• 0 Posts
  • 21 Comments
Joined 3 years ago
cake
Cake day: July 31st, 2023

help-circle
  • And to lean even more into that, now that AI is genuinely cracked at discovering vulnerabilities, security by obscurity is way less effective.

    Funnily enough this means now the effectiveness of crowd sourced security is actually the only remaining factor for security, so that would make open source software the safer variant. So in a way, AI not only open sources everything technically, it also makes open source software the more secure variant regardless.


  • It doesn’t. I wouldn’t say I made a generalization but a very high level explanation so it’s easier to understand for non-tech people.

    I suspect modern models are trained specifically on these kinds of questions to spare the embarrassment of this question for future models.

    that does mean they stochastically understand the occurrence of other numbers for this question and if you wrote tokens similar to this it would answer with 4 or whatever the specific case. I’m just trying to explain that no matter whether it gets it right or wrong, it goes off of statistics and semantics, not math. Because these things are not made to do math, that’s what algorithms are for, and we know very well how to write them.


  • hoshikarakitaridia@lemmy.worldtoLocalLLaMA@sh.itjust.works•Qwen27b vs GPT 6
    link
    fedilink
    English
    arrow-up
    7
    arrow-down
    1
    ·
    20 days ago

    Kind of but not really. AIs learn to predict text but they can’t really do math the traditional way. They learn the tokens, which are more than syllables and less than a word long. So because they don’t do math, LLMs in particular speculate “something like 3 comes after this question” because in training it has seen similar questions answered with “3”. That’s the jist of it.



  • Let me answer what happens currently:

    When a person is dead, trust & wills happens. Either someone is explicitly given ownership of the estate (often it is split between children) but if not, the estate forms it’s own legal entity that operates to the best understanding of the decedents wishes.

    With the accounts it’s a bit more complicated but generally from what I know a lot of platforms offer to reassign an account if a death certificate is provided and a family member has come forward to request control. However in practice I would assume this only happens to the accounts that family members are aware of, and the rest of them basically become stale.

    With regards to privacy, I would assume it is preserved for the most part. If relevant details about the decedent come up for legal reasons there might be some good cause to breach these but I can’t imagine all person information gets revealed just because. Though I’m also not a lawyer, so take all of this with a grain of salt.










  • MCP is a really dumb server that just takes requests and sends back data.

    Somewhere in your frontend settings (Open-WebUI has it under “integrations” I believe, some backends can set these things too) you see a section called “MCP server”. You need to keep your MCP server running somewhere (it’s very slim, so not a huge deal usually) and then in those settings you put your IP and port for the server, for example “http://192.168.1.15:8080/”. Usually if you did it right (sometimes you can check if it works in the settings), you can ask the model if it has access to the tools and it will tell you.

    What’s happening in the background is that every conversation the AI does, it gets a short preamble with some context about what it can do, and if you connect web search, custom skills or an MCP server (it’s just skills but they are run through their own server basically), the AI sees all that in this preamble if it’s set up correctly. And then when it feels like it needs to use those tools, it will check the description on how to do that, and then it uses them.


  • 1070ti and some hexacore something cpu that’s alright with around 24gb ddr4, I think one of the slower speeds. I also got a faster PC with another llama.cpp instance and I vibe coded a quick and dirty proxy which I can use to switch between the different backends. It’s genuinely useful because I got code editors but also my Open-WebUI using it.

    Sounds elaborate but generally it’s all pretty primitive.

    Also fyi long preloading steps because my vram is obv not good enough to hold the full model and kv cache and stuff, generation at about 15t/s tops.

    My experience is right now we’re at the stage where it’s a good assistant but not great for actually taking over bigger tasks completely.

    Am looking forward to:

    • experimenting with existing non-uniform quants; I believe those quants are the next step to more bang for your buck on those low-grade consumer rigs
    • hoping for new kinds of quants to come out and be more widely adopted
    • maybe a more universal or consistent harness that is less prone to broken tool-calls etc.
    • better models, specifically a new version for 35b a3b.