HNHacker News
TopNewBestAskShowJobs

ngneer

634 karma · joined June 8, 2015

submissionscomments
ngneer··on Meta torrented & seeded 81.7 TB dataset containing copyrighted data
Sounds just like how Facebook got started, harvesting photos without permission. From the Wikipedia article, the Facebook precursor was known as Facemash. On Zuckerberg, "He hacked into the online intranets of Harvard Houses to obtain photos, developing algorithms and codes along the way. He referred to his hacking as "child's play.""

If I were younger, I would be livid.

ngneer··on Understanding Reasoning LLMs
You might be conflating the epistemological point with Turing's test, et cetera. I could not agree more that indistinguishability is a key metric. These days, it is quite possible (at least for me) to distinguish LLM outputs from those of a thinking human, but in the future that could change. Whether LLMs "think" is not an interesting question because these are algorithms, people. Algorithms do not think.
ngneer··on Understanding Reasoning LLMs
Dijkstra myopic. Got it.
ngneer··on Meta torrented & seeded 81.7 TB dataset containing copyrighted data
That one is tough, because they are blind to the risk. I try to only work with people who have been burned before or have been around long enough to have seen the aftermath. Let me guess, they are probably telling you "show me the vulnerability", but refuse to delay shipping or fund the PoC.

Best advice is to communicate in writing the most likely risk and threat scenarios, with as much data or extrapolated data as possible. When the security flaws are later discovered, that is data you can refer to.

From what I read, this is what Zoom was like early on. They had amateur hour security and then when s*t hit the fan they beefed it up and retained a security team. I guess you could say it worked for them?

ngneer··on Understanding Reasoning LLMs
Nice article.

>Whether and how an LLM actually "thinks" is a separate discussion.

The "whether" is hardly a discussion at all. Or, at least one that was settled long ago.

"The question of whether a computer can think is no more interesting than the question of whether a submarine can swim."

--Edsger Dijkstra

ngneer··on AMD: Microcode Signature Verification Vulnerability
Yes, remote attestation is tantamount to the hardware snitching on the execution, and not necessarily to the benefit of the platform owner. I agree with you, that from a game theoretic perspective, as soon as the vast majority of platforms betray user trust, those seeking to use other machines will be left stranded. However, despite the above dynamics, you could still buy any machine you like and use it any way you like, just like app stores have alternatives. You may not like the alternative, but that is a different story.

Knowing nothing about you, it is safe to assume that you may use the dollar or similar currency, either directly or indirectly. Why would you choose to do that? Such currencies come with lots of rules and regulations about what you can and cannot do. Society must keep track of who has how much in banks, and individuals are held liable should there be an overdraft. Remote attestation may certainly put a damper on doing your own thing with your computer, but you can choose whether you wish to pay the price of participating in the economy. Not sure if the above is a great analogy, probably a terrible one, but the bygone ideal of having your computer acting solely in your interest is simply not very realistic. At the very least, controlling the microcode will not give you the control you seek, because there are too many other layers in the stack that are routinely broken, and you are always one exploit away from joining a botnet. No matter how you wish to spin it, your computer is not yours, and obeys many masters. If you feel otherwise, please offer a counterexample to the preceding statement.

ngneer··on AMD: Microcode Signature Verification Vulnerability
I am all for Right to Repair, and am upset with the "evil overlords" as much as the next guy. But, to present the treacherous computing argument all on its own, is, well, incomplete. I am a huge fan of Ross Anderson, RIP. However, the reality is that while DRTM was originally envisioned primarily for DRM applications, it is nowadays leveraged to protect PCs against bootkits and the like. YOU may like to own your computer through and through, down to the microcode, but, for most people, their computer is viewed as a product, that they trust vendors to secure for them. If ultimate control is so important to you, can you not buy any machine you like, including one based on a RISC-V open-source processor?

https://community.amd.com/t5/business/amd-and-microsoft-secu...

ngneer··on AMD: Microcode Signature Verification Vulnerability
There are multiple privilege levels. Microcode has higher privilege than an OS/VMM does.
ngneer··on AMD: Microcode Signature Verification Vulnerability
More likely a SHA variant that is susceptible to a second preimage (collision) attack.
ngneer··on New speculative attacks on Apple CPUs
Your friend is genuine in their interpretation, but there is definitely more to the discussion than the zero sum game they allude to. One can have both performance and security, but sometimes it boils down to clever and nuanced design, and careful analysis as you point out.
ngneer··on Reverse engineering Call of Duty anti-cheat
If compilation uses jmp2ret mitigation, a trailing ret instruction will be replaced by a jmp to a return thunk. It is up to the return thunk to do as it pleases with program state.
ngneer··on AI founders will learn the bitter lesson
I knew this comment would come.
ngneer··on AI founders will learn the bitter lesson
All of what you wrote is mostly true, except that "not 100% perfect in a number of fields and usecases" is quite an understatement. You mention the cybersecurity vertical. As a datapoint, I have put the simplest code security analysis question to ChatGPT (4o mini, for those who might say wait until the next one comes out). I made a novel vulnerable function, so that it would have never been seen before. I chose a very simple and easy vulnerability. Scores of security researchers in my vicinity spotted the vulnerability trivially and instantly. ChatGPT was more than useless, failing miserably to perform any meaningful analysis. The above is anecdotal data. Could be that a different tool would perform better. However, even if such models were harnessed by a startup to solve a specific problem, there is absolutely no way for present capabilities to yield a 30-50% HC reduction in this subdomain.
ngneer··on White House unveils Cyber Trust Mark program for consumer devices
The true problem is that the world is addicted to additive security. "What can we add to make our systems more secure?" Since the InfoSec industry sprung from the IT industry, the financial incentives are backwards. At the moment, people pay for security solutions that add complexity. Walk the floor in any commercial security conference and you will find these to be the most prevalent. "Take my tech" is the mantra. If people only paid for security solutions that remove complexity, then it would be a very different story. "Let me do away with your complexity" would be the better mantra. Albeit a gross simplification, the industry must flip the script to subtractive security, asking what can we remove or revise to make our systems more secure.

Same thinking afflicts consumer devices. New IoT device? Which known-good security validated and battle tested software stack is the vendor reusing? Oh, your own homebrew stack that phones home? How novel! You mean you slapped on whatever it took to ship? Terrific! There is a bug in ROM with no way to securely remediate? Shocking! /s

ngneer··on White House unveils Cyber Trust Mark program for consumer devices
Thanks for that context, quite interesting and very much appreciated. I ought to have looked them up proper, but their name is unrecognizable in the security scene.

I am not sure I understand why UL are "well suited" to provide these services. Is it that they have a compatible business model? I do not see how 130 years of solid experience in one domain means (much of) anything for an entirely different domain. Sounds like they have a nice round 0 years of experience in security. To make a crude analogy, this is like telling CERN to go start building spaceships. I mean, CERN has a proven track record for building complex things, right?

ngneer··on White House unveils Cyber Trust Mark program for consumer devices
Who are these UL Solutions? They seem to have come out of nowhere and hit the jackpot, inserting themselves as arbiters for security. Smells a bit like how Common Criteria proffered independent certification labs, which were no panacea either.
ngneer··on DOOM CAPTCHA
Strange. Wouldn't computers be better at this than humans?
ngneer··on Morris Chang and the Origins of TSMC
There is no shortage of material on the topic, but I would recommend "Chip War" by Chris Miller as a complementary read.
ngneer··on Masks, Smoke, and Mirrors: The story of EgyptAir flight 804
I was afraid you were going to say that ;). Thanks for the sage advice. I think that "safety inspector" would not be a good career choice for people like myself, then. Methinks Boeing and OceanGate had been in the news recently with similar safety attitudes. Oh, well. Live and learn.
ngneer··on Masks, Smoke, and Mirrors: The story of EgyptAir flight 804
Very interesting. Thank you for making the distinction explicit and for helping me to understand the other mindset. You are totally right, in that my mindset is closer to a forensics mindset in such instances, trying to get as close as possible to the "truth", so as to avoid future similar defects and improve system reliability. I do agree that some people prefer to manufacture truth. Any advice on how to get along with these?
ngneer··on Masks, Smoke, and Mirrors: The story of EgyptAir flight 804
With all due respect, (B) is logically unsound in my mind. You may have meant considering the hypotheses, and then using available data and only available data to rule in or rule out certain scenarios. In my mind, based on decades of studying engineering defects and failures, starting with a conclusion is not a way to solve a mystery at all. Rather, it is only a way to convince oneself of a falsehood. To give an example that is familiar to the HN audience, how many times have you had to debug a bug or problem in a complex system that you initially thought was caused by one thing only to discover it was caused by something completely different?
ngneer··on Masks, Smoke, and Mirrors: The story of EgyptAir flight 804
You make a valid point. I stand corrected. "Hacks" is not an accurate term, and fails to account for the full circumstance. I was merely appalled at how willing EAAID were to jump to conclusions and twist facts towards a convenient narrative. France is a democracy, and that makes for an unfair comparison between the two agencies. I am sure that even the most intellectually honest individual will choose their own safety if faced with the reality of imprisonment.
ngneer··on Masks, Smoke, and Mirrors: The story of EgyptAir flight 804
A rather well-written piece. My takeaway is that the French investigators are pros and the Egyptians are hacks. And that safety culture matters. One must not bend the facts to draw a desired conclusion. One must review the data without bias, or else recuse oneself.
ngneer··on AMD's trusted execution environment blown wide open by new BadRAM attack
True. Notwithstanding, I can verify that the technology stack is too complex to secure.
ngneer··on Palm’s CEO emails Steve Jobs (2007)
I suspect his ultimate reward for the call was sticking to his integrity, and knowing that his integrity cannot be compromised by a bully. Some people prefer that to money or fame. Agreed it would be nice if capitalism saw that as a positive and not a hindrance.
ngneer··on Converting untrusted PDFs into trusted ones: The Qubes Way (2013)
Cute idea. Reminds me of the format conversions typically used to lessen the risk of steganography. But boy, the article took forever to get to the PDF -> RGB idea, talking about Simple Representation and everything in between. RGB has less complex parsing, ergo less attack surface.
ngneer··on AMD's trusted execution environment blown wide open by new BadRAM attack
That whole song and dance about protection from an untrusted cloud operator sounds like a red herring to me. CSPs are the ones to deploy the technology. I suspect their main desire is to (a) avoid the hassle of law enforcement requests, as one cannot grant access to keys one does not have and (b) allay would-be customer concerns around cloud hosting versus on-prem deployments. Maybe (c) is to support compliance and regulatory checkboxes. TEEs are not about security, rather about security liability. Academic attacks are indeed just that, and not necessarily expected to be seen in the wild or affect security posture.
ngneer··on One of the last Navajo code-talkers died on October 19th, aged 107
For this to work, x=D(E(x)) must hold, where x is the message in military English, and E and D are the encoding and decoding functions. Ergo, the Navajo language must have as many possible utterances as military English does. This is neither surprising nor interesting. What is interesting is that E and D were not known by the Germans and could not be effectively reverse engineered in time.
ngneer··on HowStuffWorks founder Marshall Brain sent final email before sudden death
I used to love HowStuffWorks in 1998 as a teenager. This was the true power of the Internet. Not scrolling through cat memes, nor making billions on commerce. Just access to a wealth of good information produced by good passionate people. May he rest in peace.
ngneer··on Ryugu asteroid sample colonized by terrestrial life despite strict control
I could not tell by the comment whether the article was read or not. I try to give the benefit of the doubt. Your sarcasm is hilarious, but you are attacking an argument I never made (i.e., strawman).
← PreviousPage 3 of 16Next →