• 3 Posts
  • 1.68K Comments
Joined 3 years ago
cake
Cake day: June 29th, 2023

help-circle
  • Excrubulent@slrpnk.nettoTechnology@lemmy.worldAn AI hate wave is here
    link
    fedilink
    English
    arrow-up
    1
    ·
    edit-2
    21 hours ago

    whilst it is adding some productivity

    Is it though? Like what’s the evidence of that? If it just feels like it must be true, I have some bad news about that:

    https://arstechnica.com/ai/2025/07/study-finds-ai-tools-made-open-source-software-developers-19-percent-slower/

    The most interesting part of this isn’t that it slowed them down when they expected to be faster, it’s that even after it slowed them down, they couldn’t tell and were fooled that they had been faster.

    Look at the graph, especially the last two lines:

    https://cdn.arstechnica.net/wp-content/uploads/2025/07/aicodingchart-1024x507.png

    My theory about this is that LLMs were tasked with giving useful output, but they couldn’t do that, because they have no fidelity, so instead they found a shortcut, which was to trick people into thinking they were being useful. They found the same loophole that conmen have used for millenia, and automated it. It’s the AI alignment problem, only for some reason people aren’t talking about it, maybe because they don’t want to believe that we’re this easily manipulated.

    There’s no reason to believe LLMs have gotten any better at actually doing useful work in the meantime in the absence of any objective measure of it. I think the best explanation for their “improvement” is that they have simply gotten better at fooling us.







  • Documents can be replaced. They get lost, stolen, or destroyed all the time. You can explain to a clerk that your documents were stolen - technically the truth since they’re withheld from you without your consent - and if they press the issue you can explain that your abusive parents won’t give you access to them and you cannot ask them for access.

    Think about how you frame the information - don’t say “my parents have them and won’t let me access them” because that sounds potentially benign to a bored clerk just trying to get through the day and not really paying attention. Say, “they were stolen by my abusive parents who I no longer live with,” since that front-loads the problem and frames you as the victim, rather than as someone’s child. If you can have a friend with you when you go to get your documents that can help. You haven’t said how old you are that I saw, so I’m assuming you’re still a teen, but even if you’re a young adult this can still matter. Okay, I see you said you’re an adult.

    In an ideal world it shouldn’t make a difference, but the way you present what is technically the same information really does matter in getting bureaucrats to help you properly. They are people, and they don’t just follow rigid rules, they will be swayed by their emotions and learning to navigate that is a big part of getting the system to work for you.

    Also documents like that usually have serial numbers. That’s so if they are stolen, they can be registered as invalid, so the thief can’t use them to steal your identity. So whatever is in the safe can be made worthless if you get that done. Getting replacements should automatically invalidate the old ones but not every system works the same, so double-check that the old copies will be invalidated.

    It depends a little bit on where you are but in general I wouldn’t trust the cops to be helpful, unless you somehow know for a fact that they will help and not just return you to your family. I hate to say that but they fundamentally exist to protect property and a lot of them accept society’s logic that children are the property of their parents, and if you’re striking out on your own it’s important for you to learn that cops aren’t your frends. People like your dad & brother become cops specifically because it gives them power over others.

    It’s also likely your parents will simply lie and try to convince the cops that you should be back with them. Not to say they will be successful, but once you’re away from the home I would absolutely try to eliminate any contact you make after that. I don’t want to scare you too much, but also these people have a pattern. They usually know how to talk to cops, since they tend to talk the same language. Your parents likely keep the documents away from you in order to keep you controlled, so they will know that this is an opportunity for them to reel you back in. I wouldn’t give them the chance.

    I would look up teen shelters, and if you can find a group of people who you believe have your interests at heart then you can ask them for help. All of this will be a lot easier if you can find allies. If you can find any mutual aid organisations near you - “food not bombs” is a common name to look for - they may know other orgs willing & able to help, or they may just have people who are willing themselves.















  • Yes, they try to prevent unwanted outputs with filters that prevent the LLM from seeing your input, not by teaching the LLM, because they can’t actually do that, it doesn’t truly learn.

    Hypotheticals and such work because the LLM has no capacity to understand context. The idea that “A is inside B”, on a conceptual level, is lost on them. So the idea that a recipe for napalm is the same whether it’s framed within a hypothetical or not is impossible for them to decode. To an LLM, just wrapping the same idea in a new context makes it seem like a different thing.

    They don’t have any capacity to self-censor, so telling them not to say something is like telling a human not to think of an elephant. It doesn’t work. You can tell a human not to speak about the elephant, because that’s guarded by our internal filter, but the LLM is more like our internal processes that operate before our filters go to work. There is no separation between “thought” and output (quotes around “thought” because they don’t actually think).

    Solving this problem I think means making a conscious agent, which means the people who make these things are incentivised to work towards something that might become conscious. People are already working on something called agentic AI which has an internal self-censor, and to my thinking that’s one of the steps towards a conscious model.