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 Blacksmith – Rowhammer bit flips on all DRAM devices today despite mitigations
You get dedicated time on a core while your task is running. For instance t3.medium is 2 vCPUs (because hyperthreading) but as you can see in [0] it's only one physical core.

[0]: https://aws.amazon.com/ec2/physicalcores/

vngzs··on Show HN: You.com, private search engine that summarizes the web – built for devs
I installed the extension, tested a search, then switched my default back to another search engine. Future searches on you.com failed because I did not have the extension installed - even though I did.

I uninstalled the extension. I don't want to use tools that give me less power over my browsing experience, arbitrarily breaking although I'd done what was explicitly documented. I want tools that give me more power.

vngzs··on HashiCorp – S1
A lot of big finance companies use Nomad for all their compute scheduling. Citadel, for instance. They desire the ability to schedule Windows workloads, containers, regular processes, etc. through a common interface. They might not want or need to go all-in on containers.

Vault has a similar target market. Big high-paying institutions. It's not the average market of your tech company, and 100-200 person startups generally won't need it. If you're in the fintech space, maybe you do.

vngzs··on The wealthiest 10% of Americans own a record 89% of all U.S. stocks
That CNBC article is unbelievably misleading. If you dig into the sources, it links to the actual report [0] where they state that 80% of households paid federal income or payroll taxes:

> Remember, this reflects only those who pay no federal individual income tax. TPC estimates that while the number of households paying neither payroll nor income taxes also rose significantly, roughly 4 out of every 5 did pay one of these two taxes.

The trick is that "income tax" is used as a technical term in the headline, while the colloquial use of it is quite different (average readers would probably consider their payroll tax to be "income tax").

[0]: https://www.taxpolicycenter.org/taxvox/covid-19-pandemic-dro...

vngzs··on Cloudflare Uses HashiCorp Nomad (2020)
Yes, you can be up and running with a multi-machine lab install in a couple hours, even if you're rolling the component installations and configuration yourself. In K8s production cloud workloads, this is usually offloaded to GKE/EKS/Kops, but the equivalent Kubernetes manual installation is appreciably more heavy-handed.
vngzs··on Cloudflare Uses HashiCorp Nomad (2020)
Nomad is very lightweight. It also does nothing near as much as what Kubernetes does. But if all you need is to schedule containers (or other binaries) it does so quite nicely.

You generally want to pair it with Consul (as Cloudflare has done) to get service discovery. Consul is also quite lightweight.

vngzs··on Governor vows criminal prosecution of reporter who found flaw in state website
The EFF will likely take this case. Have the reporters contacted them?
vngzs··on Practice Problems for Hardware Engineers
Agreed. The quality is all over the place. "MISR the BIST" sounds like a class in-joke that made its way into the document. But it makes the questions just that much more useless for non-students of this professor.
vngzs··on Practice Problems for Hardware Engineers
I guess the term is slightly overloaded. At least from my academic background (day job is in more traditional software), AC tends toward the more "electrical engineer" side of the specialization, while IMO these exercises are more "computer engineer" in nature. Not that one won't learn about AC in the inevitable EE classes, but digital electronics tend to be the focus.
vngzs··on Practice Problems for Hardware Engineers
As someone who studied hardware engineering in college, the practice problems cover the core competencies in a degree on electrical/computer engineering. Reading through the practice problems I saw very few subjects that weren't at least touched on in undergrad. The fact that they come with solutions makes this a handy manual for students.

Considering the practice problems were hosted on Arxiv, this was exactly what I expected. What did you expect?

vngzs··on Coinbase Breach Notification
A bit late, but their new flow requires ID verification for SMS 2FA resets [0].

[0]: https://help.coinbase.com/en/pro/managing-my-account/account...

vngzs··on The Xterm Window Manager Homepage
It's easier and more lightweight to configure an SSH session than X11 forwarding or remote desktops. So if you do all your real work on a server or desktop that you access remotely, these kind of tools can be helpful.

That said, all I really need is tmux to do my work. I use a tiling window manager as my daily driver. Don't see any reason I'd adopt a floating WM in the terminal except that it looks cool.

vngzs··on Alexa, Ring, and Astro: Where's My Privacy, Amazon?
The entire point of these devices is to provide a window for the corporation into your home, so that Amazon can better sell you things. They're not doing you a favor by making the product; the product is an excuse to uniquely position them to observe your behavior in a domain previously unavailable (meatspace).

Expecting "privacy" from such a device is absurd. You waived it when you bought the device.

vngzs··on Coinbase Breach Notification
With only the pdf to go on, I address the "flaw" in more detail in these comment threads [0] [1]. In short, I believe the "flaw" is likely to be "we used SMS for identity verification, without additional necessary scrutiny."

The technical barrier to entry for accruing and using breach databases is near-zero [2], same with the barrier to SMS fraud. Both are routine and easy methods for criminal groups with no special technical abilities, and therefore they are likely. Since the onus is on Coinbase to do identity verification in account recovery, a large number of successful takeovers would be a "flaw" in their process, even if it's not a technical flaw (which I would expect to be expressed in language like "vulnerability").

Accepting untrusted, unauthenticated user input as a SMS verification number would be a serious login-related flaw, and certainly Coinbase pentests their login pages. Any competent pentester would discover such a flaw. So between "Coinbase shipped a critical and obvious login flaw to prod" and "a routine and common criminal tactic was employed successfully against them," I find the latter more likely.

[0]: https://news.ycombinator.com/item?id=28720101

[1]: https://news.ycombinator.com/item?id=28720520

[2]: https://xkcd.com/2176/

vngzs··on Coinbase Breach Notification
I haven't been able to verify these sort of claims any more than I've been able to speculate it was blanket telco Letters-of-Authorization (LoAs) [0][1] or classic SIM swaps that resulted in the account takeovers. I'm not claiming you're wrong, but given the timing of the LoA fraud and the attacks, it seemed likely to me that this was not an actual web vulnerability.

What makes you believe a specific exploit like that existed against Coinbase's 2FA? And if it existed, then why wasn't that caught in a routine pentest?

[0]: https://krebsonsecurity.com/2021/03/can-we-stop-pretending-s...

[1]: https://lucky225.medium.com/its-time-to-stop-using-sms-for-a...

vngzs··on Coinbase Breach Notification
I agree. From Coinbase's perspective, they ought to defend their infrastructure against fraud, whether that is a direct attack on the users, an attack on the users' telcos, or insider activity directly.

From the telco's perspective, they have a responsibility to stop SMS and SIM fraud, and our regulations have failed to properly hold them accountable in this domain.

I would add that the users have some responsibility for losing their emails/passwords, but my initial framing insufficiently demands responsibility for the service providers in this instance. The service providers should be expected to take all reasonable steps to prevent fraud on their platforms, and that should include extra scrutiny of SMS-based authentication mechanisms (e.g., identity verification). This is why Coinbase paid them back, accepting some responsibility for the fraud.

vngzs··on Coinbase Breach Notification
There is some speculation in another comment that their SMS verification server may have actually had a technical flaw, and the issue was not a lack of separate identity verification on SMS [0].

However, around the time of the breach date (March - May 2021), there were a number of "B2B" services that offered a "type in any SMS number and you will get all text messages to that number," type feature intended for customer support teams to use for shared SMS access. Those systems often had privileged access to telcos and were regularly exploited by attackers to break 2FA without even a SIM swap [1]. With those tools, stealing all SMS to a number required only intent, not conversations with telco support personnel.

[0]: https://news.ycombinator.com/item?id=28720280

[1]: https://krebsonsecurity.com/2021/03/can-we-stop-pretending-s...

vngzs··on Coinbase Breach Notification
I bet that control of email address + SMS 2FA was sufficient, alone, to recover the Coinbase account password. Lots of systems permit this kind of recovery, and while I may tell a technical crowd "if you use SMS for 2FA, that's on you" less technical users may not have the requisite background to understand the security tradeoff they make in doing so.

The "flaw," in my reading of it, was to support SMS-based account recovery at all. But I'm not necessarily right here, and open to alternatives.

vngzs··on Coinbase Breach Notification
Anyone care to speculate what the flaw in their SMS recovery flow actually was? It's hard for me to think there's even a safe way to implement SMS based account recovery. They would be smarter to just turn it off.
vngzs··on Coinbase Breach Notification
In the linked PDF, Coinbase does not claim to have knowledge of a vulnerability in their system (edit: though it does note "the third party took advantage of a flaw in Coinbase’s SMS Account Recovery process," I interpreted that as "we supported SMS account recovery at all" which is inherently broken [0]). The requisite two-factor bypass is detailed in the linked pdf:

> Even with the information described above, additional authentication is required in order to access your Coinbase account. However, in this incident, for customers who use SMS texts for two-factor authentication, the third party took advantage of a flaw in Coinbase’s SMS Account Recovery process in order to receive an SMS two-factor authentication token and gain access to your account.

My guess is, because funds were stolen from users' accounts, the CA breach notification laws apply and this needed to be disclosed as such. However, that doesn't necessarily mean that Coinbase was technically "breached," only that customer accounts were compromised.

If the attacker controls your personal email associated with Coinbase, accompanying passwords, and phone number, and you use SMS 2FA, then your funds were stolen. Otherwise, they were safe. That's my reading of the article.

[0]: https://krebsonsecurity.com/2019/08/who-owns-your-wireless-s...

vngzs··on Coinbase Breach Notification
Coinbase made everyone whole, and the attackers stole the credentials (not because of Coinbase's fault) ahead of time, and the attackers had to perform a "SIM swap" type attack on the users. "Breach" may be the required term for the Californian government, but this wouldn't qualify to most people as a traditional breach (i.e., compromise of Coinbase's infrastructure).

Edit: California, not Canada. My bad.

vngzs··on ValidScript – a modest proposal for app security
Unclear to me whether "better languages" is the issue or just better type systems. We should be able to express "untrusted input" in any sufficiently extensible type system (e.g., Rust would probably let you express this). Typescript may let you express this; I don't know because I've never worked with it.

At a high level, I agree - our tools should do this for us, we shouldn't need people to be experts in untrusted input handling to work effectively with it.

Google internalized this and built their own modular library that they use for all UI development that just makes this completely opaque to the developers. So they don't have most frontend input handling bugs, or if they do they get fixed for everyone at once in the framework layer.

vngzs··on Podman in Linux
I am generally conservative about these things, but recently the consensus appears to be that unprivileged user namespaces are stable enough for GA. If you work in a targeted environment (e.g., you are a reporter covering civil liberties), maybe you should wait a bit longer, but for average people the advantages of rootless containers probably exceed the advantages of keeping the unprivileged_userns_clone code out of your attack surface.

I would encourage most security-conscious users to enable it and migrate to recent podman over using Docker, assuming a sufficiently recent kernel. The latest batch of major Linux OS releases have all enabled kernel.unprivileged_userns_clone, so Red Hat, Canonical et al seem to agree.

For those interested, though, you can read the anatomy of a userns clone() vulnerability here:

https://lwn.net/Articles/543273/

vngzs··on Court issues permanent injunction in Epic vs. Apple case
Just tried this. I had to talk to a support staff on the website, and they made one attempt to offer me a lower rate for a year (which I declined).

In my opinion, it should be as easy to unsubscribe as it is to subscribe. I interpret anything else as consumer-hostile. It's very strange coming from NY Times ... I know they find it hard to finance journalism nowadays, but dark patterns are never the answer.

vngzs··on First Edition Unix Manual 'Miscellaneous' Section (1971) [pdf]
See also: historical Unix system manuals, in web format [0].

[0]: http://man.cat-v.org/unix-1st/

vngzs··on Surveilance bill rushed through Australian parliament in 24 hours
Stop allowing email providers to control the domain of your email. Buy a $10/year domain and point it at GitHub pages (so it gains some trustworthiness). Send your email to that domain, and point the MX records at a mail provider (Fastmail or GMail or ProtonMail, doesn't matter).
vngzs··on Bad News
I'm pretty sure this is the joke in multiple John Oliver web segments [0]. It's pretty well-understood that the YouTube design is not conducive to long-form discussion.

[0]: https://www.youtube.com/watch?v=knbw0gJHHBk

vngzs··on T-Mobile Confirms It Was Hacked
The price goes down as a target is in more breaches. If your SSN is previously leaked, it's worth less.
vngzs··on T-Mobile Confirms It Was Hacked
From Flashpoint's analysis [0], in 2019 a SSN is $5.

[0]: https://go.flashpoint-intel.com/docs/pricing-analysis-of-goo...

vngzs··on Federal Court Orders BitMEX to Pay $100M over Illegal Crypto Trading
Yeah, but trading volume is not liquid cash. It's not the exchange's money. They make money on fees / commissions, but they can't just dip into the customer funds and use them to pay the government.
← PreviousPage 5 of 10Next →