• 1 Post
  • 72 Comments
Joined 3 years ago
cake
Cake day: June 20th, 2023

help-circle

  • THANK YOU.

    I migrated services from LXC to kubernetes. One of these services has been exhibiting concerning memory footprint issues. Everyone immediately went “REEEEEEEE KUBERNETES BAD EVERYTHING WAS FINE BEFORE WHAT IS ALL THIS ABSTRACTION >:(((((”.

    I just spent three months doing optimization work. For memory/resource leaks in that old C codebase. Kubernetes didn’t have fuck-all to do with any of those (which is obvious to literally anyone who has any clue how containerization works under the hood). The codebase just had very old-fashioned manual memory management leaks as well as a weird interaction between jemalloc and RHEL’s default kernel settings.

    The only reason I spent all that time optimizing and we aren’t just throwing more RAM at the problem? Due to incredible levels of incompetence business-side I’ll spare you the details of, our 30 day growth predictions have error bars so many orders of magnitude wide that we are stuck in a stupid loop of “won’t order hardware we probably won’t need but if we do get a best-case user influx the lead time on new hardware is too long to get you the RAM we need”. Basically the virtual price of RAM is super high because the suits keep pinky-promising that we’ll get a bunch of users soon but are also constantly wrong about that.



  • It’s like every other media industry. The monoculture is dying. Everyone’s who’s “about it” is into niche subcultures and micro-celebrities you’ll probably never hear of.

    There was a weird period of time from the mid-20th through the early 21st century where radio and TV had very strongly concentrated media production which made up most people’s media consumption.
    For the last 15 years or so the tools of professional-looking media production for mass consumption have been available to anyone with a few hundred bucks to spare.

    In some ways it’s a communist utopia. The means of production have been commodified so much virtually anyone can afford them. However capitalists have moved on from owning the means of production to owning the means of distribution (the platforms).


  • I don’t trust the US government to do literally anything right with this, and I’m kinda surprised Google didn’t already gift an underage child to Trump so he’d make the problem go away.

    However a perfectly viable option that I’m sure the previous government looked into would be to entrust Chromium (which is Open-Source though not copyleft) to a new, independent nonprofit made of Google’s former chromium team led and paid for by a consortium of the major commercial chromium users (Google, Microsoft, etc.). It would be in everyone’s best interest to share the relatively small financial burden so that Chromium can remain decent and competitive.

    This wouldn’t be anything revolutionary. This approach of financing an independent open-source project as a “common good” is basically how the Linux kernel has been developed for many years now, most Linux code is written by corporate sponsors.


  • music-library $ du -h -d 0 .
    270G	.
    

    I am not looking for a compromise. I listen to my high-quality digital library on shuffle most of the time, and am very well aware that my phone allows me to access orders of magnitude more music than even the most compact CD player.

    When I do listen to my favorite albums as LPs, the clunkiness and the artifacts are part of an Experience. I can listen to exact copies of the digital masters of those songs any time I want to, but sometimes we do things BECAUSE they are not maximally optimal. Sometimes I want to take a walk alongside the river and get my feet a little bit wet even though I could have worn boots. Feel a little something, you know?



  • This kind of shit happens with a similar frequency… on Arch Linux. It’s rolling release, shit happens sometimes. archlinux.org’s homepage actually lists past major packaging issues.

    Debian however is rock-fucking-solid. But so is Windows Server, I hear. The problem is that Microsoft is treating Windows Home/Pro like a rolling release distro, and the users are guinea pigs. I guess Microsoft is right though, their users will eat it up 'till shit is spilling out from both ends, so why bother?



  • You’ve completely missed the point.

    You grew up in a world where the quirks of analog formats were nothing but technical limitations to be overcome. It is true that a FLAC is literally superior in every way to a Vinyl if your value function only takes in cost, quality, and convenience.

    HOWEVER Gen Z grew up in a world where music was always cheap and convenient to access. We also (mostly) grew up in a world of touchscreens and always-online gadgets and doodads. My generation’s first portable music player was often the iPod touch. You know what all of that does to a person? It creates a deep craving for tactile feedback. For technology that doesn’t nag with software updates, for music that can’t be “unlicensed” and pulled from your library remotely, for a music player that you can touch and feel and interact with in a more meaningful way than tapping on the little square of glass that already runs our lives. For the little rituals that have been stripped away, like flipping a vinyl at the midway point or rewinding a tape.

    The entire point of analog is that it’s “worse”. It’s un-clinical, it’s raw, it’s tactile, it’s physical. Listening to my favorite albums on vinyl is such a better experience than through the disembodied shuffle of my phone. I don’t crave maximum audio fidelity or convenience because I always could have those things literally whenever I want.


  • The hard part isn’t processing payment… They already basically do that for themselves with the steam wallet.

    The problem is getting the ability to withdraw funds from your customers’ bank accounts. That requires a commercial relationship with your customer’s bank and going through an insane amount of red tape. And there is no standard worldwide protocol for this, you’ll be starting from zero in every market by cold-calling major banks.

    The only viable approach is to have an army of salespeople, accountants, and project managers to do all those individual negotiations.

    The EU has been trying for years to have an indigenous continent-wide payment processor. The first attempt failed, now Wero is poised to succeed in the next few years but that’s building off negotiations that started a few years ago with pressure from the EU and buy-in from the financial sector, and still only a handful of European markets have been integrated at this point.

    Now imagine all this difficulty but you have to also get active buy-in from every market worldwide. There’s a reason Visa/MC have a near monopoly on international payments in the western world, and it’s not that no-one else thought to get a piece of that very juicy pie that’s making them literally billions in profit every year.



  • Wikimedians discussed ways that AI/machine-generated remixing of the already created content can be used to make Wikipedia more accessible and easier to learn from

    The entire mistake right there. Look no further. They saw a solution (LLMs) and started hunting for a problem.

    Had they done it the right way round there might have been some useful, though less flashy, outcome. I agree many article summaries are badly written. So why not experiment with an AI that flags those articles for review? Or even just organize a community drive to clean up article summaries?

    The questions are rhetorical of course. Like every GenAI peddler they don’t have an interest in the problem they purport to solve, they just want to play with or sell you this shiny toy that pretends really convincingly that it is clever.


  • This is separate from A-GPS. Google seems to be using WiFi rather than Bluetooth, but the broader point remains the same. No one is stopping any vendor from crowdsourcing the location of every BT device… which is what Apple has done, for Airtags which don’t have the battery capacity to run a GPS chip.

    Sure without GPS it wouldn’t be very effective to rely on only nearby devices to guess the current location. But an attacker only has to get lucky once to get your home address. So the only safe approach is to hide nearby devices/networks from unauthorized apps.


  • Every Bluetooth device has a unique identifier. Any phone that has seen that Bluetooth device in the past could have told google/apple/whoever “hey BTW this device is at those coordinates”.

    Google already uses this with WiFi to help “bootstrap” GPS localization. It is much faster to get a GPS fix if you already know roughly where you are (a few seconds vs a couple minutes), so they use nearby WiFi/Bluetooth devices to determine that. Remember 10-15 years ago when getting a GPS fix took forever? GPS didn’t change, this did.
    Apple went further and does this with Airtags now. Every Bluetooth device that ever went near an iPhone is in Apple’s database with GPS coordinates.

    So unless you live alone in a mountain cabin that has never been visited by someone with a smartphone before and you didn’t disable the “enhanced localization” feature on your phone, yes your Bluetooth is at risk of giving up your location.


  • Plenty of cars flash their brake lights when ABS(/ESP?) engages, which is reasonable and should be a legal requirement IMO.

    There’s lots of room to give additional info in between that and “brake light is on because the driver doesn’t understand that they can do mild adjustments by letting off the gas / stupid bitch-ass VW PHEV computer thinks using cruise control downhill with electric regen requires the motherfucking brake lights”. It’s like no-one realizes or cares that brake lights lose all purpose if they’re on when the car isn’t meaningfully decelerating. ARGH.



  • It’s because the ~*~tech~*~ sector fundamentally relies on different economics than most engineering companies, and that has investors absolutely bricked up.

    What investors being sold by “tech” companies is infinite ROI. Sure, [YouTube/Twitter/Uber/whoever] has never been profitable more than a few quarters in a row (if that), but think! They have virtually no fixed costs! That means if we just inject a few more millions in R&D we will finally reach the threshold where we can scale deployments to hundreds of millions of users who will be paying us MRR! Hosting costs are virtually nothing and at that scale R&D is basically free as well! And if push comes to shove, we can reduce costs to nearly zero by firing all the engineers! The economies of scale are practically infinite, they say.

    It’s the rare instance where capitalists actually care about long-terms gain a bit too much. The tech industry tends to be single-mindedly chasing monthly user counts first and revenue second or third. Then at some point reality catches up, the accountants start getting their way, the product starts getting enshittified, and the users leave for something else. Did the product actually turn a net profit over its lifetime? Who knows, who cares. Everyone who made those early business decisions has long since cashed out.

    Where the markets are unbelievably irrational is that this frenzy has spilled over into industries where the the sales pitch for infinite economies of scale doesn’t even make theoretical sense. Tesla sells physical products, so why are they worth more than every other automotive company combined? OpenAI operates at an enormous loss because LLMs are just expensive to train and run by nature, so they cannot be profitable under the current business model at any scale. Yet here we are. Just because it’s labeled as “tech”, investors are throwing our retirement funds into it. And any time the markets are being irrational, there’s a risk that investors wise up to the bad fundamentals and the whole thing comes crashing down.


    In Europe we’ve been spared some of the worst of the craziness. Although venture capitalism is alive and well in the software sector, I would wager that European companies tend to have stronger fundamentals on average (but that’s just a gut feeling, I’m not an economist).



  • I know people in that predicament and they’re, charitably, helpless little babies when you tell them to read two paragraphs of documentation on how to run one command in a Linux CLI.

    Fundamentally nothing out there really caters to the needs of resellers. Your average resale company couldn’t automate a backup job to save itself from bankruptcy if it doesn’t come with a neat GUI, a 24/7 support contract, and preferably a Microsoft or oracle logo somewhere in the corner to inspire confidence.

    Like I jest but there are Microsoft outfits and FOSS outfits and there is essentially zero professional overlap even though they both sell IT products/solutions. The disconnect is a mile wide. Which translates to wildly different business models where the FOSS people have been running shit in containers for 15 years while the Microsoft slaves are still licensing their monolithic solutions by the CPU Core and doing weird-ass shit like buy 4-core xeons because it’s more economical with these archaic licensing models.

    So sure Proxmox/Suse are certainly very happy with their sales number right now but anecdotally I’m not seeing the migration frenzy that one would expect under such intense price gouging. Broadcom correctly identified that it will take years for these super corporate structures to steer away from “the way we’ve always done things” and in the meantime that’s untold millions in additional short-term profits.