Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

DOGE worker’s code supports NLRB whistleblower

krebsonsecurity.com

Post reply on HN
Print view View original

491–500 of 586 posts

Re: DOGE worker’s code supports NLRB whistleblower

#491

Post by throwaway290 »

somenameforme wrote: post #457
dragonwriter wrote: post #433

Earlier quoted context omitted.

It's uses a simple, purpose-focused template of a type that is a common recommendation for clear communication, outline numbering, and highlights keywords using monospaced text, as is common practice in technical writing. None of that is unusual for a human, especially writing something that they know is going to be high visibility, to do. Modestly competent presentation is now getting portrayed as an "AI tell".

It has excellent presentation, excess verbosity, and is wholly nonsensical. Read the code. It uses excessive whitespace doing things like function calls/declarations with one parameter per line, and so it's probably like 100 lines "real" code of mostly tight functions -- the presentation/objections make no sense whatsoever. I was able to generate extremely comparable output from ChatGPT by telling it to create a hype…

> I was able to generate extremely comparable output from ChatGPT by telling it

Just to check, you know that ChatGPT is fully built on human writing right?

Would it be ironic if I claim "what you write looks like what the tool can output, so you used the tool" if the tool was built to output stuff that looks like what you write.

Fun fact: anything you or me write looks like ChatGPT too. It could be surprising if people didn't spend billions and stole truckloads of scraped unlicensed content including content created by you and me to get the tool to literally do just this.

Re: DOGE worker’s code supports NLRB whistleblower

#492

Post by jiggawatts »

watwut wrote: post #483
jiggawatts wrote: post #369

Earlier quoted context omitted.

A normal audit, sure. This isn’t that. This is the prison guards flipping the mattress looking for contraband. All of the public complaining is by staff that don’t understand their new position in the pecking order. There is a King in charge and he cares not for the wailing of the petty nobles.

>This is the prison guards flipping the mattress looking for contraband. No its not. These prison searches in fact do tend to find knives and what not and do in fact have some role in managing prison violence. This is not about anything like that at all.

It’s an analogy about established civilised audit procedure versus whatever DOGE is doing, which is… not that.

It’s not about literal shivs.

Speaking of finding things under the mattress, they did find corruption and waste, which they feel justifies their approach.

In my opinion the tiny amount of “waste” they uncovered is arguably not worth the damage and risk done, but that’s my opinion, not theirs.

Again: the damage and humiliation is part of their agenda. They feel that departments are “disloyal” to the King… err… President and hence ill treatment is not only justified but warranted.

Re: DOGE worker’s code supports NLRB whistleblower

#493

Post by odo1242 »

dragonwriter wrote: post #433
captainkrtek wrote: post #427

Earlier quoted context omitted.

Yea clearly AI with the keyword bolding, numbered arguments, and so on. Feel like lots of AI produced content follow this structured response pattern.

It's uses a simple, purpose-focused template of a type that is a common recommendation for clear communication, outline numbering, and highlights keywords using monospaced text, as is common practice in technical writing. None of that is unusual for a human, especially writing something that they know is going to be high visibility, to do. Modestly competent presentation is now getting portrayed as an "AI tell".

The format doesn’t itself indicate AI, but when combined with the fact that the critique is mostly nonsense it does appear to strongly suggest it.

Re: DOGE worker’s code supports NLRB whistleblower

#494

Post by bryanrasmussen »

ahwelatif wrote: post #446
dragonwriter wrote: post #433

Earlier quoted context omitted.

It's uses a simple, purpose-focused template of a type that is a common recommendation for clear communication, outline numbering, and highlights keywords using monospaced text, as is common practice in technical writing. None of that is unusual for a human, especially writing something that they know is going to be high visibility, to do. Modestly competent presentation is now getting portrayed as an "AI tell".

I'm relatively confident this critique is AI-powered. The dead giveaways: 1. Verbosity. Developers are busy people and security researcher devs are busy even moreso. Someone so skilled wouldn't spend more than 2-3 sentences of time in critiquing this repo. 2. Hostility. Writing bug free code is hard, even impossible for most. Unless your name is Linus Torvalds, Richard Hipp, or maybe Dan Abramov, most devs are not co…

a propos number 2, I think this is only a feature of seasoned developers who have managed to outgrow their own high opinions of themselves. I've met plenty of younger devs who would totally write something like this taking down the work of someone whose style did not align exactly with what they considered "good".

Re: DOGE worker’s code supports NLRB whistleblower

#495

Post by anang »

frumplestlatz wrote: post #351
threeseed wrote: post #342

Earlier quoted context omitted.

Occam's razor would suggest someone from Russia could just use their own IP because people like you would think it's a hoax anyway.

Why does someone from Russia want access to NLRB data, and why would DOGE be immediately leaking just-granted NLRB login credentials to Russian assets when it would be trivially traceable back to them, and if they were in fact granted untraceable/unlogged admin credentials, could legitimately download the data themselves and simply hand it over to said Russian assets if that was their actual intention? It's not behav…

> Whoever was attempting to log in was using one of the newly created accounts that were used in the other DOGE related activities and it appeared they had the correct username and password due to the authentication flow only stopping them due to our no-out-of-country logins policy activating.

Explains this:

> why would DOGE be immediately leaking just-granted NLRB login credential

The implication is that the credentials were for more than this specific system. It's entirely feasible that a bad actor would immediately try to vacuum up as much data from as many systems as possible, it's just that this system had a geo block that made it clear this was happening.

I don't think we need to assume that this was a targeted attack on this specific NLRB system, just that this specific NLRB system was the one that caught the attempts.

So, what systems DIDN'T block authentication?

Re: DOGE worker’s code supports NLRB whistleblower

#496

Post by arrowsmith »

ahwelatif wrote: post #446
dragonwriter wrote: post #433

Earlier quoted context omitted.

It's uses a simple, purpose-focused template of a type that is a common recommendation for clear communication, outline numbering, and highlights keywords using monospaced text, as is common practice in technical writing. None of that is unusual for a human, especially writing something that they know is going to be high visibility, to do. Modestly competent presentation is now getting portrayed as an "AI tell".

I'm relatively confident this critique is AI-powered. The dead giveaways: 1. Verbosity. Developers are busy people and security researcher devs are busy even moreso. Someone so skilled wouldn't spend more than 2-3 sentences of time in critiquing this repo. 2. Hostility. Writing bug free code is hard, even impossible for most. Unless your name is Linus Torvalds, Richard Hipp, or maybe Dan Abramov, most devs are not co…

Once you've read enough ChatGPT slop, you know it when you see it:

- Massive verbosity.

- Flawless spelling and grammar.

- Grandiose tone.

- Robotic cadence where every paragraph and sentence has similar length (particularly obvious in longer text.)

- Em dashes everywhere.

- The same few stock phrases or sentence structures used over and over - e.g. "This isn't X—it's Y", which that issue uses twice in two paragraphs:

    There is nothing "hardcore" about writing fragile, insecure, and unscalable code. This isn’t pushing boundaries—it’s demonstrating a lack of engineering fundamentals.

    If this is what was learned at previous jobs, then it’s time to unlearn it and start following best practices. Because right now, this is not just bad engineering—it’s reckless.
If AI didn't write that snippet then I'll permanently retire from internet commenting.

(None of what I just wrote is intended as a defence of DOGE.)

Re: DOGE worker’s code supports NLRB whistleblower

#497

Post by throwaway290 »

marcusb wrote: post #117
nativeit wrote: post #59

Earlier quoted context omitted.

Why wonder? The user who wrote it seems to be a pretty well established user, and their public repositories suggest that they work in adjacent contexts, so it's entirely plausible they attempted to use async-ip-rotator in one of their projects.

??? The public repos for this person that I could find that weren't forks with no activity to upstream consisted of a dice-rolling guessing game, rock-paper-scissors, and some kind of framework for downloading and transcribing audio files that does not yet download or transcribe, but implements a whole bunch of boilerplate. I find it rather difficult to believe this person engaged in a good-faith review of the async-…

I wouldn't expect somebody to use their main non throwaway account, which probably ties to their job or school to write this today... when if someone in the gov doesn't like what you say they do things like cancel your visa or sanction your employer.

Re: DOGE worker’s code supports NLRB whistleblower

#498

Post by odo1242 »

Ferret7446 wrote: post #488
progbits wrote: post #45

> Ge0rg3’s code is “open source,” in that anyone can copy it and reuse it non-commercially. As it happens, there is a newer version of this project that was derived or “forked” from Ge0rg3’s code — called “async-ip-rotator” — and it was committed to GitHub in January 2025 by DOGE captain Marko Elez. Original code: https://github.com/Ge0rg3/requests-ip-rotator Forked: https://github.com/markoelez/async-ip-rotator Code…

The GitHub part makes it... weird. You are only required to keep the GPL3 license if you re-distribute it. Putting it in a GitHub repo, is ambiguous whether or not it is re-distributing it, at least morally. If you want to delete the license in a personal copy, that is perfectly valid according to the license terms. If you then happen to upload that to a private GitHub repo, also perfectly valid. If you then happen t…

Putting it on a GitHub repo IS redistributing it. By putting it on GitHub you agree in the ToS that you have the rights to distribute the code. Which you only have if you don’t violate the license.

Re: DOGE worker’s code supports NLRB whistleblower

#499

Post by Ukv »

stevenwoo wrote: post #287
mythrowaway49 wrote: post #263

this part of the whistleblower complaint seem way worse: " On or about March 11, 2025, NxGen metrics indicated abnormal usage at points the prior week. I saw way above baseline response times, and resource utilization showed increased network output above anywhere it had been historically – as far back as I could look. I noted that this lined up closely with the data out event. I also notice increased logins blocked…

Any guesses for best possible interpretion? The Russians have infiltrated their PCs with keyloggers and DOGE are working from insecure open networks. The worst possible interpretation is straightforward - they are working for the Russians as agents and let the Russians in or installed the keyloggers for Russia.

> Any guesses for best possible interpretion? The Russians have infiltrated their PCs with keyloggers [...]

Best possible case I see would be that the whistleblower has made some mistake (or is being intentionally dishonest). Seems plausible for instance that "it appeared they had the correct username and password" based on "our no-out-of-country logins policy activating" could just be a misunderstanding of how/when the policy triggers. Not to say it's the most likely explanation, just the least concerning one.

I think less concerning than keyloggers, while still assuming the whistleblower is correct, would be that a DOGE employee was using a VPN/proxy/Tor. Probably not a great idea to have traffic going through a hostile nation state even with encryption, but less bad than keyloggers on their machines stealing and trying credentials within minutes.

Definitely concerning though, to be clear - just steelmanning/answering the question of best possible interpretation.

Re: DOGE worker’s code supports NLRB whistleblower

#500

Post by arrowsmith »

roywiggins wrote: post #399
grandempire wrote: post #317

Earlier quoted context omitted.

> I dont think POTUS can What data in a federal agency could the chief executive not have authorization to access?

I am fairly sure it would be a crime for the President to pull up someone's VA health records on a whim, or at least it would be a crime for anyone at the VA to facilitate him doing that. We can also add to that IRS data. The articles of impeachment against Nixon included the following: "He has, acting personally and through his subordinates and agents, endeavoured to obtain from the Internal Revenue Service, in viol…

"When the president does it, that means that it is not illegal" - Nixon

I wonder, if he was alive today would he stand by those words?

Post reply on HN