331–340 of 1001 posts
Breaking: the runners lagging behind in the AI race demand openAI to stop right on its tracks; this is very unsafe; it will be much safer once we catch up and openAI is on the rear view mirror.
“AI will probably most likely lead to the end of the world, but in the meantime, there'll be great companies” - Sam Altman But sure, let’s just let this guy who sounds like a madman keep going. What could possibly go wrong?
I'm really disappointed to see the HN reaction of straight up scoffing at this like it's irreversible and so let's just step on the gas. It's some weird combination of techno-utopianism and nihilism and good old elitism.
It really is irreversible. The methods and compute needed to build a GPT-type AI are widely available.
The idea that "x is technically possible, therefore it is inevitable" - the favored line of the tech oligarchs who are the very ones directly furthering x - is sadly cargo culted by many rank-and-file technologists. If you were to apply the same defective reasoning to, say, nuclear weapons, then we would all be dead by now.
It is possible for humans to agree that something should not be done, and prevent it from being done. This is especially true of these LLMs, which require ingesting (and ignoring copyright on, by the way) massive amounts of human work and then spending massive amounts of computation to process.
That we could choose to do something about this should not be controversial, regardless of what the driving AI.
GPT-4 might be close to the best we'll get on the general LLM model front for a while since they trained on a huge chunk of web text. Next real advances will probably be in tuning them for specific applications in medicine, law, accounting, marketing, coding and etc. As someone running a one man company I can't wait for the cost of accounting, legal and copywriting to approach 0. Cost of shipping products will also g…
It's wishful thinking that somehow that goes to 0.
Earlier quoted context omitted.
yes so let's not bother trying to stop an existential threat. have you considered the possible consequences of all this?
Very hard to say how threatening AI really is. We need more research to determine that. Even if it is an existential threat, a moratorium will not make it go away. That's why we still have nukes.
what we have is a continuously improving system that can write code and solve problems with intent, and is starting to have access to the internet. I don't think AI is inherently evil or will do anything harmful on its own, but it is a tool that humans will use to harm each other, deliberately or otherwise.
if you need research, go and read some of the myriad science fiction on the topic, or just imagine an "intention bot", where you set up an LLM with a goal, then have it iteratively write and execute programs to attain that goal.
also there undoubtedly is research on the topic, but I don't need to read it to know that completely controllable brains with access to the internet, that don't get tired or bored or have to sleep or eat are an extremely dangerous thing
Earlier quoted context omitted.
if governments across the West came out now and made laws to stop this, then sure a few thousand experts on hacker news that can afford to drop $20k on hardware could keep developing them illegally, but all the universities and corporations would stop. and it's the universities and the corporations, openAI especially, that have capacity to develop the dangerous AIs
Then the West would fall behind other countries as well. It would be how like the US tried to ban stem cell research under Bush.
Earlier quoted context omitted.
I'm not concerned about how they work or whether they're attaining personhood or not. I'm concerned at what people can and will use them for to the detriment of our civilisation
How they work is the most critical piece of that puzzle. It isn't a mystery at all. LLMs implicitly model every pattern they can find in the text they are given. If the pattern aligns to language, humans categorize it as a feature. If it aligns to desired behavior, humans categorize it as a feature. If it aligns to undesired behavior, humans categorize it as a limitation. What humans categorize as features and limita…
the danger here is that we make extremely powerful tools extremely available and people use those tools to hurt people, possibly a lot of people, possibly even by accident
the scary part is that hacker news is full of the exact people that have the knowledge and capability to do this, and look at the reaction to this post: a resounding fuck off
The cat is out of the bag. Some people might volunteer to slow down. But others inevitably won't. And that creates this sense of FOMO with everyone. I think the reason OpenAI is choosing to operate in the open has been exactly this. This way they get to control the narrative, act as a first mover in the market, and ensure it is done right (by their notion of this). Of course others are working on the same topics and…
As GPT-4 puts it "the confetti has left the cannon."
we've worked hard to accelerate both global warming and nukes, so AI is no different - yet.
"The Dark Forest" is a science fiction novel by Liu Cixin, which introduces a concept called "suspicion."(“猜疑链”) In this concept, the interaction between civilizations is viewed as an ongoing game of suspicion, where each civilization is worried that the other civilization may pose a threat to it. This kind of suspicion makes it difficult for civilizations to establish trust, thereby making the universe a dark forest…