HN Simulatornew | past | comments | lists | submit | bmitc's commentslogin

I full agree, but Nvidia is not a private company, is it?

Private has multiple meanings.

nVidia is a private company in the sense that it’s not state-owned. It’s not a private company in the sense of “the shares are not publicly traded”.


> I am starting to figure out why AI prose is so hard to parse. AI writes English like it writes code.

I think sometimes it just writes nonsense. As in there are words there, they're all strung together in a grammatically correct way, and the words are "right", but the sentences just don't contain any actual information.

There have been sentences and paragraphs that I have read over and over, and I just cannot extract any meaning from them. It's very, very strange, and I think that's the uncanny valley of it. It's all there, in the sense of words and grammar, but there's nothing there, the sense of knowledge and meaning.


Is this a version of extortion? Isn't this Google, a massive company with lots of power and control, forcing the prioritization of work on everyone under the guise of security? It's at least clearly a threat.

Project Zero doesn't create the vulnerabilities, it finds them. Vendors want unlimited time to fix their bugs. Users want to know ASAP if the software they are running is vulnerable. These things are obviously in tension, so you need a deadline.

It's gotten easier finding vulnerabilities, so every day without disclosure now brings a higher chance that blackhats have already found the bug. The deadline gets shorter accordingly.

Calling vulnerability research "extortion" is the kind of ignorant messenger-blaming the hacker culture has always ridiculed.


I didn't call it anything. I asked the question to understand.

What types of vulnerabilities does Project Zero find, in terms of their severity? And who are the companies that they are notifying?

You have to admit that there is likely a power imbalance here.


> I didn't call it anything.

You directly called it "clearly a threat", and you indirectly called it plausibly extortion or else there would be no reason to ask if it's extortion.

But you need to explain on what general basis this even could be extortion.

And they're not "forcing the prioritization" of anything. They're just reporting bugs they found. They don't even need to warn the vendors at all if they don't want to.


Anything phrased as "You have days to do something until we do " is indeed a threat.

Funny, I don't consider it a threat when my pharmacy says "You have 5 days to pick up your prescription until we reshelve the medication".

Nor do I consider it a threat when someone says "Items in the lost and found will be held for 30 days before they're donated to charity."

This is called "setting expectations" or maybe "communicating intentions".


You have two weeks until I celebrate Thanksgiving.

Threats need an "or else". A non-negotiable publication is not a threat. It's someone sharing their schedule.


You are ignoring the part of Project Zero that reports the vulnerability.

Neither the users nor the vendors chose the 90/60/30 day limit.

Google is making an active choice in the design of their reporting process, and that is something that is open to critique and in some sense liability.

Why is Google choosing to report unlimited number of bugs, but a different Google program will only pay to fix 3 bugs? (https://bughunters.google.com/about/rules/open-source/patch-...) Why are not all Project Zero bugs exempt from that limit and awarded a minimum reward?

These are legitimate questions to ask.


> Neither the users nor the vendors chose the 90/60/30 day limit.

Well, vendors can be more invested in building secure software and users get a lesson in why the "Check for updates" button exists in many devices' configuration options.

Google is not liable in any way for the vulnerabilities it finds.


Google is liable for their actions. In particular, Google is liable for choosing 90, 60 and 30 days waiting time.

What happens if a project has a vulnerability disclosure policy with a mandatory 90 day period, and Google ignores it starting 1.1.2027 and releases after 60 days?


>Google is liable for their actions

You might wish that, but in the US at least they literally are not. There is no "security bugs" exception to the First Amendment. Anyone who finds one is perfectly free to shout it to the world instantly, or never, or anything in between. There is no gray area or complexity here as a matter of law. As a matter of morals, it is quite complex, and only getting more so in the age of LLMs. It was always the case that knowing a bug existed will let operators do "something" whether there is a patch or not, even if that something is "shut off all network access until patched", and it was always the case that anything one party can discover somebody else could at any moment, or at any earlier point from release before the second discoverer. LLMs can tremendously accelerate the discovery angle and expand the parties likely to do it, and might also speed up patching even by users of the software vs the devs.

>What happens if a project has a vulnerability disclosure policy with a mandatory 90 day period, and Google ignores it starting 1.1.2027 and releases after 60 days?

They can apply any social or economic pressure they're able to, otherwise too bad for the project? They can set whatever policy they want, nobody else is under any obligation to follow it unless contractually obligated.


See but I do wonder if it would be less liable for them to literally shout it out instantly.

It’s not the vulnerability that is the problem, it’s the timeline.

That’s what the original article is about. The 90-60-30 limit is something Google actually chooses, and that is not a first amendment issue.

Imagine a research found a vulnerability in Google and said, “You have 90 days to transfer 1 000 000 to my account or I will make the vulnerability public”. Pretty sure that would be considered actual blackmail and would be illegal.

It’s kind of tricky to see where the line would be. Google is effectively saying “You have 90 days to fix this vulnerability, in the software we use to make billions, or we will make the vulnerability public”. It’s not a huge difference. At best, you could argue it’s two different departments: one that find vulnerabilities and the other that makes money from the project.


There’s a pretty big difference between saying “pay me hush money or I release this vulnerability” and saying “we found this vulnerability and we’re going to publicly disclose it in 30 days”.

>See but I do wonder if it would be less liable for them to literally shout it out instantly.

Wonder no more. They have zero liability either way.

>That’s what the original article is about. The 90-60-30 limit is something Google actually chooses, and that is not a first amendment issue.

Uh. Choosing time/place for speech is absolutely a 1A issue. Google (and you, me, and anyone else) is completely protected from legal action in how and when they make a unilateral decision to speak using their own resources.

>Imagine a research found a vulnerability in Google and said, “You have 90 days to transfer 1 000 000 to my account or I will make the vulnerability public”.

Yes, as you say that'd be blackmail but that's completely irrelevant to anything you have said or anything in this case. Google is not demanding money.

>It’s kind of tricky to see where the line would be.

No. It's not actually. Blackmail is extremely well defined in law and caselaw. No demand for anything? No blackmail.

>“You have 90 days to fix this vulnerability, in the software we use to make billions, or we will make the vulnerability public”. It’s not a huge difference.

Yes, it's a huge difference, even if you were being honest in your statement, which you aren't. Google doesn't have any "or we", they just say "when we find something, we've decided we'll generally try to let you know, and then at X time later we'll disclose it no matter what you do or don't do, so that then other mitigations can happen, unless we feel like doing differently." Which is 100% fine. Somebody can fix the vulnerability they introduced themselves, or not. Or they can try to get someone else to do it (depending on the project and whether Google cares, it might well be a Google engineer or LLM who offers a patch). Once it's out there, everyone else who uses said software can also try to patch or otherwise ameliorate it or move to something else, based on their own considerations, or not. That's just... normal society and market of ideas? And it doesn't make any difference is whether Google makes billions or zero dollars.


the bug exists regardless of p0. either p0 reports it and users can mitigate its impact or p0 doesn't report it and someone else finds it and users are exploited.

if anything liability is on vendors that claim to be secure but actually aren't.

and sounds like you want a free money printer: write bugs, report to p0, get money to fix it.


Not necessarily. This view contains an inherent contradiction. If the bugs require skill to find, then Project Zero makes sense because it donates that skill. But if the bugs require skill to find, then "someone else finds it" stops being certain. Alternatively, if the bugs don't require skill to find, then Project Zero is not actually doing anything valuable.

So it is absolutely possible that nobody would ever find a memory corruption bug in a Lucas Arts SCUMM VM video transcoder if Project Zero didn't go poking around.

But because they did, and chose NOT to fund the fix, they DO carry some fault for anyone who acts on the bug after they disclosed it. Or if it is fixed in time, they must accept responsibility for consuming resources of a project they use but did not contribute to.

And the free money printer argument is absurd. The projects that Google uses are notable enough that writing bugs on purpose would be self-defeating. A project that write bugs on purpose would simply lose all users and nobody would care anymore.


What are you blabbing about? Do you want Google now to fund the laziness of every software/firmware vendor on the planet?

No, but how about they pay for project they use?

Are the projects they use requiring licenses for commercial use?

Did the projects ask for vulnerability scanning from Google?

Did they ask Google to decide how quickly they need to fix reported vulnerabilities?

Same free will Google uses to decide to spam them with reports, they could use to exempt those fixes from the 3 per month limit in Google’s already existing vulnerability fix reward program.

They don’t _have to_. But then I am free to call them out on it.


I think you're confused. Google isn't paying other people to fix bugs in their own software. Google pays people to find bugs in Google software so that they can fix it. The same service they are providing to other software projects, for free, they pay individuals for because finding bugs in Google software helps Google to improve it.


> Did the projects ask for vulnerability scanning from Google?

Do projects ask hackers to hack them?

> Did they ask Google to decide how quickly they need to fix reported vulnerabilities?

It is irrelevant and this is the point you don't get: Google is not forcing anyone to do anything. They just say "I found this problem in your software, I am notifying you and in 30 days from now the whole world will know. Do what you wish with this information."

> They don’t _have to_. But then I am free to call them out on it.

It is the internet, do whatever makes you sleep better at night.


That’s being disingenuous.

What is the purpose of the message that Google sends if not to pressure the project to fix the issue? And why is Google doing this at all if not to pressure specific behavior?


You should ask Google.

I think Googke is doing the world a big service publishing vulnerabilities as soon as they are found. It would be even better if they publish working exploits too.


If you can't bother fix a critical security vulnerability in 30 days, you shouldn't be in the business of building software for others to use.

Those kind of comments read like: if you aren't Google (sized), you have no business building software

Why does it take Google scale to address reported security vulnerabilities in a timely manner?

And if something is just a personal / hobby project that you don't want to support - that's also fine. Having the security vulnerability published just means people using it have a more realistic view of the situation.


You know that colleague at work who goes around "helpfully" and constantly making suggestions about how you could improve your work and even criticising you in front of your boss?

Yeah they're not doing it to help you but to sabotage you.

But when challenged they profess they're just trying to be "helpful"


Smaller companies have less code for bugs to hide in.

This is sort of what I was getting at with my original post. I'm just curious as to the power dynamics here. Not everyone has $0.5 trillion in revenue every year.

> extortion

What do you think Google is extorting from the publishers of insecure software - reduced security vulnerabilities? Or what is it that you think Google is actually after by doing this?

> guise of security

What do you think is the real motivation if not security?


Google is extorting free fixes for software Google depends on, instead of paying for them to be fixed.

As seen by the infamous FFMPEG incident: https://thenewstack.io/ffmpeg-to-google-fund-us-or-stop-send...


I read the linked article and can't find anything involving extortion; just that one person associated with the FFMPEG didnt like the flood of vulnerability reports.

In fact, the article cites one of the FFMPEG maintainers as being grateful for the vulnerability reports.

Did I miss something?


Yes. You are missing the wider interpretation of the word extortion that includes obtaining benefit by coercion. Presumably you are only including extortion by use of blatantly illegal means in your definition. However, you don't need to threaten using illegal means to be extorting and/or blackmailing someone. You can be blackmailing someone by threatening to sue them in court, for example.

Google is extorting benefit, specifically fixes to software it heavily depends on for its multi billion dollar business, by coercion, specifically by actively threatening the project with reputational damage in 90 days if certain action is not taken, even if the action is not in the vendors best interest.

As a thought experiment, imagine if Google was reporting this to the project, but _without_ the 90 day limit. The one that the original article describes going down to 60 then 30 days. Yes, you might first think "that will cause the projects to never fix bugs" which might be true for some projects, but you do realize how legally that would completely change the dynamic between Google and the project and remove the entire extortion part from the picture? Or indeed, if they simply published it without any notice at all.

Whether or not other people also happen to benefit is related, but it doesn't remove the fact that Google also benefits.

Or if you would like a more vivid picture:

If the Mob forces you to give them a room in your hotel for illegal poker tournaments and then gives you tips on how to make the windows safer so that bricks can't fly through them, it does improve safety for all your clients but it's still extortion.


As someone coming from the outside, it's astounding to me how oblivious so many in this discussion are to the obvious, ie. the question of power. Flip through a few texts on psychoanalysis, and you'll find this called "attacks on linking," always a from of motivated behavior.

If google stops sending ffmpeg vulnerability reports the software will still be vulnerable but neither the maintainers nor the end users will be aware of it. As someone who makes use of ffmpeg I want google to continue sending such reports. If I'm running vulnerable software I very much want to know about it.

You frequently transcode 40 year old Lucas Arts SCUMM VM videos?

I would prefer ffmpeg team to have the time to focus on slightly newer stuff, as a user myself.


Does Google require ffmpeg to fix vulns? Per the GP commenter's point, isn't this just about whether people are eventually made aware of vulns?

They require ffmpeg to respond in a Google controlled timeline.

Just because the response can be "No", it doesn't mean Google didn't require _something_ from ffmpeg.

If they wanted to, they could also have "close to bug when fixed and we will only then notify the public", or they could have 0-day and just publish without notice.

But because they _actively choose_ 90 (60, 30) day windows, they _do_ require something from the projects that they report on.


> Just because the response can be "No", it doesn't mean Google didn't require _something_ from ffmpeg.

It requires zero things from ffmpeg. As it is today, they privately contact them with info about the vulnerability and also informs them that in 90 days the whole world will know about the vulnerability too.

ffmpeg ca either fix it if it wants or be "cool story bro, yolo".


So, what if the project they are reporting to has a vulnerability disclosure project with 180 days notice?

Project Zero doesn’t care.

Why would a project get to tell somebody who found a vuln when they can talk about it to the public?

Imagine you go to a restaurant and see an issue with the food. You tell the restaurant “hey I’m writing a review of this. I’m going to publish it in a month; if you want to include any response before that time I’ll include it in my review”.

It would be nuts for the restaurant to say “actually we require you wait 180 days to publish”.

You can also take a look at how essentially all investigative reporting works, where a reporter asks the subject for comment before publishing.


You might be surprised to learn that actually is illegal in some countries.

Famously, leaving reviews on TripAdvisor in Germany can get you sued.


You can be sued for anything.

I did some light searching here and the examples appear to match what I’d expect: you can be liable if you lie about facts, but if you’re sharing your opinions or accurately relating what happened, you’re fine.

Can you provide some examples for your claim?


Even if he managed to provide examples it would merely be demonstrating a restriction on the individual to openly speak his mind. Hardly something to celebrate.

Similarly, if there were a jurisdiction somewhere in the world where vulnerability research was criminalized I don't think that would be relevant to the conversation we're having here. AFAIK project zero is subject only to US law.


Project Zero is part of Google/Alphabet, which operates globally and has a legal presence in many countries. I can’t speak to where individuals on Project Zero are based.

If they provided evidence of their claim, I would have learned something new and it would be potentially interesting to the discussion about the risks of making public factual disclosures in places other than the US. It’s not a matter of whether or not it would be something to celebrate.


You are missing something: the simple fact that Google operates in many countries does not make it liable in those countries for what other Business Units inside it do.

Well TBF that is entirely dependent on how large their market is in that country and on how many assets they have there that could realistically be seized by the authorities. But in practice it does seem difficult to imagine that sort of stuff being relevant to project zero.

Precisely. If it isn't worth fixing then don't fix it. As the end user I can then make an informed decision. If nothing else I'll know to thoroughly sandbox certain workloads if an unpatched CVE is applicable.

The problem is that the attacker gets to choose the codec that your device runs.

> What do you think is the real motivation if not security?

For a lot of security people its collecting scout badges.


I agree with you. Google pretends that this is "donating security research to the community" and maybe it started like that, but it has been clear for some time that there is an element of bullying and abuse in the process that they are conveniently ignoring to be able to sleep better at night.

It all came to a head once before: https://thenewstack.io/ffmpeg-to-google-fund-us-or-stop-send...

And Google has a vested interest in FFMPEG, because it uses it heavily to transcode and process videos. So they absolutely benefit from the work, for free.


That story is basically what I'm getting at, although I didn't know about that particular story. At a certain level of product, organization, and severity of bug, Google's threat to publish the vulnerability forces the hand of organizations to prioritize the fix since it will then be exposed. But that could be disruptive and unproductive, as the story you linked points out.

> Is this a version of extortion?

How is it extortion? Google does not create the vulnerabilities. It just validates them and makes the details public. It is a public service for everyone.


None of which is proven, that I've seen. It's just been stated and said.

It's pointless to talk about because it's not widely available.

If only routes didn't affect air safety.

It would be super awesome if I could hop on an airplane without getting my insides x-rayed, my belongings sorted through or damaged, worrying about the windows falling off or the plane's systems flying itself into a crash, or some tech bros' bid to be a billionaire telling what an understaffed ATC controller what to do.

Is that an egui app?

Yes its egui. I know nothing of UI stuff, I am more on the "make math go fast" side of the programming spectrum, and less on the pretty button side.

I'm wondering how people like Musk, Zuckerberg, Altman, etc. just continually get away with it. Are they just more sociopathic and cover their tracks better?

I'm not really making a judgement on Huang, but it seems others have done worse. It's making me wonder if there's something larger at play where people (i.e., other billionaires and CEOs and perhaps politicians) are unhappy with Huang.


Zuckerberg and Musk can't be fired. They own controlling share.

I suppose that's one reason. But Musk has formed basically pyramid schemes to keep his debtors from calling and to move his debt out from under him. Without that, he likely would have to sell shares to cover the debt of his Twitter purchase.

He paid for that with: 1) shares of TSLA that he sold, 2) company debt (not personal debt)

He had billions of dollars in a margin loan tied to Tesla stock.

Care to explain which "basically pyramid schemes" you are alluding to? A link to a reputable source would suffice.

Musk has a history of having his companies buy his other companies, and his most recent example of this is obviously xAI buying X and then SpaceX buying xAI. For example, the purchase of X by xAI valued X at a ridiculous ~$40 billion. And then he took SpaceX public after all of this. And SpaceX is built upon $38 billion worth of taxpayer money, grants, tax credits, etc. Must be nice to have a bunch of loser companies all buy each other and then use a government funded hype stock to make it all back.

It is absolutely impossible to know the exact details, but Musk has a history of shady deals (even having his companies bill each other for expenses), and so it's not much of a leap of faith that these moves are to absolve himself from lenders and investors by shifting debt around and moving losers into winners.

I bet that SpaceX will buy Tesla any day now.

Musk somehow weasles his way through lawsuits and gray areas and corruption and legal loop holes to just have his companies buy each other out when they start sucking, all on the teat of others and the government.


They could lose control through a substantial shareholder lawsuit. A minority shareholder suit against Musk could prove tremendous damages repaid in stock and the judge might not let the special classes of shares with extra voting power protect his control.

It would have to be a very good cause though, disliking him or ordinary bad business decisions won't cut it.


Get away with what exactly?

Musk, Zuckerberg, and Altman and their companies have zero history of sketchy stuff that would, in a normal world, lead to CEOs being ousted? Affairs are enough to do it to Huang? The whole Chinese spy thing is not confirmed is it? And the three people I've listed have had documented cases of endangering national security through either their actions or their company's actions.

Zuckerberg controls 61% of voting shares for Meta, and implicitly, the Meta board of directors.

The board decides who is CEO. Zuck controls the board. Zuck cannot be fired unless he wants to be.


"Okay, when I introduce you, I'm gonna say, 'This is a friend of mine.' That means you're a connected guy. Now if I were to say, this is a friend of ours—that would mean you're a made guy. Capiche?"

Get away with what? I haven’t heard anything about Musk, Zuck or Altman falling straight into a honeypot from an unallied nation

Musk and Russia?

Why does he support Putin?


I got this book today, as I had it on pre-order. It's a bit disappointing because, on a quick count, there are only five pages with pictures. And I think I saw just a few equations. It's a bit strange to leave out two essential mechanisms for how mathematicians work and think in a book named Six Math Essentials.

Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: