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 1Password for SSH and Git (Beta)
I'll concede that it's not nearly as crazy to share keys across devices for home use or low-risk things. I must admit I was speaking mostly from an enterprise perspective.
vngzs··on 1Password for SSH and Git (Beta)
I think this is a bad idea for users. I don't think SSH keys are things you should share across machines in a password manager. If you have two devices, then you should have two keys (though this is the subject of some debate; see [0]). Using the 1Password SSH agent encourages people to have "one" SSH key across devices, which means that any leaks will disproportionately impact them.

It's unfortunate, because there is some real innovation around the per-application usage permissions:

> 1Password will ask for your consent before an SSH client can use your SSH key. Because of this, there's no concept of adding or removing keys like with the OpenSSH agent.

If an organization wishes to solve the SSH pubkey distribution problem (the main reason one would copy a private key across machines), then they should use SSH certificate authorities like [1]. In fact, I think that would be a far more interesting 1Password product—HashiCorp Vault could use some competition for this kind of use-case.

[0]: https://security.stackexchange.com/a/40061

[1]: https://www.vaultproject.io/docs/secrets/ssh/signed-ssh-cert...

vngzs··on Heart-disease risk soars after Covid, even with a mild case
The CDC reports 16.3 flu/pneumonia deaths per 100,000 [0]. That's 0.0163 percent.

Johns Hopkins University reports covid-19 has 278.89 deaths per 100,000 - a 1.2% mortality rate in the US [1].

That would put covid-19 at 278.89/16.3 = ~17 times more deadly than the flu. Just because the numbers don't hold for a particular age group doesn't make the aggregate statement false. In fact, it would be misleading to categorize "17 times more deadly" as anything other than "more deadly."

Edit: I made a mistake dividing the percentages the first go around, reporting covid-19 as 73 times more deadly than influenza. I have updated the 3rd paragraph to show the calculations directly.

[0]: https://www.cdc.gov/nchs/fastats/flu.htm

[1]: https://coronavirus.jhu.edu/data/mortality

vngzs··on Twitter misses ad revenue and user growth estimates
It's addressed in the article, though absent specific numbers:

> The company said the impact from privacy changes by Apple Inc (AAPL.O) remained modest. Last year, Apple began requiring apps to receive permission from iOS users to track their activity on apps and websites owned by other companies.

> The Apple changes could impact Twitter in the future as it grows its performance advertising business, Segal said, referring to ads that seek to drive sales or other consumer actions. He said Twitter is working to mitigate future negative impacts from Apple's changes.

vngzs··on Exposing a web service with Cloudflare Tunnel
Quick word of warning: I found it striking that even Cloudflare's Teams product, which supports Tunnels as a feature, does not make Tunnels private (e.g., by enforcing authentication, or restricting who can reach an exposed tunnel to your organization) by default. Anyone on the Internet with the Cloudflare Warp client can reach a Tunnel configured with default settings, a quirk that is not called out in their official documentation.
vngzs··on Write it in Garamond
Computer Modern looks terrible on computers (ironically!). It's only tolerable on relatively high-DPI printed documents which are capable of rendering the intricate, narrow parts of the legends. I wish people would stop using it for digitally distributed PDFs.
vngzs··on Tesla to recall vehicles that may disobey stop signs
At the time of comment, a LiDAR rig would cost around $10,000. A few years before that, they were more like $100,000. Presumably the cameras are much cheaper.

I would be willing to bet that production efficiencies will be found that will eventually drive that cost down significantly.

vngzs··on Tesla to recall vehicles that may disobey stop signs
Tesla didn't use LIDAR because it is more expensive [0]. Quoting Musk:

> Anyone relying on LIDAR is doomed. Doomed. Expensive sensors that are unnecessary. It’s like having a whole bunch of expensive appendices... you’ll see.

[0]: https://www.theverge.com/2019/4/24/18512580/elon-musk-tesla-...

vngzs··on Socrates on the blessing of being refuted
While I agree that the content of one's apology is largely irrelevant here (actions speak louder than words), the video makes two substantive claims about changes when controversial opinions about about covid-19 are presented:

1. A disclaimer at the beginning of the video will instruct people to consult their physician and provide a warning that the views expressed are "against the current consensus of experts."

2. When a controversial expert is interviewed, the following videos will feature experts who hold the mainstream viewpoint, serving to counterbalance the claims of a small minority.

Time will tell if this happens, but I find the above points far more meaningful than any apology.

vngzs··on Decryption through LUKS2 reencryption crash recovery
According to [0] attacks from a cold device require two steps of physical access. That would bypass signed contents of a boot partition, since you're messing with unprotected LUKS header contents.

If the contents of the boot partition are unsigned there are probably easier attacks available. You could, for instance, deploy a kernel with crafted reencryption logic baked in over the default boot kernel. That wouldn't require this vulnerability, because the victim will be typing the passphrase into the attacker's (unprotected) kernel from /boot.

[0]: https://bugzilla.redhat.com/show_bug.cgi?id=2032401

vngzs··on A Tesla on autopilot killed two people. Is the driver guilty of manslaughter?
In my experience driving Teslas on autopilot (a few dozen hours), you need to keep watching the road. The whole time. Especially on on/off ramps and around traffic lights. I almost watched a Tesla I was operating drive into a meridian on a highway off-ramp because the concrete was approximately the same color as the road. I slowed down to see how long it would take to detect the coming collision ... and if I hadn't been watching carefully, it would have been too late.

Yes, drivers must still pay attention in self-driving cars. Probably a future exists where that is no longer necessary, but that's not today.

vngzs··on Mortgages are a manufactured product
Nit: "personal property" traditionally means any property except for real estate. Using it to mean "real estate" is confusing.
vngzs··on Ask HN: Should I publish my research code?
Please don't use this license. Copy the language from the preamble and put it in your README if you'd like, but the permissions granted are so severely restricted as to make this license essentially useless for anything besides "validation of scientific claims." It's not an open-source license - if someone wished to include code from CRAPL in their MIT-licensed program, the license does not grant them the permission to do so. Nor does it grant permission for "general use" - the software might not be legally "suitable for any purpose," but it should be the user's choice to decide if they want to use the software for something that isn't validation of scientific claims.

I am not a lawyer, just a hardcore open-source advocate in a former life.

vngzs··on SSH Bastion Host Best Practices
From a defense standpoint, one should consider "shell on a box" to usually mean attackers can get root on a box. If they can get persistence, they can wait for a kernel CVE to abuse.

Now, if you're just using a bastion as a jump host, you don't need to offer shells on it. Just allow people to proxy a port to behind the bastion and be done with it.

    PermitTTY no
    ForceCommand /usr/sbin/nologin
    AllowTcpForwarding yes
    AllowAgentForwarding no
vngzs··on Switching to an N95 mask gives a 75x boost in Covid protection
I'm not aware of airborne particle size distribution data existing specifically for covid-19 (please share it if you have it), but the majority of particles that might transmit an airborne virus are usually 0.3 microns or greater in size. That is, while some virus particles may linger and those may be smaller than 0.3 microns, the quantity of particles is highly relevant in determining whether someone will get infected - if we can cut a patient's exposure to virus particles by a significant factor, this may indeed prevent infection.

> The count median diameter (CMD) of the cough aerosol particles was fairly similar for all subjects and did not vary significantly during and after illness (p = 0.9340). When the subjects had influenza, the CMDs of all the cough aerosols were between 0.57 and 0.71 μm (average 0.63 μm, SD 0.05), and the geometric standard deviations (GSD) were between 1.54 and 1.83. After the subjects had recovered, the cough aerosol CMDs ranged from 0.57 to 0.89 μm (average 0.63 μm, SD 0.100), and the GSDs were between 1.53 and 2.28. Similar results were seen for the volume median diameters (VMD); the average cough aerosol VMD while ill was 2.44 μm (SD 1.17) and was 2.24 μm (SD 1.50) after recovery. The GSDs ranged from 1.66 to 2.31 while ill and 1.54 to 2.21 after recovery. The difference in VMD while ill and after recovery also was not statistically significant (p = 0.7780).

It is notable that the particle sizes did not differ significantly when the patient was infected with a virus versus the control group. This makes me think that covid-19 is similar.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4676262/

vngzs··on Switching to an N95 mask gives a 75x boost in Covid protection
In case anyone is interested, the fit test procedures are documented here: https://multimedia.3m.com/mws/media/252309O/3m-adflo-qualita...
vngzs··on Switching to an N95 mask gives a 75x boost in Covid protection
Also notable is that proper use of an N95 takes training. Unlike with surgical masks, N95 fit is specific (you normally get them sized when issued at a hospital). In addition, most people have bad habits around mask removal - you should always remove masks by the bands, but people tend to touch the mask surface itself. And with any mask removal, you should always wash your hands (in a particular manner that I'm sure readers are familiar with by now). Finally if, when removing a mask, the user touches the interior of the mask after touching an infected surface then reapplying the mask can lead to infection.

Not doing any of these steps can lead to infections. Simply issuing N95s to a general population might not help, if the population is improperly trained.

Surgical masks "work" by preventing the wearer from spreading diseases to others. In aggregate, this can help a population, but in specific cases it might not protect you. N95s work by protecting the wearer, with only secondary benefits to others (the N95s with "nozzles" on the mask surface provide no protection to other parties, as outflow is unfiltered).

vngzs··on Venmo, PayPal, Cash App must report payments of $600 or more to IRS
People who buy and resell collectibles or clothing for considerable profit with no intention of using the items themselves. The term is frequently associated with concert tickets, where people with no intent of going to a concert will drive the ticket prices up and skim a profit.
vngzs··on Venmo, PayPal, Cash App must report payments of $600 or more to IRS
I was concerned because I couldn't find the specifics on what payment tags would lead to income reporting. Here's a better article that addresses that clarification in specific terms [0]:

> It’s important to note that this change — which will be applied during 2022’s tax year and moving forward — applies only to payments received for sales of goods and services, and not to friends and family payments.

[0]: https://finance.yahoo.com/news/tax-law-sell-more-600-1548205...

vngzs··on Venmo, PayPal, Cash App must report payments of $600 or more to IRS
My roommate sends me $700 via Venmo every month. I funnel it directly to the landlord in a direct deposit on the same day. It's frustrating to realize these transactions are going to be clearly reported as income, because I have never received income from Venmo - all I do with it is split bills. I realize I can probably sort this out with the IRS, but it seems relatively inappropriate that I would have to do that in this situation.
vngzs··on Lawsuit filed alleging Google is paying Apple to stay out of the search business
Some non-default browsers (e.g., Snowhaze, Firefox) support adding new search engines. However, if you highlight and "look up" a word on-page, the search link will always take you to the OS-default search engine.
vngzs··on Great engineering teams focus on milestones instead of projects
This is a good point. Scrum's sprint length inflexibility frequently means over or under-allocating time on projects, which messes up metrics (e.g., ticket completion per sprint) ... but is otherwise irrelevant to the work itself. This leads to organizations making sub-optimal choices that help with scrum, but don't help with work.
vngzs··on Face recognition is being banned, but it’s still everywhere
I don't know how long the records persist, but presumably for things like video surveillance footage the decay period would be quite fast. For full-text contents, less speedy but still fast. I can only truly envision long-term collection and storage of metadata - and even then, it's a big question how much is feasible and reasonable to store indefinitely.

It's likely that for your security camera footage to be accessed you would have to be targeted. It's likely that such targeting would only affect footage in the future or very recent past. But I'd grant that pervasive attackers can probably capture exponentially-decaying single video frames from millions of cameras if they were appropriately motivated, and those single frames could go back quite far.

The real problem here is the existence of such a system creates a kind of panopticon [0], with chilling effects not only on activity and discussion contrarian to the current administration, but also any future administration that may have access to electronic surveillance records. Without knowing how long records are kept, it is quite plausible that a future authoritarian state will misuse past records to target civilians.

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

vngzs··on Alternative privacy-respecting front ends for popular services
Collecting and sharing massive sets of data on people has a societal cost when that data leaks and is used to target individuals with identity theft. Stalkers abuse these datasets to find and attack their targets. Data breaches are especially common in the ad-tech, marketing, and sales industries because they are largely unregulated and exist to collect data on millions of people at scale.

Not all users of anonymity services are people who have a choice in the matter. And if you're not currently a victim of stalking or identity theft, why invite attacks by openly sharing personal information? It's a personal risk calculus, and everyone's needs are different.

- 622 million leaked email addresses, employers, geographic locations, job titles, names, phone numbers, & social media profiles from PDL [0]. According to the PDL website, the purpose of this data was for PDL clients to "supercharge [their] software with over 3 billion profiles. Tap into the resume, contact, social, and demographic information for over 3 billion unique individuals, delivered to you at the scale you need it."

- 125 million leaked email addresses and 9 billion data points on users from Apollo [1]. Dataset includes email addresses, employers, geographic locations, job titles, names, phone numbers, salutations, & social media profiles.

[0]: https://www.troyhunt.com/data-enrichment-people-data-labs-an...

[2] https://www.wired.com/story/apollo-breach-linkedin-salesforc...

vngzs··on Alternative privacy-respecting front ends for popular services
Very similar. Startpage tries to learn as little as possible about its users (which I believe, after seeing their error page). It sends searches to Google on users' behalf and returns results without personalization.

In Whoogle's case, Google can still track the searches you send to it. With Startpage, only Startpage can track them in detail - and Google only in aggregate.

vngzs··on Bitwarden compatible server written in Rust
In Rust, abstractions tend to be zero-cost, but things that actually impact runtime are as fast as you write them.

Concurrency problems don't magically disappear because your language can check for them. You still need to use memory-safe patterns, but the compiler will force your hand and prevent you from doing unsafe things accidentally.

In the end, this makes it far easier to build correct, concurrent applications. If you decide to try some fancy new pattern that doesn't require locking, the compiler can guarantee the memory safety of your new implementation.

vngzs··on Clever uses of pass, the Unix password manager
Besides that, pass leaks metadata.

If your passwords are the only part of your password manager you think are secret, then you're probably not considering the utility of an enumeration of all your network handles. Mine aren't very interesting, but I wouldn't post that on GitHub, even in a private repo.

I love pass and use it. But when I need to carry my password database around, I don't rely on pass alone - I throw the pass database in a Veracrypt container. That way if my flash drive is stolen, it's just an encrypted blob.

vngzs··on Crypto wallet security as seen by security engineers
It's not obviously explicit (to me, at least) whether they are discussing rotation of passwords specifically or wallet-related authentication flows more generally.

In Bitcoin wallets, for instance, multi-signature wallets do not allow rotation of the signing keys. If one of your singing keys gets owned, you need to burn the wallet.

vngzs··on Ask HN: I'm looking for a good book on the fundamentals of CS
Surprised no one seems to have mentioned the godfather of CS books, The C Programming Language by Brian Kernighan and Dennis Ritchie [0]. At first glance, it's a programming language book rather than a "computer science" book, but you probably won't think that by the time you're finished ...

First, every operating system you'll use - likely in your lifetime - is a giant mountain of C code. From the kernel all the way to the userspace utilities. As a programmer, your career will be forged on the shoulders of giants, and those giants wrote in C. So to understand how your machine works, you must learn C.

Second, K&R covers far more than just a language - it presents a system of thinking about computing. It covers algorithms for line counting, memory allocation, and the construction of linked lists. The techniques you learn for pointer manipulation will enable you to author implementations of data structures that are considerably more elegant and simple than the naive methods [2].

It contains fundamental, influential insights about the computer science field on nearly every page.

Still, I'd discourage you from writing C professionally if you can help it. Rust retains nearly all the advantages of C while abandoning its unsafe memory model and replacing it with compile-time determined allocations and de-allocations. As roughly 70% of CVEs in C or C++ projects are caused by memory unsafety, this is a significant feature [1]. But the insights you make in studying K&R will translate to everything you build, regardless of language.

You'll walk away with a mental model for computers so powerful that it was the genesis of our modern technology era.

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

[1]: https://www.chromium.org/Home/chromium-security/memory-safet...

[2]: https://grisha.org/blog/2013/04/02/linus-on-understanding-po...

vngzs··on An open-source, private, & secure iOS web browser
Snowhaze came out on top of an iOS browser comparison I recently read [0], so I decided to give it a shot. It is far and beyond the privacy features of most other mobile iOS browsers.

In addition to sandboxed tabs, reasonably secure defaults, and robust content blocking built-in, it also offers some interesting features like user-agent randomization, tab masking, and URL tracking depersonalization.

I'm completely unaffiliated with Snowhaze, just sharing a recent discovery with HN.

[0]: https://web.archive.org/web/20210827144742/https://blog.then... (site's down so archive.org it is)

← PreviousPage 4 of 10Next →