• 0 Posts
  • 18 Comments
Joined 1 year ago
cake
Cake day: June 24th, 2023

help-circle






  • I’ve been an immigrant for about 25 years now. Not sure if that counts, but I’ve been slacking and haven’t gotten a citizenship yet, so… probably?

    I can confirm that I’ve up until now always been on one of the health plans my employer made available to me. It certainly made things easier that I was never out of job, and all of those jobs provided great benefits (typical white collar computer nerd stuff.)

    However, I stopped working last year, and my 18 months of COBRA (a continuation of employer-provided coverage after leaving a job, except you pay yourself the premiums your employer was paying (about ~$2000/month for me)) are running out very soon, so I’m discovering the bizarro world that is US healthcare without an employer plan.

    I’ve contacted some insurance brokers to help me find a new plan, and each one of them has tried to push weird non-ACA-compliant plans to me under false pretenses (ie. they’ve actively lied to me about what the plans were.)
    Sometimes the awful stereotypes about a profession are awful for a reason.

    Which leaves me with the ACA marketplace, where every single plan is significantly worse than anything any of my employers ever offered, both in terms of breath of network, prescription coverage, and geographical coverage. I didn’t mention the famously terrible mental health insurance coverage because it was already impossible for me to get in-network care there even with my employer plan.

    And then if you figure out which is the least bad plan in the 100+ sad plans offered to you, and you commit the faux-pas of googling them, you’ll get a deluge of screaming victims of those plans wishing they had picked anything else because their experience was a literal nightmare.

    So that’s encouraging.

    In specific terms, the ACA healthcare.gov site I linked above lets you put a list of doctors and medications to see which plans support them, and the answer for me is “none.” None of the plans available would cover all the medical care me and my wife are getting on an ongoing basis.
    So it becomes a matter of picking and choosing what I’m going to pay out of pocket.
    For example, right now I pay $0 for various insulin pens, but a great number of those plans won’t cover those, or cover a little bottle of insulin instead you’re expected to use with disposable needles each time you’ll fill yourself and inject yourself with, and hopefully not fuck it up. Out of pocket, with some “discount card” (GoodRX or whatever), a month supply of the pens would add up to roughly $800. So something that was “free” to us (if you ignore the large insurance premiums) is going to feel like quite the luxury instead.

    One of the aforementioned lovable insurance brokers suggested that I create a fake company in order to be eligible for reasonable employer-sponsored plans and avoid this nonsense. Sounds great, except for the whole fraud thing and the risk of getting found out and denied at the time when I’d need it the most (which would probably also be when an insurance provider would look the closest to try to find any reasons to deny a large claim.)

    And then, there’s the quasi-scam that targets religious (and/or desperate) people, known as “health care sharing ministries.”
    They appear to be very affordable plans with great coverage, managed by “faith-based organizations.”
    They are not insurance, and ostensibly claim to simply “share the burden” of healthcare across all their members.
    Notably, they are not actually obligated to meet any of the (low) bars set by the ACA, or to simply pay any of the insurance claims their members send them, and so sometimes they don’t. Tough break.

    At the end of the day, I’m still going to pick an ACA plan and just pay out of pocket whatever isn’t covered. I just have to settle on a plan, which feels like picking from a set of shrunken and half torn blankets the one to use on my bed.

    Anyway… what would I change? Nothing obviously. All is for the best in the best of possible worlds.



  • I was watching the network traffic sent by Twitter the other day, as one does, and apparently whenever you stop scrolling for a few seconds, whatever post is visible on screen at that time gets added to a little pile that then gets “subscribed to” because it generated “engagement”, no click needed.
    This whole insidious recommendation nonsense was probably a subplot in the classic sci-fi novel Don’t Create The Torment Nexus.

    Almost entirely unrelated, but I’ve been playing The Algorithm (part of the Tenet OST, by Ludwig Göransson) on repeat for a bit now. It’s also become my ring tone, and if I can infect at least one other hapless soul with it, I’ll be satisfied.








  • Running strange software grabbed from unknown sources will never not be a risky proposition.

    Uploading the .exe you just grabbed to virustotal and getting the all clear can indicate two very different things: It’s either actually safe, or it hasn’t yet been detected as malware.

    You should expect that malware writers had already uploaded some variant of their work to virustotal before seeding it to ensure maximum impact.
    Getting happy results from virustotal could simply mean the malware author simply tweaked their work until they saw those same results.

    Notice I said “yet” above. Malware tends to eventually get flagged as such, even when it has a headstart of not being recognized correctly.
    You can use that to somewhat lower the odds of getting infected, by waiting. Don’t grab the latest crack that just dropped for the hottest game or whatever.
    Wait a few weeks. Let other people get infected first and have antiviruses DBs recognize a new malware. Then maybe give it a shot.

    And of course, the notion that keygens will often be flagged as “bad” software by unhelpful antivirus just further muddies the waters since it teaches you to ignore or altogether disable your antivirus in one of the most risky situation you’ll put yourself into.

    Let’s be clear: There’s nothing safe about any of this, and if you do this on a computer that has access to anything you wouldn’t want to lose, you are living dangerously indeed.


  • There are a near infinity of those out there, many of which just grab other scanlation groups’ output and slap their ads on top of it.

    Mangadex is generally my happy place, but you’ll have to wander out and about for various specific mangas.

    Several of the groups that post on Mangadex also have their own website and you may find more stuff there.

    For example right now I’ve landed on asurascans.com, which has a bunch of Korean and Chinese long strips, with generally good quality translations.

    The usual sticky points with all those manga sites is the ability to track where you are in a series and continue where you left off when new chapters are posted.
    Even Mangadex struggles with that, their “Updates” page is the closest thing they have to doing that and it’s still not very good.

    If you’re going to stick to one site for any length of time, and you happen to be comfortable with userscripts, Id’ suggest you head over to greasyfork.org, search for the manga domain you’re using, and look for scripts that might improve your binging experience there.



  • itsnotlupus@lemmy.worldtoaww@lemmy.worldDiscussion: AI Posts
    link
    fedilink
    English
    arrow-up
    0
    arrow-down
    1
    ·
    1 year ago

    You can either go full bonzai and aggressively trim any little branch that pokes out of place to try to keep a meticulously maintained tree at all times, or you can just let the tree grows as it will, and if a branch becomes an obvious issue, then just cut the entire branch and graft it somewhere else.

    If I was a mod here, I would do the latter, maybe even setup an /c/AIAww or whatever in anticipation for what might come.
    That’s probably my laziness speaking.


  • There have been efforts to build reputation systems that don’t rely on central servers, like early day bitcoin’s Web of Trust, which allowed folks to rate other folks with public key crypto, thus ensuring an accurate and fair trust rating for participants, without the possibility of a middle-man putting their thumb on the scale.

    One problem with it is that it was still perfectly practical for bad actors to accumulate good ratings, then cash out their hard-earned reputation into large scams, such as the “Bitcoin Savings & Trust” (for $40 million in that particular case), which quite possibly made it measurably worse than not having a system that induced participants into making faulty judgments in the first place.

    I think the main practical value of something like reddit’s karma is an indication of age and account activity, both of which can probably be measured in other, if less gamified ways.