HNHacker News
TopNewBestAskShowJobs

kbenson

31,676 karma · joined July 26, 2012

k e n t r a k # google's public mail service
submissionscomments
kbenson··on XScreenSaver is available on Android. Google requires a privacy policy
> It's just a useful example of something that plays well in the HN comments section but not in the real world

I take it you're referring to the earlier, short one, of "We don't store your data, period." If that's the case, I think you're saying the same thing I did in a slightly different way and emphasizing different points? I agree, there's a prevalence of people that see simple and direct and make the mistake of equating simple with useful. Something being simple can be good, but not at the expense of greatly increasing the ambiguity of what it applies to.

kbenson··on XScreenSaver is available on Android. Google requires a privacy policy
I appreciate that the version they indirectly encouraged you to use by rejecting the other one is much more clear about how you not only don't store anything, but don't collect anything. For the pedanticists here which might argue you can't collect anything without storing it, I suspect there's a lot of wiggle room there (even if only in the developers mind and not legally), and the additional information is useful to guard against that.

For example, getting paid to receive and forward data, or to send it directly to a third party may be something a developer could do and thing they are complying with "we don't store your data". Technically correct, probably not legally correct, but unlikely to matter legally unless noticed and they are sued.

Much better to be explicit about things as long as it's still easily readable.

kbenson··on Dmv.org
Every time I search up a DMV question I get a DMV FAQ link that answers my question. Nine out of ten times when researching a tax question (I do my own taxes) I get a link to the IRS or the franchise tax board for my state with the specific directions I need.

When I'm searching for in for that the ultimate answer is coming from a government entity, usually it's because I want authoritative information, and I'm not willing to rely on a third party hopefully getting it right.

The only thing I want from Google when making these searches is to deep link me into the appropriate section of the official site. The only reason I'll visit a third party site is if I don't have a good official link and I'm hoping to get a clue to refine my search by reading specific laws or statute id's they might reference, since I'm not going to make tax filing decisions or assume directions given regarding a DMV process are correct from some random online website.

kbenson··on Microsoft AI spying scandal: time to rethink privacy standards
> I still don't understand why you're trying to shoehorn this into a narrative about being poor

Well, part of it was that a sibling reply[1] was already present, and the idea that "oh, we can all just run our own AI models at home" seemed ludicrously costly in both money, and time to me. It's not that I was trying to shoehorn it in as much as juxtapose it against what was already being stated.

> But it's not like once someone reaches "not poor", time for self-actualization abounds.

I have no idea why you you are interpreting the inverse of my statement that being poor makes this harder to mean that not being poor makes it easy or trivial, which is how it appears to me you're interpreting my statements.

Privacy is hard in the current climate. It takes time and effort or money to alleviate some of the privacy concerns, to different degrees depending on the specific concern. Those are all things the working poor have less of than other economic groups, so solutions that require them will likely be less used by them. That doesn't mean don't offer them, but we should keep that in mind when advocating for people switching to open source solutions, so we make sure the solutions actually solve the needs of the people intended to use them, and not just the needs of a subset of those people.

It would be a real shame if what I think is one of if not the best option for solving this problem in an egalitarian way failed to take into account the needs of that economic group and we actaully ended up with a solution that many are protected by, but disproportionally not the poor for multiple reasons. A world where 90% of people out of poverty are protected in some manner but only 50% of people in poverty are isn't necessarily better IMO, and may actually be much worse, since I'm not sure there will be much incentive to fix it at that point.

No, I wasn't very clear in my initial comment (and wasn't really going this deep), I just wanted people to consider that running your own LLM for your own needs isn't really feasible for most people and in a more general way than a reply to that person specifically might have communicated, and I was slightly inebriated when I wrote it, so didn't think it warranted much explanation.

1: https://news.ycombinator.com/item?id=40593085

kbenson··on Microsoft AI spying scandal: time to rethink privacy standards
> While middle class people are more likely to pay for things that need skilled repair (due to knowing fewer people who can do such work informally), they're not paying for things they can do themselves like housecleaning, landscaping, etc

It's not that they're paying for all of them, but that they have the option of paying for them and I would hazard that a good portion of middle class families with both parents fully employed opt for offloading at least some of that some of the time if they live in a portion of the country where middle class means you actually have disposable income. E.g. many people pay for larger landscaping projects, or hire a handyman to do cleanup around the outside of the house or even might have a maid come in once a month. This might be less common now that the middle class has eroded to some degree, but I think that's a problem of a shrinking middle class, not of those being things the middle class doesn't do.

> And apart from some relatively affordable table stakes, I don't think the financial cost is really one of those.

I'm not sure I agree with that. Having spare hardware to run something, a stable place to put it, and paying the power for it (a minor but increasing cost) all play a part.

> The attention cost of self-actualizing and the ambiguity of using a non-advertised solution are though.

I agree, and I think the working poor have less attention to spare because of less free time, but also that they generally have more and larger worries that make this problem seem insignificant by comparison. If your major worried are making rent, having enough money for food, repairing your car so you can effectively get to work and take kids to school, spend quality time with your family, and set up a solution so that your privacy is protected from companies that want to monetize you, which one will get the least attention? Middle class families probably have at least one less of those concerns, and every major concern that's more important than using free software to avoid a small but persistent exploitation is something that competes for attention.

kbenson··on Microsoft AI spying scandal: time to rethink privacy standards
Poor, as I intended it, was meant to encompass many things. Working poor people have less money, but they also often have less free time because they have less money. If you have plenty of money you can shop around for jobs that fit your schedule better, or if your job allows for it take time off, or outsource some work you would have to do yourselves (housecleaning, maintenance, landscaping, child care when you're unavailable, cooking, etc). Additionally, if you're in a family, there's a higher chance that both parents will need to work full time, meaning there's less free time for all those things mentioned above, which if one parent worked part time or not at all could allow for a lot of those things to be done while one is working, leading to more free time in the evening for both.

Given that, who's more likely to look into free alternatives to apparently free services online? The middle-class person with a spouse that works part time and takes care of many of the chores and that pays for a handyman or contractor or repair man to fix appliances and household problems or the poor person that works full time, their spouse works full time, and when they're done working they're busy doing the chores that life requires because paying someone else is not feasible for them?

Time is money, and the working poor have neither.

kbenson··on Microsoft AI spying scandal: time to rethink privacy standards
Caring about privacy has always been a rich person's game, since the poor have little to pay with other than their data. That's worth keeping in mind.
kbenson··on Microsoft AI spying scandal: time to rethink privacy standards
> I think many people here on hackernews live in a bubble and just don't understand what is an average human being.

Plenty of us understand, but it's depressing and usually trying to explain to others that have a rosier view of how people will react feels a lot like trying to get the people themselves to care, which is to say it's hard, thankless, and depressing, so few people bother.

kbenson··on Debian's /tmpest in a teapot
> The essence of the problem is that there's no standard pathname for a personal directory that's guaranteed to be on local disk, even if $HOME isn't. Consequently, people have relied on /var/tmp/$USER for this. There are realistically affected users who can't change the new defaults.

How is that not a site specific problem they introduced? There's no standard pathname because there are valid reasons why you want to have nothing on the local system, such as when it doesn't have a local disk.

If you (read as "the admins") have configured the system where there's very few acceptable locations to store local data, and that's a specific need, then it's up to the admins to provide a solution to the problem, not the Debian distro, which is flexible enough to handle this fine.

It's not that hard to create your own location for data if you need to, and symlink to it from other locations if required. Need a /tmp/$USER to be a local disk location and /tmp is a ramdisk? Create a script that sets up the correct local location and created a symlink to it in /tmp and make sure it's run on boot and maybe once daily. Both cron and systemd solutions could work for this. Worried that user tmp files will be cleaned out when they shouldn't be? Put in the correct exceptions for tmpwatch or whatever debian uses, or disable that service.

Not only is this not rocket science, it's literally the job of the admins that design the OS deployment so that it works for the intended use cases.

kbenson··on Nginx Unit: open-source, lightweight and versatile application runtime
Windows has the same thing, it's just much less exposed, and none of the paths are magical, they're well defined and mostly adhered to for all major distros. The core of how it works in Linux is fairly straightforward.

The main difference is in how additional software is handled. Windows, because of its history with mostly third party software being installed, generally installed applications into a folder and that folder contained the application... Mostly. Uninstalling was never as simple as that might imply.

Linux distros had an existing filesystem layout (from Unix) to conform to, so when they started developing package managers, they had to support files all over the place, so they make sure packages include manifests. Want to know where use executable are? Check bin. Superuser.com executables? Check sbin (don't want those cluttering the available utils in the path of regular users). Libs for in libs.

/bin and /usr/bin and the others are holdovers from the long past when disks were small, and recent distros often symlink usr to / so they're different in name only. /usr/local is for admin local modifications that are not handled through a package. /opt is for whatever, and often used for software installed into a contained folder, like in windows.

Just know what bin, sbin, lib, opt and etc are for and most the rest is irrelevant as long as you know how to query the package manager for what files a package provides or as it what package a specific filer belongs to. If you looked I to windows and the various places it puts things I suspect you'd find it at least complicated, if not much more.

Note: what I said may not match the LSB (which someone else usefully posted) perfectly, but for the most part it should work as a simple primer.

kbenson··on The Pumpkin Eclipse
I didn't mean to imply it was exactly analogous, just a lot simpler, and there is a lot to be gleaned from that data. In fact, I would assume the more data you pass over it (e.g. if you proxy all your traffic across TOR) the easier it is to make assumptions about the source, unless it actively splits it across different exit nodes, which seems like it could be problematic in a lot of cases. If you have a DigitalOcean personal server you run some services on and you access it through TOR... well you might have just made the job of anyone trying to deanonymize you much easier.
kbenson··on The Pumpkin Eclipse
> If these guys can do it, remind me again how Tor is secure because nobody could possibly be able to follow packets from your machine, through the onion hops, to the exit node where the same packet is available unencrypted...

You're supposed to be protected by the fact that you're going through multiple nodes before exiting TOR, and traffic should be mixed. Can you find some streams if you have most/all the nodes within your network and can analyze the traffic? Probably some, but the more traffic a node handles the harder it would be.

There is a simpler approach though, which is to just run exit nodes.[1]

1: https://en.wikipedia.org/wiki/Tor_(network)#Exit_node_eavesd...

kbenson··on I got tired of hearing that YC fired Sam, so here's what actually happened
Eh, it's an average of one a day, but it could be rolled up in a report once a month and there's 20-40 to look through, and depending how hands on he is (he might have delegated some of this work) that may or may not be something he looks closely into. I imagine there's an order of magnitude more companies that apply that don't get funded. Where does he spend his time? Mentoring companies that have been accepted, and if so, all of them or ones that align with some prior interest, or in vetting and looking for unicorns in the applications?

I don't know how any of their stuff works, but I'm sure most people can appreciate that the stuff that gets lost between the cracks is the stuff that doesn't come in and get tracked through the normal channels, and an OpenAI investment where one of the main people in the company is heading it might be something that wasn't through normal channels. Maybe that was on purpose so pg didn't see it, or maybe that stuff is mostly handled by someone else. Or maybe pg knew about it and decided to lie about it on twitter. /shrug

kbenson··on I got tired of hearing that YC fired Sam, so here's what actually happened
That is weird. Although OpenAI isn't even listed in their startup directory, so maybe the investment was done differently, and not through their normal channels. Or maybe they divested their share after a short period. Whether that means pg would have been more likely or less likely to know about it I have no idea.
kbenson··on The Pumpkin Eclipse
> and it's been priced out of the market.

Whether that's likely is entirely based on the cost of the device. Some things are simple and cheap and extra hardware cuts deeply into the profit. Others are not but this sort of thing is also important because they are remote and you don't want to have a person go out on site. When the device is expensive enough or sending someone to the site is expensive enough, "just ship a replacement" is not really a viable solution, unless you're installing it in a high-availability capacity where you can fail over to it without physical intervention.

Obviously it's not a solution for every circumstance. Nothing really is. I don't think it's useful for us to assume that a solution has to be, as that doesn't really help us in the many instances when it's good enough.

kbenson··on I got tired of hearing that YC fired Sam, so here's what actually happened
Is it? They state they've funded 5000 companies since 2005[1], and if that was evenly spread per year, that's about a company a day. Maybe pg only pays attention to the ones that other people bring to his attention given that amount, and maybe he "knew" about OpenAI in that it was in some report he skimmed or was mentioned but maybe it was seen as entirely handled since it was a project for someone else there. It could entirely have been out of mind within a week or two of him "knowing" it and then it's quickly forgotten, and may seem like new information when it comes up years later.

1: https://www.ycombinator.com/companies

kbenson··on The Pumpkin Eclipse
> The full storage/kernel/network/UI stack required to make that happen isn't ever going to run under "ROM" in the sense of truly immutable storage.

Why not? I'm essentially describing a specialized OOB system, and it would just use a carved out small chunk of system RAM or ship with a minimal amount RAM of its own. If you mean actually impossible to change because it's physical ROM ("truly immutable"), that's less important to the design than there's no mechanism that allows that storage area to be written to from the system itself, whether that's just the very locked down and minimal recovery kernel it houses not allowing it, or a jumper.

kbenson··on The Pumpkin Eclipse
You could but a base level firmware on ROM, with a hardware trigger, and all that does on boot is listen and receive a signed firmware to write to the system. It needs a way to be triggered through hardware examining traffic and that also needs to require the seen command be signed. That recovery boot system needs to be as simple and minimal as possibly so you can have good assurance that there aren't problems with it, and should be written in the safest language you can get away with. Guard that signing key with your life, and lock it away for a rainy day, only to be used if much of your fleet of devices is hosed entirely. It should not be the same as a firmware signing key which needs to be pulled out and used sometimes.

I think that could work, to a degree. There's always the risk that your recovery mechanism itself it exploited, so you need to make it as small and hardened a target as possible and reduce its complexity to the bare minimum. That doesn't solve the problem, which might be inherently unsolvable, but it may reduce that likelihood of it to levels where it's not a problem until long past the lifecycle of the devices.

kbenson··on Toxoplasma Gondii significantly alters wolf behavior
I'm not sure that says what you claimed?

Overall, eating raw or undercooked beef, lamb, or other meats; contact with soil; and travel outside the country were major sources of infection.

Contact with soil is from cats shedding oocysts, as covered earlier in the same extract, and while that doesn't necessitate owning a cat, I would classify it as exposure to cats (your neighbors outdoor cat shitting in your yard is still cat exposure to my sensibilities).

The association of cats and human toxoplasmosis is difficult to assess by epidemiological surveys because soil, not the cats, is the main culprit. Oocysts are not found on cat fur and are often buried in soil along with cat faeces. Therefore, direct contact with cats is irrelevant with respect to T gondii transmission, and soil contact is universal and difficult to avoid.

Which doesn't say cats don't matter, just that direct exposure doesn't matter. Soil exposure is exposure to cats through an intermediary, just not direct. It will still likely be affected by prevalence of cats in a population, even if not necessarily at a household level.

So, I think the article answers my original question to a degree, but I also think your claim that almost everyone that has it got it through eating meat with it in it is too strong given the sources you quoted (thanks a lot for posting them though, it's always nice to have data for a discussion!). If you think my reading of it is incorrect or I'm missing something, please let me know though.

kbenson··on Toxoplasma Gondii significantly alters wolf behavior
No judgement here. If I was dating, that would be my dating pool...

In all seriousness, since this topic comes around every couple years, I've always wondered if we tested different populations of people based on career if we'd see some correlation. Cats are fairly ubiquitous across populations (but knowing if there's a higher prevalence in some would be important), so there's at least a chance for most people to be exposed I think.

kbenson··on Researchers cracked an 11-year-old password to a $3M crypto wallet
It was both, like it usually is. All that investigating and reverse engineering would have been for nought if the program didn't have the problems in the first place. Hard work is often how you capitalize on luck. Sometimes the work is enough by itself, and sometimes it's not and the luck is integral.
kbenson··on Researchers cracked an 11-year-old password to a $3M crypto wallet
Are you thinking of the one shown in Jurassic park? The scene in hackers was much more CGI, and while I don't doubt it was inspired by fsn, I'd be very surprised if it actually was fsn.
kbenson··on The Nonprofit Industrial Complex and the Corruption of the American City
Really? I find that for some people the only criteria they need for a stance, whichever side they're on, is just to know how to stand so they're opposite the other side.
kbenson··on The Nonprofit Industrial Complex and the Corruption of the American City
That's a horrible story, but pulling it out seems to indicate you didn't understand what the GP was trying to convey at all. The point was that to use a crime from twenty years ago for which a person was convicted and served time for as indicative of problems with the current project he's the head of seems to indicate they could not not find anything relevant to the project itself to criticize. Doubling down on the details of that crime doesn't change that or counter the claim at all.

If you want to make a claim that rehabilitation is impossible, or that he's not rehabilitated, then do so. If you want to state that it goes to the point being made in the article, that hiring felons with a violent history is common, then do so. I don't think the article is indefensible, but these additional details without context to make them relevant don't add much.

kbenson··on More nuclear energy means less mining
"Best" is relative to how you're measuring, and cheapest doesn't mean that's where you want to put all your investment. The cheapest option that can work for some small about of the time can be best by one measurement and not by another.

There are other problems with solar that make it beneficial to offset with other energy sources. For example, lack of ability to generate any energy for 50% of the year, and reduced efficiency depending on weather.

Wind has its own problems as well, but are map to somewhat different circumstances than solar, so is a good supplement for some of the power.

Nuclear has its own problems as well, but unlike wind and solar, many of those aren't inherent to the source, but to how we've developed methods to harness it and deal with produced byproduct (waste). I suspect a first principles approach to nuclear energy might eventually be a very useful source of energy. I think it would be beneficial if we still kept some nuclear around to keep developing it towards a better source. It does well to offset some of the requirements of other systems, which are limited by time, geography or climate.

kbenson··on Amber: Programming language compiled to Bash
Seems to render fine on Firefox mobile for me, but maybe I'm just not noticing what's broken.
kbenson··on Lumina's legal threats and my about-face
I'm mostly going by the author's own notes on how they responded, in which they state this:

• You assumed we'd never gene sequenced the bacteria, even though I posted the sequence publicly.

I assumed you did not regularly sequence the bacteria because you did not say that you did and you did not report anything about following manufacturing regulations.

That is a very weird defense to take in my opinion, and is the crux of my problem with the author in this argument. It is not acceptable to make a negative inference and then express it publicly as fact because the other party didn't mention something. That is not responsibly journalism, and to me it explains why the CEO was struggling to see good faith initially. That's not usually the behavior you see when people are communicating in good faith.

kbenson··on Lumina's legal threats and my about-face
And good faith conversations don't come away with one party publishing assertions of fact just because the other side didn't directly refute the assertion in question.

If the company thinks the journalist isn't working in good faith, are they not then justified in that?

Context is everything. I think there's poor behavior on both sides, but to my eyes it starts with a journalist making unfounded assertions. Would Lumina have gotten upset and threatened over poor press even if it wasn't to the point where someone was making statement of fact that were both wrong and perceived as hurtful to their business? We'll never know, because it appears those statements were made.

We can dumb this down and make it very simple. If someone interviews me, and asks if I beat my wife and I don't answer (for whatever reason. Maybe I perceive it as a joke that doesn't need a response, maybe I don't want to engage with bad-faith questions), and they then print that I'm a wifebeater, should I not be upset enough to threaten to sue initially? Should I assume they're working in good faith, or is that enough to assume bad faith? Personally, I think my choices would be to assume bad faith or that the person is inept and uneducated at the job, and if the person didn't seem stupid, that seems unlikely.

kbenson··on Lumina's legal threats and my about-face
That's probably true, but I think that's actually kind of sad. We've seen enough legal threatening to prevent the truth from getting out that I think our reactions are to assume that's the case. If indeed this is Lumina being upset that a journalist went to far and made false claims about them as they indicate, and the truth is better than indicated, would we not expect a legal proceeding if the journalist would not retract false claims?

I don't know who is ultimately in the right here (but at least one of Trevor's defenses of his assertions seems flimsy at best to me, namely the "you didn't say otherwise so I assumed this was true" one), and it would have been better if they could come to an agreement without the legal system about how best to handle this (Lumina does seem to be going harder at this than possibly needed), but I think we should assess the situation on the merits, not that there is a legal case at all.

kbenson··on Lumina's legal threats and my about-face
Although this part (as noted above):

• You assumed we'd never gene sequenced the bacteria, even though I posted the sequence publicly.

I assumed you did not regularly sequence the bacteria because you did not say that you did and you did not report anything about following manufacturing regulations. I’m glad that you did and do. Sequencing the bacteria regularly is critical to make sure what people are putting in their mouths is what they think is putting in their mouths. And by “post the sequence publicly”, you mean on Manifold, rather than on your company’s website, for whatever reason. Sorry I didn’t check all of Manifold.

really doesn't paint Trevor Klee favorably to me. I'm not sure how you go from "they didn't say anything about it" to "I can assert they aren't doing it". Even if you ask a question directly and don't get an answer, all you should feel comfortable asserting without evidence otherwise is "they did not address this when asked directly". In this case it's unclear whether they didn't address a question asked or the question wasn't really asked at all.

> Especially the whole idea that they wouldn't answer any questions over email, which is a red flag for me about them not wanting a record of a discussion about this.

Or it could just be that they wanted to make sure their statements were taken in context and they got a chance to explain if they thought the other person was misinterpreting the question. I'm not sure it's a *good( idea then, but I can at least imagine scenarios where they think it might be, especially if they've had prior experience with reporters misreporting because of that problem. Sometimes over-corrections happen and cause their own problems.

← PreviousPage 7 of 34Next →