• 0 Posts
  • 13 Comments
Joined 1 year ago
Cake day: April 23rd, 2023
  • Onihikage@beehaw.orgtoPrivacy@lemmy.ml•Routing all traffic through Mullvad?English
    8·
    1 month ago

    Router-level VPN is going to be more difficult to configure and cause more problems than just having it on all your devices. There are some games where online play just refuses to work if connecting through a VPN. Some mobile apps are the same. When a website blocks your currently selected server, and the usual solution is switching to another server, that’s going to be more difficult and more tedious when it’s configured at the router level. In addition, if you do something like using a self-hosted VPN in order to connect remotely to a media server on your home network, that becomes more difficult if your home router is on a different VPN.

    If you’re trying to keep local devices in the building from phoning home and being tracked, a PiHole or router-level firewall might be a better solution. I think if you’re running a pfsense or opnsense router and are a dab hand with VLANs then maybe you could get what you’re looking for with router-level VPN, but it’s a huge hassle otherwise. Just put Mullvad on your computers and phones and call it a day.

  • Onihikage@beehaw.orgtoLinux@lemmy.ml•Alpaca Simplifies Running Advanced AI Language Models on LinuxEnglish
    2·
    1 month ago

    Unfortunately I can’t even test Llama 3.1 in Alpaca because it refuses to download, showing some error message with the important bits cut off.

    That said, the Alpaca download interface seems much more robust, allowing me to select a model and then select any version of it for download, not just apparently picking whatever version it thinks I should use. That’s an improvement for sure. On GPT4All I basically have to download the model manually if I want one that’s not the default, and when I do that there’s a decent chance it doesn’t run on GPU.

    However, GPT4All allows me to plainly see how I can edit the system prompt and many other parameters the model is run with, and even configure multiple sets of parameters for the same model. That allows me to effectively pre-configure a model in much more creative ways, such as programming it to be a specific character with a specific background and mindset. I can get the Mistral model from earlier to act like anything from a very curt and emotionally neutral virtual intelligence named Jarvis to a grumpy fantasy monster whose behavior is transcribed by a narrator. GPT4All can even present an API endpoint to localhost for other programs to use.

    Alpaca seems to have some degree of model customization, but I can’t tell how well it compares, probably because I’m not familiar with using ollama and I don’t feel like tinkering with it since it doesn’t want to use my GPU. The one thing I can see that’s better in it is the use of multiple models at the same time; right now GPT4All will unload one model before it loads another.

  • Onihikage@beehaw.orgtoLinux@lemmy.ml•Alpaca Simplifies Running Advanced AI Language Models on LinuxEnglish
    3·
    1 month ago

    I have a fairly substantial 16gb AMD GPU, and when I load in Llama 3.1 8B Instruct 128k (Q4_0), it gives me about 12 tokens per second. That’s reasonably fast enough for me, but only 50% faster than CPU (which I test by loading mlabonne’s abliterated Q4_K_M version, which runs on CPU in GPT4All, though I have no idea if that’s actually meant to be comparable in performance).

    Then I load in Nous Hermes 2 Mistral 7B DPO (also Q4_0) and it blazes through at 50+ tokens per second. So I don’t really know what’s going on there. Seems like performance varies a lot from model to model, but I don’t know enough to speculate why. I can’t even try Gemma2 models, GPT4All just crashes with them. I should probably test Alpaca to see if these perform any different there…

  • Onihikage@beehaw.orgtoLinux Gaming@lemmy.ml•Hardware for Linux gaming PCEnglish
    1·
    1 month ago

    PCIe gen 5 is for the PCIe slots and NVMe storage slots, but they’re backwards compatible; you can put a gen 3 component in a gen 5 slot and it will work at gen 3 speeds. Similarly, if you put a gen 5 component in a gen 4 slot, it will be limited to gen 4 speeds. Right now there’s very little appreciable difference between gen 4 and gen 5 unless you’re spending a lot of money on the component (GPU/storage). Another thing to note is that Gen 5 requires that both the CPU and motherboard support it; a CPU with gen 4 support in a gen 5 motherboard will limit all the slots to gen 4 speeds.

    RAM is a totally different standard that must be matched exactly for what the motherboard has; if it’s a DDR5 motherboard then you have to use DDR5 RAM or it won’t even fit in the slots. You can get a PCIe gen 5 motherboard and just use gen 4 SSDs or GPUs, that’s perfectly fine and leaves you room to upgrade later.

  • Onihikage@beehaw.orgtoLinux Gaming@lemmy.ml•Hardware for Linux gaming PCEnglish
    1·
    1 month ago

    Seems mostly fine to me, I game all the time on Linux (Bazzite gang 🤘) with a 3900X + 7900GRE, haven’t had any significant issues aside from needing to make sure clock speeds were configured correctly on the GPU. Two ram sticks is the way to go with these systems as sometimes they don’t support 4 sticks at full speed.

    You’re right that GPU passthrough is definitely more for tinkering or advanced users with very specific needs (usually professionals who need Windows/Nvidia and choose to run it in a VM rather than dual-boot), with a budget to match. For a gamer couple, having fully separate systems is going to be much less hassle and more resilient against failure.

    The one thing I would recommend changing is the power supply, it’s unironically the most important component in the computer because if it fails it can kill everything else, and the System Power 10 is known enough for being low-quality that discussions of that come up in web searches. Poor quality power supplies can damage your hardware and otherwise cause weird, intermittent issues even if everything seems to work fine most of the time, and will fail and shut off the computer when a good power supply would have just kept on chugging. Seasonic and Corsair are considered the best brands and have 10 year warranties - they’re more expensive, but they’re worth it. You want 80+ Gold or better these days, this is a buy once, cry once component.

    If you don’t have a UPS, I would also recommend getting one at some point, either one big shared unit (if they’ll be close together) or two individual units. Having backup power will allow you to shut down the computers gracefully during a power outage, and prevents the worst-case scenario where the power goes out while the computer is installing updates and it turns into a brick.

  • Onihikage@beehaw.orgtoTechnology@beehaw.org•What if Everyone Knew Who You Really Were on the InternetEnglish
    1·
    1 year ago

    Everyone knowing your identity? The drawbacks would far outweigh the benefits. However, there may be a path to the benefits of a Real ID sign-up system that mitigates the possible harms.

    First of all, let’s get this out of the way - this “minimal harm” approach would only be feasible if the government could either reach some level of technical competency or farm out the task to heavily restricted private corporations that do have that competence. If we presume that’s the case (unlikely), the question becomes whether the people would be willing to accept it. If we presume the majority of citizens also want such a thing (a tall order to be sure, I certainly don’t want it), then the question becomes what sort of system would be able to maximize privacy, and thus safety, while still requiring your real identity to be involved in creating online accounts? What would that system look like?

    (Collapsed for your convenience because I wrote way too much about this hypothetical)

    We’d absolutely need a level of abstraction. The government knows who you are anyway, but the business entity you’re interfacing with would get a unique token from the government that is not your actual Real ID number but which is a hash generated from the business’s (salted) ID number and your own salted ID number (idk I’m not a cryptographer).

    Signing up for an account would resemble using Google or Facebook to create an account; you’d be redirected to some third party Identity Verification System (IVS) which would handle identity verification and redirect you back to the account creation with the extra piece of information provided by the third party. You’d still pick a username, password, etc.; the government database would only be used to generate that unique token.

    More specifically, the website or service would only be passed a token from the IVS, uniquely generated based on the company ID and the person’s ID, and the government database would only keep the token, not any of the data used to compute it. (That’s not counting China and other authoritarian states, of course - they’d definitely retain all that information and have a list of all the sites you have accounts with. This wouldn’t solve that problem.) This would make the IVS database virtually useless on its own, as an attacker who compromises the database has no way of knowing which token is associated with which website, and cannot derive it themselves unless they’ve also compromised one or more target websites at the same time. The cryptographic stuff would be rotated once it’s known that a breach has occurred, so such breaches would likely be limited to state actors or black-hat groups that hoard zero-days.

    Now, what would all this accomplish? What would it make possible that currently isn’t outside of China?

    • Unique website signups - one person, one account, and if it’s banned, that’s it, you don’t get to log in to that site ever again until you’re unbanned. Your only option to get around a ban would be to commit identity fraud, which would be quickly traced back to you if everything really was using this system.
    • If you block someone, they can’t just make a new account and keep harassing you; they’d have to start committing crimes, and the pattern of behavior would be easily traced back to their original account, and with it, their original identity.
    • No more sock puppets. If you say something on a platform, you only get one account to say it with. Troll farms would have to openly pay thousands of people to support a particular view, which many websites would likely consider a bannable offense. Troll farms are non-viable.
    • A website doesn’t need your email address or any personal information from you in order to verify your identity for password resets. If the IVS returns the correct token, that’s good enough.
    • If a user has committed a crime, and evidence of this is visible on a website or platform, a government with jurisdiction can, with a warrant, request that user’s token. That gives them a specific identity in the ID database to investigate further.
    • If the government is investigating a particular individual over whom they have jurisdiction, they can query websites or businesses over which they also have jurisdiction for information on whether any of the tokens in their database match a user account’s identity token, and request data from the matching account. It would be a much more focused process than queries based on IP addresses which judges keep having to say are not proof of identity.

    What would this system not do? What doesn’t change compared to now?

    • Companies using this system would still only know for sure who you are if you tell them; at most, they know with certainty what country your identity is associated with, but little more.
    • Companies could still coordinate information on data such as which accounts sign in from the same IP addresses, which would tell them more about specific users and potentially let them profile you.
    • Companies will still give up any information they have on you to the government if compelled by a warrant, sometimes even without one.
    • Websites can be hacked and your data on that website exposed to the world, requiring you to reset your password, etc.
    • The government can be hacked and information about your identity exposed
    • Accounts can be hacked, and nefarious people can do nefarious things under your name without having to commit identity fraud (though this act could itself be considered a crime under such a system)
    • Stalkers can still figure out who you are based on information you post, and go after you in the real world
    • The government doesn’t know which websites you visit unless they’re actively spying on you.
    • Oppressive governments can and will continue to monitor and log everything they can about you, and attempt to weaponize this against dissenters or those otherwise deemed “undesirable”

    Even in the grandest, best-possible-case scenario I can think of, it still comes down to “Can I trust my government to not take more information than they’re allowed to, and can I trust that they will not abuse the information they do obtain?” For many, I suspect the answer to both questions is no.