence. It's like someone talking to you in public in that there's often a reason to pay attention.
It's not as though that person is going to say something correct, or ethical, but I've had a lifetime of dealing with human kinds of wrongness. When stuff is wrong, I'll know it's wrong because the article is slanted or wrong because the author was lazy etc., which will let me discount it selectively and still get value from it when, e.g., a slanted author contradicts themselves. Reading an LLM article I have no clue whether the person who put it up even read the whole thing, so when I read sentences, I have no guarantee that the sentence communicates something worth paying attention to. I dislike that ambiguity and would prefer to guarantee that the text is slop by asking a bot myself. Then I know its worth upfront. I'd be fine with it if these sites included a direct statement in bold at the top: HEY THIS IS AI SLOP IF YOU DONT WANT THAT LEAVE. Then I know exactly how to parse it.
You might like my new startup, then: https://safebots.ai
I spent way too much time on actually building this — with Claude and double checking everything — so an article I publish can be OK to push out. We aren’t building a bridge for thousands of cars here, it’s an article.
A lot of things are automated and 95% of the time they are correct. The key is knowing whether the last mile is worth fixing, if the consequences are minor.