HNHacker News
TopNewBestAskShowJobs

querez

976 karma · joined April 11, 2020

submissionscomments
querez··on Bluesky's at Protocol: Pros and Cons for Developers
As someone who's currently on the brink of signing up to bluesky: what's the recommended way of picking a handle, then? Are there any downsides to going with a normal `@name.bsky.social` thing first and later switching if I feel there is a need? Or should I best use my own domain from the get go?
querez··on I got dysentery in a human challenge trial
> The presence of that phrase alone is enough to instantly downvote and flag any and every post on sight, even those that are otherwise neutral or even positive

I feel you're severely overreacting. If everyone would act that way, this community would definitely be much a worse place. Kant's Imperative and all that.

querez··on I got dysentery in a human challenge trial
The discussion is pointless and was meant as a jest, but

> And the question was about the number of people. You included statistics regarding the number of events.

The poster just said "more common". It's your interpretation that this was "a question about the number of people". But that's just one of many possible framings of the question. "The number of events" is an equally valid one.

querez··on I got dysentery in a human challenge trial
Oh you sweet, sweet summer child... While it's possible that you're right, I don't think that's as clear as you make it out to be. A quick google didn't give me any concrete numbers, but at least for Urine, according to Wikpedia, " watersports (meaning urolagnia) was ranked ninth in popularity among sexual fetishes in the United Kingdom"[0], and " 1999 study of 164 males in Finland from two SM clubs found that 18.2% had engaged in coprophilia"[1] (of course, this number is for a sub-population that is more likely to try out kinks).

I don't have good numbers for birthing without any anesthesia either, but the Australian Institute of Health and Wealthfare claims that "In 2022, 4 in 5 (80%) women who had labour in Australia received pain relief"[2].

Giving birth is a "0-3 times in a lifetime" event, while sex is a much more high-frequency thing. I'm assuming even people who engage in these fetishes do that every time they have sex, but still many more times than they'd give birth.

So let's try to come up with some numbers: lets say we have 8 billion people on earth, and 1% of the adult population engages in watersports or scat, and they do so 10 times a year for 30 years of their life. That's 300 * 80M = 24B times someone eats shit / drinks piss in the next 30 years.

On the other hand, let's say we have ~4 billion women on earth, each gives an average of 2 children in the next 30 years, and 20% do that without pain relief. That's 4B * 2 * 0.2 = 2B births without pain relief in the next 30 years.

I don't think that what you're saying adds up.

Sources:

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

[1] https://en.wikipedia.org/wiki/Coprophilia

[2] https://www.aihw.gov.au/reports/mothers-babies/australias-mo...

querez··on Nobel Prize in Physics awarded to John Hopfield and Geoffrey Hinton [pdf]
what makes you say/think that?
querez··on FFmpeg 7.1 release: a tons of codecs
> Getting to shrink my DVD and Blu Ray rip collection down by moving it all to AV1 at rest would massively reduce my storage cost

SVT-AV1 is pretty fast these days, Almost every minor release added last year added an additional 10-50% speedup. So if you haven't checked it out for a year or so, you'll be amazed by how much quicker it has gotten. Just saying.

querez··on [dead]
So, if I understood correctly, they had a liquidity event where people could cash in their options/shares, and people did. They seemed to expect this expense, IIUC they even took on capital (Series I funding) to make sure they can afford it. The loss is not due to bad financial performance at all (the article mentions increased sales by 30%). Hats off to Stripe for making sure that their employees, who had to sit on their options, had a chance to finally cash in on the success of the company.

(If my understanding is correct, the title is quite misleading, because it makes it sound like Stripe was struggling. When in fact they flexed their muscle by taking on cash to hand to their employees).

querez··on 13ft – A site similar to 12ft.io but self-hosted
The problem is that this might end with a bad incentive structure that I think would lead to bad quality: it would push you to write the kind of articles that everyone wants to click on/pay for. So mostly clickbait. Emotional content instead of factual one. It's unlikely that this could finance the long-term, high-quality investigative journalism that actually defines high quality journals.
querez··on 13ft – A site similar to 12ft.io but self-hosted
I disagree, I think the current approach actually makes for a better and more open web long term. The fact is that either you pay for content, or the content has to pay for itself (which means it's either sponsored content or full of ads). Real journalism costs money, there's no way around that. So we're left with a few options:

Option a) NYT and other news sites makes their news open to everyone without paywall. To finance itself it will become full of ads and disgusting.

Option b) NYT and other news sites become fully walled gardens, letting no-one in (including Google bots). It won't be indexed by Google and search sites, we won't be able to find its context freely. It's like a discord site or facebook groups: there's a treasure trove of information out there, but you won't be able to find it when you need it.

Option c): NYT and other news sites let Google and search sites index their content, but asks the user to pay to access it.

querez··on The Greatest Educational Life Hack: Learning Math Ahead of Time
My daycare has a lot of non-native people who do not speak the local, native language with their child, at all. Still, all children (age 3, they're usually in daycare since age 1) speak the local language fluidly, thanks to how much they they spend in daycare.
querez··on Fastest rate of natural carbon dioxide rise over the last 50k years
It's the consumers first and foremost. The multinationals are just doing what's in their best interest.
querez··on xLSTM: Extended Long Short-Term Memory
In this specific case, it's fairly well known that Hochreiter was the major brain behind the original LSTM.
querez··on I'm no longer involved at Signal
ah thank you, I wasn't aware!
querez··on Tell HN: You can't search your WhatsApp without exposing yourself to metaAI
Without any context this is a super-unspecific claim. Do you mean, like "They use an LLM (in inference mode) somewhere in their search-pipeline"? If so, then: well, no, d'oh. I'd be surprised (and disappointed) if in 2024, they didn't, given that they're one of the leaders in the field. I see no issue with this.

The question is if they store your search strings (that's an issue separate of AI), and if so, then whether that data is used for AI training. And in the latter, the question is should they (ethically)?, could they (legally)?, and are they?

Without more details, your post as extremely high entropy wrt. how much we should care.

querez··on I'm no longer involved at Signal
Well this is certainly an unexpected and weird announcement. I actually thought Moxie was (the main person behind) Signal. Not giving any details about this feels weird to me. How are we supposed to trust an organization that prides itself to be the beacon of free communication, when they communicate so... cryptically? Does anyone have more information on what happened here? (And what/if any are the current alternatives?)
querez··on Signs you're about to be acquired (2020)
Insider trading, or starting to look for a new job.
querez··on Delta Lake vs. Parquet: A Comparison
I'm not well versed in these things, but at this point, aren't you re-inventing database systems? Talking about things like ACID transactions, schema evolution, dropping columns, ... in the context of a file-format feels bizarre to me.
querez··on FAQ on Leaving Google
Unfortunately, that's exactly the thought that killed Google Brain :(
querez··on Bad scientific code beats code following "best practices" (2014)
It's a decent measure of complexity: It's not that "opening files" themselves is work-intensive. But having a lot of files smells of overengineered code. One long, yet simple function has less cognitive overhead than spreading the function across multiple classes or functions or call hierarchies (themselves spread over multiple files).
querez··on ArXiv now offers papers in HTML format
Am researcher, care about responsiveness way more than pretty. I am super glad for the option. Downloading PDFs is super annoying. I'm stoked.
querez··on Spotify will reduce total headcount by approximately 17%
Of course it is. That's why I'm pointing out that everyone learns the lesson eventually. If it were obvious from the outset it wouldn't require learning. If it was easy to discover, it would be learned immediately. But as it stands, it's a difficult thing to learn because it's not in the employer's interest: they definitely do want people who are are better aligned with the goals of the company outside of "work for payment". They want people that are passionate about the job and the company, because ceteris paribus those people will do a better job. And a way of getting there is trying to be a company worth working for. To have a good culture, great goals, or great colleagues. But taking a more cynical view, that's just a different form of payment, in the sense that 60k/year at a great place to work is probably worth more to you than 70k/year at a not-so-great job. But as soon as a company enters some hardship and needs to cut cost, it needs to take a sober and realistic look at its expenses. And personal costs tend to be a huge expense.
querez··on Spotify will reduce total headcount by approximately 17%
> I feel like I've woken up to the nature of the employer/employee relationship with all these layoffs. I was blind but now I see, we ARE human resources

I think that's just a normal process of growing older. Also: yes, of course you are a resource to your employer. An annoying one (stell or laptops don't usually talk back and has demands). I also don't think there's anything wrong with companies behaving this way. By all intents, purposes and employment contracts, it was always meant to be a "work for payment" type deals, nothing more. Of course, ideally your interests and the companies' align beyond that narrow scope; e.g. you're both interested in you becoming an expert in X, or in solving problems in area Y. But it really comes down to you being a resource for the company to produce stuff. Thinking back I realized that my elders and mentors and everyone worth listening to always tries to tell me that (e.g. by telling me "you're not your job", "look out for your WLB, noone else will", "a job is just a means to an end", "HR protects the company, not you", "always keep in mind that you're disposable"...). And yet it seems to be a lesion that everyone needs to rediscover for themselves eventually.

querez··on Guide to Adopting AV1 Encoding
av1an is essentially a "wrapper" around other encoders. I've played around with it lot in the past, but have never seen any real quality gains from using it (as measured by VMAF) instead of just using the encoder directly. Am I doing something wrong? What exactly does av1an buy me (except maybe for better independent threads?)
querez··on Ways to teach kids to code (2016)
> Now if only I could get him to care about anything more than his hotwheels collection...

he's a kid, es supposed to care more about that than anything else.

querez··on A dive into the AMD driver workflow
> G/A can't be used for large scale training, nobody is going to give their data to them. Major trust issues there.

And yet, that's what a lot of the big AI startups are doing. Granted, it's not what everyday business are doing (yet). But TPUs offer pretty impressive perf/cost ratio, so I'd be surprised if it's actually "nobody".

querez··on The damaging results of mandated return to office
To each their own. I find that WFH makes me incredibly unproductive, I'm glad I'm able to return to to the office. I think there are a lot of factors that affect this (e.g. I have a family with small children at home, and my commute is <15 minutes by bike).
querez··on Livestream of Reddit subreddits going private in protest
That's not really decided yet. Without going into too much detail, a lot of our moderators are not as engaged as they once were: Eternal September has washed out many of our core contributers long ago, and the subreddit's focus has shifted over the years. Mostly to topics that (IMO) most mods don't care too much about (self-promotion, "beginner friendly content" instead of expert discussion, fluff...). So in my view, most of the subreddit is in strong decline and is just not "the place it used to be". With several million subscribers, I feel like it's very hard to foster a real sense of community. And once you hit a certain popularity, the amount of newcomers will always be larger than the "core", and good content erodes from there. I guess that's true of a lot of reddit, though. We could've tried harder to prevent this (or guide our community), but at the time we felt that "hands off" was a better moderation approach, and now it's likely impossible to change that course.

Anyways, IMO most of the really interesting discussion on the topic happens on twitter already. My best guess is that instead of finding a new home for the subreddit, we'll eventually just either give control to some other body (or reddit takes it from us) or we'll just set the subreddit to public again and let it go its way without being very involved (Most of the current mods weren't really involved anymore anyways). I'd love to find a new place devoted to this topic (twitter definitely sucks for long form discussion), but I haven't found such a place, and don't have the time/energy to set one up myself.

querez··on Livestream of Reddit subreddits going private in protest
Meh, I wouldn't be so quick to judge. I'm a mod of a several-million-subscriber subreddit (one that's likely highly popular among HN subscribers). When we asked our community if they wanted to join the protest, the answer was a very resounding "yes, but 2 days aren't going to cut it"...

So we're still debating when/IF we're going to return after going dark. To add, all the mods are tired and don't have enough time to dedicate to the subreddit as it is (with mod-tools working, which won't be the case after the API changes). Only time will tell, but we're most likely only reactivating our subreddit once reddit signals clear willingness to ease on their squeeze. By which time a lot of damage (in terms of subscribers moving on) will be done. Also, some of the mods have said they won't come back after this, as this was the final drop in the bucket and they'll delete their account today.

So all in all, there is some hope that this does serious damage to reddit (at this point, they're clearly seen as the enemy by a lot of us). Talking to mods at other very large subreddits, we're not the only ones thinking along these lines.

querez··on CPR's true survival rate is lower than many people think
Not sure what the situation is in the US, but I've taken 3-4 first aid courses in my life (it's compulsory to get a driver's license where I live). Every time, they told us that CPR survival rates were abysmal if we aren't also using an AED. IIRC, with an AED the survival rate goes up significantly -- though this article states that in-hospital survival rates are only at 17%, which indicates it's not all that great with AED, either.
querez··on CPR's true survival rate is lower than many people think
When I was thought it, it was clearly pointed out to me that chance of survival is ~5%, but increases significantly if an AED can be used in time. I've also been told about the rib cracking etc. Just as a counterpoint.
← PreviousPage 2 of 7Next →