This wasn't just capable last year, it was capable prior to ChatGPT existing entirely through specialized automation harnesses that predate the terminology of a "harness" for models. This is something you could have done in 2018 if you had a spare couple of billion dollars of compute to waste. It was a subject discussed at security conferences even before that.
Automated dynamic exploit chains and credential discovery has been something every red team worth their salt does at a lesser scale for 20 years, why would you ever think that the capability didn't exist until it only cost a few thousand dollars to do?
Existed does not mean was easy, or was cheap, or was thought of as reasonable to attempt.
Also, IQ is not a term that applies to language models, and the term you are looking for is "Long Horizon Coherence".
Your meaning was not clear, which was why I made certain to explain why your two comments do not mean the same thing.
If that was not enough to explain why what you said was incredibly inaccurate and toes the line of fascist propaganda, I would be happy to have an amicable discussion with you one on one - as comment threads on HN only go so deep.
Your phrasing in the post I replied to is not the phrasing you are using now.
People frequently do bad things to people that they think are the right thing and that they still know are not “for [the victim’s] own good”. For example, removing an entire group of people from society is often presented as “for the good of everyone else”.
This is an incredibly important distinction when talking about the horrific ways gay people have been treated in the past couple centuries.
Coping with what? Your reply is too full of buzzwords and vague assumptions to understand what point you are trying to make or what you believe you have inferred from me.
Why on earth do you think an LLM would be capable of detecting problems with a listing like that without a human doing so first? LLMs are just as gameable as SEO, if not more so.
The "steering assist" feature in new cars is terrifying, and has almost severely injured or killed me once.
When my car was last receiving service, I was given a loaner vehicle with this new feature. I was driving home in the rain, and an aggressive driver passed me very closely, triggering the collision avoidance system - this forced the steering wheel to turn away from the car, made the car hydroplane and nearly flung me into a ditch at 55MPH. I had to force back against the steering and creatively spin my vehicle against the turn.
Anything you might want to do with a radio or IR device but don’t have specialized hardware for. It’s kind of a swiss knife/leatherman tool for short range communications standards.
Showing what you have done to help yourself is by far the most effective way to get help in my experience - not only does it show that you aren’t just helpless and will be a drain on resources, it also gives the person helping you a jumping-off point, avoiding potential troubleshooting and back-and-forth questions.
I make sure to do this at work, when I’m asking questions about things outside of my domain. I do it when talking to tech support for my ISP or phone, when I contact maintenance for my apartment and when I take my car in for service.
> myself, i think in terms of a mono/dialogue if i am preparing to disseminate.
> if i am thinking for my own benefit i percieve it as idetic imagery and run thought experiments to qualify for IRL candidacy.
I’m very similar personally, though I often have background dialogues going over “sub threads” of whatever I’m thinking about.
I have also wondered if people who say they don’t have this kind of thought process actually do - but the real difference is that for some reason it persists in memory for us, whereas it is immediately discarded for them, leaving only the result.
Animals don’t have language the same way we do, but many of them do have some kind of semi-structured communication. Cats for example, when domesticated, develop an entirely parallel method of communication specifically to interact with humans - one that is neotenous in genesis, but nonetheless absent from their feral counterparts.
This is in the same vein to what I wanted to say here as well - I find it very likely that people with no internal monologue are executing similar mental processes, just without them being so “front and center” in the mind.
Automated dynamic exploit chains and credential discovery has been something every red team worth their salt does at a lesser scale for 20 years, why would you ever think that the capability didn't exist until it only cost a few thousand dollars to do?
Existed does not mean was easy, or was cheap, or was thought of as reasonable to attempt.
Also, IQ is not a term that applies to language models, and the term you are looking for is "Long Horizon Coherence".