• 1 Post
  • 464 Comments
Joined 1 year ago
cake
Cake day: October 4th, 2023

help-circle



  • Tales-like

    I’ve been kind of out of the RPG loop for a while, probably not the best person to suggest, and haven’t played the series, but I’m thinking that if you could expand a bit on that, it might help provide suggestions…I mean, not clear to me what you’re looking for that’s specific to that relative to other RPGs. Similar setting? A long-running RPG series with many entries? The combat system (absent the real-time aspect)?

    You mention “depth of story”, so maybe something with a similar level of storytelling?



  • I don’t know whether Altman or the board is better from a leadership standpoint, but I don’t think that it makes sense to rely on boards to avoid existential dangers for humanity. A board runs one company. If that board takes action that is a good move in terms of an existential risk for humanity but disadvantageous to the company, they’ll tend to be outcompeted by and replaced by those who do not. Anyone doing that has to be in a position to span multiple companies. I doubt that market regulators in a single market could do it, even – that’s getting into international treaty territory.

    The only way in which a board is going to be able to effectively do that is if one company, theirs, effectively has a monopoly on all AI development that could pose a risk.






  • using an admin portal’s default credentials on an IBM AIX server.

    I think that there are two ways to solve that.

    The first is to have the admins actually complete setups.

    But, humans being humans, maybe the second is a better approach:

    When creating a computer system, don’t let a system be used, at all, until all default credentials have been replaced with real ones. If you do, someone is invariably gonna screw it up.

    Your directions may say “Before pulling lever 2, pull lever 1 so that machine does not explode”. And maybe you feel that as the manufacturer, that’s covered your hind end; you can say that the user ignored your setup instructions if they get into trouble. But instead of doing that, maybe it’s better to not permit for a situation where the machine explodes in the first place; have pulling lever 2 also trigger lever 1.


  • If you’re talking about that VAE tiling feature or or Tiled Diffusion or whatever it’s called, I think that it shows up in the text below the image in A1111, and I think that anything that shows up there is also stored in the generated image’s comment metadata.

    I don’t normally use Tiled Diffusion, if that’s what you’re referring to, but let me see if I can go generate something with it and check.

    checks

    Yeah, text: Tiled Diffusion: {"Method": "MultiDiffusion", "Tile tile width": 96, "Tile tile height": 96, "Tile Overlap": 48, "Tile batch size": 4}, gets added to the text below the image and to the image metadata.

    That being said, I don’t know how far I’d trust the image metadata for reproducibility if this is a hard requirement you’re looking for. I have definitely seen various settings that mention that they induce non-deterministic behavior, and I’m not sure that all of those are encoded in the metadata. Also, while the version (and looks like git hash of built version) is encoded, I’m sure that not everyone is using the same version, and I don’t know what compatibility is like across versions.

    EDIT: For example, see the “Optimizations” here:

    https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Optimizations

    You have a bunch of A1111 command-line optimizations options that have descriptions like:

    --opt-sdp-attention May results in faster speeds than using xFormers on some systems but requires more VRAM. (non-deterministic)

    And those are not encoded in the image metadata, and that’ll make a given output non-reproducible.




  • tal@lemmy.todaytoSelfhosted@lemmy.worldSelfhosted chat service
    link
    fedilink
    English
    arrow-up
    2
    arrow-down
    1
    ·
    edit-2
    15 days ago

    I have already looked in XMPP, but it required SSL certs and I did not have the mood to configure them.

    There are definitely XMPP clients that do end-to-end encryption that do not rely on TLS for key exchange, though.

    https://en.wikipedia.org/wiki/Off_the_record_messaging

    Off-the-record Messaging (OTR) is a cryptographic protocol that provides encryption for instant messaging conversations. OTR uses a combination of AES symmetric-key algorithm with 128 bits key length, the Diffie–Hellman key exchange with 1536 bits group size, and the SHA-1 hash function. In addition to authentication and encryption, OTR provides forward secrecy and malleable encryption.

    The primary motivation behind the protocol was providing deniable authentication for the conversation participants while keeping conversations confidential, like a private conversation in real life, or off the record in journalism sourcing. This is in contrast with cryptography tools that produce output which can be later used as a verifiable record of the communication event and the identities of the participants. The initial introductory paper was named “Off-the-Record Communication, or, Why Not To Use PGP”.[1]

    I’ve used Pidgin with the libOTR plugin that implements that protocol.



  • wordfreq is not just concerned with formal printed words. It collected more conversational language usage from two sources in particular: Twitter and Reddit.

    Now Twitter is gone anyway, its public APIs have shut down,

    Reddit also stopped providing public data archives, and now they sell their archives at a price that only OpenAI will pay.

    There’s still the Fediverse.

    I mean, that doesn’t solve the LLM pollution problem, but…