HNHacker News
TopNewBestAskShowJobs

vngzs

2,583 karma · joined October 2, 2018

https://github.com/vngzs https://keybase.io/vngzs

[ my public key: https://keybase.io/vngzs; my proof: https://keybase.io/vngzs/sigs/JW88RqiRCJqnuZ_59waFsjWNKkHQ8c9ejTJR4JXX_mE ]

Comments are my opinion and not intended to represent the viewpoint of my employer (past or present).

submissionscomments
vngzs··on CalyxOS – De-Googled Android Alternative
You have an army of volunteers backporting patches, in the case of Debian. It's been done, but it takes a certain amount of support.
vngzs··on CalyxOS – De-Googled Android Alternative
GrapheneOS, lacking MicroG in the default install, is therefore more private than CalyxOS. Keeping Google out of the loop entirely is necessary for true privacy.
vngzs··on CalyxOS – De-Googled Android Alternative
I know people who made it to the final rounds of interviews at Calyx. They are the real deal. I don't think much of anything could get them to compromise their values about privacy [0].

Might they miss something because they're a smaller team? Yeah, maybe. Will they sell out? I don't think so.

[0]: https://en.wikipedia.org/wiki/American_Civil_Liberties_Union...

vngzs··on An Open Letter Against Apple's Privacy-Invasive Content Scanning Technology
Sad the high DPI displays didn't go well. My eyes can't take the 1080 vertical pixel displays that are still so common on open laptops nowadays. But I really want to like the Librems; there aren't many trustworthy laptops with kill switches out there.

I have an X1 Carbon Gen 9 with a high DPI 16:10 display, 32GB RAM, and anywhere from 4 to 12 hours of battery depending on workload. It's worth a look for people who can tolerate Lenovo's history (BIOS rootkits targeting Windows in the non-ThinkPad lines).

vngzs··on The Framework Laptop is now shipping
I love 3:2 displays. They're wider than 4:3 but maintain the vertical height necessary to see a bunch of code onscreen. If I only get one display (for portable use), I'll pick 3:2 every day.
vngzs··on The Framework Laptop is now shipping
Ryzen tends to get better battery too. I read an article complaining about high temperatures and bad battery life, both of which tend to be significantly better on AMD hardware. Honestly, it's the only thing keeping me from ordering one today.
vngzs··on Valve Steam Deck
This seems to integrate tech from Steam Controllers. I.e., dual touchpads on a gamepad style layout. They went through many prototypes [0] attempting to create something as portable as a console controller but as accurate as a mouse. I certainly prefer it to a touchscreen.

[0]: https://i.imgur.com/3FskweR.jpg

vngzs··on Mozilla rolls out DoH to Canadian users of Firefox
Not with a TCP proxy. To establish a HTTPS tunnel, the client needs to know the identity of the remote end for certificate validation. Randomizing remote ends with a naive proxy would lead to TLS errors.

So you'd likely configure your browser to do DoH destined for that proxy, and have the proxy generate DoH requests to the remote ends. The downside is you'd have to somewhat manually set up certificate trust between your client and your proxy - either by trusting the proxy's certificate, or installing a root CA that issues the proxy's certificate in your browser's trust store.

vngzs··on We need a Butlerian Jihad against AI
I do believe it's reasonable to draw a distinction between what we're doing now (which is essentially just "statistics") and what Musk warns about.

It's a mistake to believe strong AI will just be a more powerful iteration of today's weak AI. He is arguing to slow developments toward general intelligence, not developments in any narrow field.

vngzs··on HIV vaccine trial starts at Oxford
Not just "new," but also interesting. An approach must be both new and interesting to be novel.
vngzs··on Open 3D Engine
Why do you say CryEngine is not very good? I don't have any experience with the dev side, but I used to follow engine tech and CryEngine was (historically, at least) ahead of the competition in rendering large, luscious outdoor environments. These are generally regarded as difficult in software compared to the more static, geometric indoor environments common in earlier technology.
vngzs··on Commit often, perfect later, publish once: Git best practices (2013)
Ever spend significant time spelunking in the Linux kernel commit history [0]? They have a guide [1] on submitting good patches, and lots of care is taken to commit message quality.

The commits are quite helpful, not just for diagnosing errors but explaining philosophy and reasoning behind code changes [2]. I'm tremendously thankful for the amount of effort those developers put into this, because it helps newcomers come up to speed and follow development. In that kind of codebase, it would be virtually unapproachable otherwise.

[0]: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/lin...

[1]: https://www.kernel.org/doc/html/v4.10/process/submitting-pat...

[2]: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/lin...

vngzs··on Group Video Calls
Many of these claims are described in [0]:

> WhatsApp still won't be able to access any of your communications or share them with Facebook. Meanwhile, WhatsApp will be able to share user account information like your phone number, logs of how long and how often you use WhatsApp, device identifiers, IP addresses, and other details about your device with Facebook. Plus, WhatsApp can share transaction and payment data, cookies, and location information with Facebook if you grant permission. All of which has been true since 2016.

More is detailed in the WhatsApp privacy policy [1] (emphasis mine):

> We collect information about your activity on our Services, like service-related, diagnostic, and performance information. This includes information about your activity (including how you use our Services, your Services settings, how you interact with others using our Services (including when you interact with a business), and the time, frequency, and duration of your activities and interactions), log files, and diagnostic, crash, website, and performance logs and reports. This also includes information about when you registered to use our Services; the features you use like our messaging, calling, Status, groups (including group name, group picture, group description), payments or business features; profile photo, "about" information; whether you are online, when you last used our Services (your "last seen"); and when you last updated your "about" information.

I suspect nokya was exaggerating somewhat, but only somewhat.

[0]: https://www.wired.com/story/whatsapp-privacy-policy-facebook...

[1]: https://www.whatsapp.com/legal/privacy-policy

vngzs··on Developing Games on Linux: An Interview with Little Red Dog Games
It wouldn't be unreasonable to officially target Proton support. Hell, I'd like that almost as much as native support. And we can give back to a community that is bringing a massive catalog of Windows software to Linux by sending our patches to the compatibility layer instead of just our proprietary source tree.
vngzs··on Signal not in compliance with new rules, say officials
The news sources I've encountered seem to imply otherwise. The linked Washington Post article was an early hint at this, and built on original material from [0]. Now we have other sources backing this up, such as [1], and detailed breakdowns of this practice with lots of evidence as seen in [2]. Just search for the word "local" or "state" in [2] if you need the details.

[0]: https://www.reuters.com/article/us-dea-sod/exclusive-u-s-dir...

[1]: https://theintercept.com/2018/01/09/dark-side-fbi-dea-illega...

[2]: https://www.hrw.org/report/2018/01/09/dark-side/secret-origi...

vngzs··on Hacker deleted all of NewsBlur’s Mongo data and is now holding the data hostage
I just want to say that I think this is absolutely a fair approach for many systems, but it sounds so radical against the backdrop of perimeter security that you're catching undue flak. You need to take a lot of care to build and design the "one layer to rule them all" in a way that's ridiculously sound, but this can actually produce a strong design if you pull it off. I didn't give this advice because I looked at where NewsBlur is today and figured they had some learning curves to get over before they could make the right tradeoffs to do it safely. If they got that much wrong with their DB, they might not know how to design a sane zero-trust network, or even how to make rational decisions in that world.

You see this "authN/authZ above all else" line of thinking in Google's security design [0], with their ubiquitous login wall. For their employees, that login wall has extra hardening - you can't do regular password resets, and you need to possess a physical security key (which acts basically as a scaled-down single-purpose HSM) and pass a suite of posture checks on the device (proportional to the sensitivity of the protected resource) to pass through it.

Then they put this login wall in front of everything, even internal services, and that tends to be OK because the system "fails closed," and the only way to access their protected resources is via physical safe rooms deep in the Google offices in which elevated privileges may be obtained.

Putting all your systems on the Internet forces you to get the incentive structure right, and that can be useful in companies where "private networks" can serve as justification to weaken security - if your login page is Internet-exposed, then you simply have no choice but to make it strong enough to withstand the chaos of the Internet. There's significant merit in that.

[0]: https://sre.google/books/building-secure-reliable-systems/

vngzs··on Signal not in compliance with new rules, say officials
> Are you saying that that the probable cause requirement still allows warrants to be granted very easily? Or that the probably cause requirement is not being followed?

Well, both.

Warrants may be granted very easily via the FISA courts [0]. They are unenforced in the warrantless wiretap programs [1]. If your wiretap happened to be illegal, and you want to work around this to win the court case anyway, you use parallel construction to build a case that does not rely on the secret wiretap [2]:

> Because the SOD's work is classified, DEA cases that began as NSA leads can't be seen to have originated from a NSA source.

> So what does the DEA do? It makes up the story of how the agency really came to the case in a process known as "parallel construction."

[0]: https://en.wikipedia.org/wiki/United_States_Foreign_Intellig...

[1]: https://en.wikipedia.org/wiki/NSA_warrantless_surveillance_(...

[2]: https://www.washingtonpost.com/news/the-switch/wp/2013/08/05...

vngzs··on Signal not in compliance with new rules, say officials
Signal has a means of delivering push notifications using their own infrastructure via websockets. It has a higher battery drain than using Google's push infrastructure, but it works.
vngzs··on Hacker deleted all of NewsBlur’s Mongo data and is now holding the data hostage
I think you misunderstood me - I mean that you want multiple layers to your security, and you need to be careful to select layers that fail independently. If Vault has a sufficiently bad unauthenticated bug, then attackers can simply use it to request database credentials and query the DB, which is now on the open internet.

An easy way to get independent failures is to layer a private network with strong auth and firewall rules. While it's certainly possible to expose your DB to the internet safely - given sufficient protections - you won't get that with just a TLS key. And even if you try to implement the "obvious" additional layers here (Vault, right?), it's easy to inadvertently include design problems that reduce to "only one failure and the system is exposed."

vngzs··on Hacker deleted all of NewsBlur’s Mongo data and is now holding the data hostage
Yeah, there's the whole "zero trust" possibility which I didn't mention. If you do authentication/authorization really well, you can stop doing (2). For things like databases, I think it's better to treat them as if they _could_ be exposed to the open internet, without actually doing so. It's generally not the case that anyone on the internet needs to query your DBs.

As described, you're _only_ relying on client TLS to protect your database. What if the TLS key leaks? So you're down to one layer again - one mistake and it's game over. So maybe you need Hashicorp Vault so you can have client certificates with very short validity periods. So do you expose Vault to the Internet, so you can fetch your client certificate to query the DB? What if Vault has a bug? And it's turtles all the way down ...

I love zero trust designs. But I think saying you can just slap client TLS on the problem and be done is laying the foundation for a repeat of this event - however you look at it, you want redundancy in your security.

vngzs··on Hacker deleted all of NewsBlur’s Mongo data and is now holding the data hostage
I think there are some good lessons here:

1. Even if you have one way to protect your database (e.g., firewall rules), you should have another. In this case, use a database password or (better) client TLS certificate to authenticate traffic. We're all human and we mess up. You should be designing systems that are graceful in response to your inevitable mistakes.

2. If you can afford another server/a hosting provider with VPC-like features, don't put your databases on the open internet. Run them in a private network (RFC 1918), behind a NAT and a load balancer/entrypoint that only routes to your internet-serving applications. Allow only those application servers to hit your databases. If you had done this, the attacker wouldn't have noticed your mistake, because all they could hit was your public server.

3. Keep regular backups. Oh, and test them! If you don't test your backups, you don't have backups - you have archives. In the GitLab data loss incident [0], they had 3 methods of data backup, but all of them failed. A regular test would have discovered this. Don't make that mistake.

Good on you for sharing the events as they happen. I think people tend to be much more forgiving in response to openness. Don't freak out, and write a public postmortem when you're done.

[0]: https://about.gitlab.com/blog/2017/02/01/gitlab-dot-com-data...

vngzs··on Learn C and build your own Lisp (2014)
I'm fairly conflicted about this. While I believe that learning C teaches bad habits - manual memory management tends to account for some absurd percentage (70-80%) of CVEs in modern programs - it's also unquestionably tethered to the history of software, and a majority of the immense foundation of tooling on a Linux system is authored in C.

There is a terrific benefit to learning on a platform where you can not only look up documentation within the system (man pages), but also "dig deep" into the code of any particular piece of the system you are interested in. So in learning C, you gain access to the body of knowledge contained in a Linux system, or contained in a BSD, and when you get curious about something like "These loops and conditionals are all fine and well, but how does a window actually get drawn to my screen?" you have everything at your fingertips to go and answer it.

Is that connection to the machine possible without learning C? You could use a toy operating system written in something else, but you miss out on the honesty that comes from teaching the same system that people are actually using (or at least able to use).

I think Go might not be an unreasonable alternative. It shares C's philosophical heritage, removes the need for manual memory management, and is capable of exposing a student to pointers, creating images, and HTTP communication all within the stdlib and the first chapter of the official book. The only downside, as I see it, is that one would need to additionally learn something like Rust to be able to apply the concepts to things like kernel drivers and close-to-the-metal programming.

vngzs··on Apple's iCloud+ “VPN”
Thanks for the "how" - I guess if you fully control the client and server, there's some extra checks you could implement client-side based on the cryptographic root of trust.

FWIW, I wasn't really trying to make a prescriptive statement about how the world ought to be, I was more trying to describe what (I think) the perspective of these corporations has been on the matter.

In the past, I've been an encryption advocate with the knowledge that we (tech) must sacrifice some ability to appease politicians in implementing it. What you're describing sounds like an innovative way to preserve privacy and provide security for at-risk people, which is a perspective I haven't heard before.

vngzs··on Apple's iCloud+ “VPN”
In the bloom filter example, what device calculates the hash inputs for the bloom filters? If it's the server, then the server needs a copy of the image to check. So is it the client? If so, how can you prevent a malicious client from forging their hashes to be those of known-safe images?

Not saying it's not possible to build an E2E image storage service that also has the protections society tends to demand. Just saying that I haven't seen anyone do it yet, because these problems are subtle.

vngzs··on Apple's iCloud+ “VPN”
From Apple's statement[0]:

> The first assigns the user an anonymous IP address that maps to their region but not their actual location. The second decrypts the web address they want to visit and forwards them to their destination. This separation of information protects the user’s privacy because no single entity can identify both who a user is and which sites they visit.

Apple is not saying nobody can deanonymize you - they are being very careful to only state that no single entity can deanonymize you. Hence you should still assume this is not a good protection against any entity with subpoena power, or the ability to compel the cooperation of Apple and their 3rd-party egress relay providers.

[0]: https://9to5mac.com/2021/06/07/apple-icloud-private-relay-fe...

vngzs··on U.S. to give ransomware hacks similar priority as terrorism, official says
The government already has baseline security standards that are created and published by the government, not private entities. NIST 800-53[0] is a good example of this, and it generally applies to critical infrastructure providers:

Regulations should not adopt a private company's baseline security standard; we should lean on the work NIST has already done and standards that already apply to (mostly defense) critical infrastructure.

[0]: https://nvlpubs.nist.gov/nistpubs/SpecialPublications/NIST.S...

vngzs··on Show HN: Flashcards to learn AWS skills
I don't want to hear people tell me only how to solve problems with AWS's legos. Sure, we use them - and sure, knowing them will help you, but those tools are a crutch that is going to prevent you from solving tough problems if it's the only thing you know. You don't want to be stuck on a problem just because it's not in Amazon's toolbox yet.
vngzs··on A new replication crisis: Research that is less likely to be true is cited more
This sounds inseparable from the replication crisis. The incentives are clearly broken: they are not structured in a manner that achieves the goal of research, which is to expand the scope and quality of human knowledge. To solve the crisis, we must change the incentives.

Does anyone have ideas on how that may be achieved - what a correct incentive structure for research might look like?

vngzs··on Ampere switching to in-house CPU designs
You can get ARM desktops now at non-crazy prices. The neat thing about them is they are 32/48 core CPUs at consumer-available prices[0]. Sure, they're on the high end, but it's around the cost of a workstation that one might use for, e.g., kernel development.

I'm not affiliated with Avantek in any way. I just did a bunch of research looking to build/buy a desktop, and ARM was a contender, but I ended up building a beefy x86 machine with a Ryzen instead.

[0]: https://store.avantek.co.uk/arm-desktops.html

vngzs··on Improving Firefox stability on Linux
Yeah, sorry, the openjdk situation sucks. It's an "extra" package, which means the maintainer is a community member rather than being part of the Arch core group.

The nice thing about Arch packages, though, is that they're basically just bash scripts. And if all you need is a simple version bump, it's usually quite easy: download the package tarball from [0] and change [1] to the version you want, then do a `makepkg -s` in the directory of the package. It will build the package in a (usually reproducible) chroot. If it builds without errors, then you'll end up with a tarball that you can `pacman -U ${pkg}.tar.zst` to install the produced files.

If you need help, makepkg documentation on the wiki[2] is pretty great. And don't forget to send a patch to arch-dev-public[3] and CC the maintainer. At the very least, it'll kick off a discussion that will get the package updated.

Rolling your own packages is easy in contrast to, say, Red Hat - where compiling an RPM is easy if you've done it a bunch, but really difficult to get bootstrapped on.

[0]: https://archlinux.org/packages/extra/x86_64/jdk-openjdk/

[1]: https://github.com/archlinux/svntogit-packages/blob/packages...

[2]: https://wiki.archlinux.org/title/Makepkg

[3]: https://lists.archlinux.org/listinfo/arch-dev-public

← PreviousPage 6 of 10Next →