Carl-Richard

@StaraX But have you tried brain training?

The way I predict the next OpenAI model names: astra -> nova -> black hole (singularity; i.e. we've reached AI singularity, probably not under the LLM moniker). Might as well add in neutron star there somewhere.

So you're saying if somebody wants to donate to malaria prevention, they should also post about it on youtube and monetize it so that they can get a ROI and so they want to do it more (unless they feel better to just donate without getting anything in return)? 😆

I think my answer to this ethical altruism conundrum is that we don't just value the sheer quantity of human life or even alleviating human suffering, but we value things like music, art, and we need different people to do different things to make society go around. We need air traffic controllers, train station workers, nurses, physicians, academics, politicians, etc. Even people who suffer, value music, art; they might go to those things to alleviate their suffering, and if you take that away from them but you assure their life through Malaria prevention, you might have made their life less worth living. So what to do?: Do what you feel called to do. A person who feels called to do something, who feels motivated to do it, will do it better than somebody else who does not feel as motivated. And that way you will contribute greater to that part of the world you choose to contribute to. If you feel very called to donate money to Malaria prevention programs, do that. If you feel very called to find a cure for cancer, do that. Charity is a way of using your time on something you think is valueable. So use your time on something you think is valueable. "But can't you donate like 5% of your savings to Malaria prevention and save a big chunk of lives that way and still do what you feel called to do?" Well, yes, but what about other programs like water sanitation, cancer prevention, other disease prevention? "Just choose one that has a high life-saving conversion rate". Well, why do you ought to use 5% of your savings on initiatives with high immediate life-saving conversion rates and not say finding a cure for cancer or just landing a recording contract?

I know you just used Puke GPT to write that because that was absolutely a load of horseshit. Imagine copypasting my post into AI and copypasting back its answer to me when I asked you to merely clarify your answer, instead of merely clarifying your answer. That explains the absolutely horseshitted AI response you just got. You get out of it what you put into it. The fact of the matter is you made a ridiculous typo and you also bombed the "them" referent and you can't just accept it and clarify so everything is clear. Tell your AI that I asked whether it's worth it to spend 5% of your income on malaria prevention OR that 5% on your art/career/calling, not that they should abandon their art/career/calling and then donate 5% of their income (uh, what?). Tell your AI that it sucks at reading comprehension and should be scrapped or 300x its token usage. And tell yourself that you're lazy and don't deserve to talk to other people in this thread if you can't even clearly state what you are saying when asked to do so.

The context problem is mainly because we give a tiny little prompt and the AI is supposed to deduce what we mean based on that (which is hard because it has to know a lot about us and our goals that we haven't communicated to it). A smarter model or more tokens only fixes this marginally.

@OBEler Man you have the conversational understanding of 4o. I talked about the progression of LLMs, how far we've come. I also talked about 5.6 Sol. I haven't tried Astra much so I can't tell you what I think, but I think the context argument holds for Astra as well (it's a much deeper problem than upping the performance a few percent on some reasoning benchmarks). And if you want to talk about understanding technology, if you think current LLMs are AGI, you have a low bar for what AGI is. Literal "AGI" means we are now literally all out of work. Yet we're not.

Oui oui baguette.

That is impossible. The world works only according to the outdated 17th century Newtonian atomistic-mechanistic worldview.

LLMs at this point are supposedly like hyper-nerds: good for book knowledge, but that's about it. I remember when people were gushing over 4o two years ago like it was some sort of genius when it was actually a high-schooler with stifled speech, and I thought it was retarded back then. I haven't tried Astra much yet, but supposedly it should be really really smart. However, I've tried 5.6 Sol a bunch for producing text, and it still has a problem of not understanding context well (it says a lot of things you would usually find irrelevant even if it is technically valid). Perhaps you could fix some of it with better prompting, but context is such a subtle thing that I believe humans will still have an edge on that for a very long time.

@yetineti "It seems, although [...]" is a fucked up typo for "it seems as though", with the comma and everything; extremely specific and multi-step typo. Besides, your referent "them" is highly ambiguous (are you talking about the curer or the cured?). I would kindly request you to try again.

The concept of a male escort, especially a high-rated one, is really fascinating. Notice the personality type. Notice the social status, the philosophy and values, the type of interactions that women want out of it. They (the good ones) seem like highly caring, sensitive heroes.

With enlightenment, there is nobody to be happy.

Y'all be practicing non-dual self-harm. You're only harming yourself. You're treating your brain like a pot of stew, stiring it into a homogeneous Oneness.

This video reminded me of this thread. Seems like a similar kind of person as the male escort above:

Somebody has to write the name and date in the AI execution prompt. Somebody has to lead the prisoner to the execution chamber. Unless you're in the singularity, in which case probably most crime and social ills will evaporate anyway (either benevolently or violently).

Feelings and intuition are counted in "functional". It's just the equation is also larger. There is that and then there is the "feeling" of living in a society that you know can take your life if it so decides (is that a good feeling or not?). And the fact that executioners suffer terrible rates of PTSD. There is also an ironic catch-22: the only people you would want as executioners would be extreme psychopaths or serial killers (people who take pleasure in or are completely fine with killing people), but you'll of course end up executing a lot of those people instead of giving them a job.

@Basman Genuine question: Do you think a sense of retribution for the victims is necessary for a functional society? I discussed this with a youth politician once, and he seemed to hint that it was, probably while hinting to Breivik.

I think the question should be what creates a functional society, not necessarily whose feelings are hurt or not (although that could be a factor).

Being stripped of autonomy is arguably the greatest punishment you could get (which is what jail is). Imagine you're just living your life and one day the government tells you to move to a different country and you are not allowed to leave. People love travelling/moving to other countries, when they are the ones deciding it. People also love solitary confinement when they are the ones deciding it — it's called meditation.

Are Norwegian jails punishment?

Describe to me what a random benchmark does to the best of your ability and I'll bet that changes absolutely nothing.

Me reading a benchmark result is like a third grader reading a tax return. Wtf am I supposed to gather from that? Yeah ok we see a graph and different models scoring differently on that graph. Ok: what does the graph actually mean? I have no idea. What does the scale of the graph entail (other than "it did x% success")? No idea. You could Rick-and-Morty it out for me and it would be just as insightful: they scored 100% dinglybops while the other models scored 55% squirblybops. That means the next models will probably score 55% sqooblebops on the sqoodle test.

"Thanks to psychedelics" would have to be qualified. But Ram Dass was famous for psychedelics and then he travelled to India and later became what people would consider a spiritual teacher. My first LSD trip was a lesson of dukkha in Buddhism (which was before I had started getting into spirituality). I remember I went out of that trip with a kind of a quest: "I need to fix this thing". It really foreshadowed all my spiritual engagement after that.

Exactly. Meanwhile Abdul El-Sayed, a Muslim American, is running for Senate and is pro gay marriage, NYC's mayor Mamdani is a Muslim and pro gay marriage, there are other Muslim Americans I know about who accept gay people (and who don't accept them, just like with Christians). I even know someone from the US who practiced Islam in Lebanon who is a feminist super lefty.