Artificial neural networks are making strides towards consciousness?
1–10 of 85 posts
Re: Artificial neural networks are making strides towards consciousness?
#2Ugh. You can’t say “strides towards consciousness” until you define consciousness AND have a metric that can provide objective measure of such things.
Re: Artificial neural networks are making strides towards consciousness?
#3>In our conversation, lamda tells me what it believes Ramesh felt that Lucy learned about what Mateo thought about Lucy’s overture. This is high order social modelling.
I am not sure the example convinces me that what GPT-3 is doing here is 'modeling' the agents in the story. It feels like the author is trying to draw a contrast between rote statistical language modeling and higher-order understanding, and it's not clear to me that the result shown could not be achieved with the former. I'd like to see a bit more skeptical probing of this. If GPT-3 has a model of the story and its characters, how consistent are its answers to questions about that model? Is it distracted by irrelevant sentences?
In the counterpoint article, they give example questions like:
>Dave & Doug: What’s the world record for walking across the English Channel?
>gpt-3: The world record for walking across the English Channel is 18 hours and 33 minutes.
Is this the output of an agent which has successfully modeled the scene?
Re: Artificial neural networks are making strides towards consciousness?
#4Ugh. You can’t say “strides towards consciousness” until you define consciousness AND have a metric that can provide objective measure of such things.
Which I wonder if we'll ever have as consciousness is a strictly subjective thing.
Re: Artificial neural networks are making strides towards consciousness?
#5No. No they're not.
Re: Artificial neural networks are making strides towards consciousness?
#6Any links for the full article ?
Re: Artificial neural networks are making strides towards consciousness?
#7Queue the hacker news haters. As someone that works with neural networks on a daily basis, I am constantly stunned by the effectiveness of modern attention based NLP models. They actually work for real world tasks, and you don't need to be an expert in them to apply them. Not only that, the newer ones are drastically better than the ones only two years ago. These are the models released in public.
The sheer amount of capitalist interest in getting these models to improve almost dictates that we will see incredible improvements over the next decade. Facebook and Google, with their war-chests of money, have massive business models that directly profit from AI/NLP based technology. I'll repeat, the best business on earth (google) directly profits from these models, and thus will hunt and find the best human talent to develop them.
What's occurring is that the thoughts of billions of people are being poured onto the internet, that data is the prerequisite we needed for AI. AI wouldnt be possible without this convergence of data in text and video form, cheap large scale compute, and the capitalist forces explained above
Re: Artificial neural networks are making strides towards consciousness?
#8It is double for me.
On the one hand, such a title is ridiculous and shameful and silly, especially when involving popular science; now the common public interested in AI starts to talk about Skynet.
On the other hand, this reads like a title which could have been written by a language model itself. So we are making strides at something. But seems humans are making strides towards acting like language models, not the other way around.
Re: Artificial neural networks are making strides towards consciousness?
#9Queue the hacker news haters. As someone that works with neural networks on a daily basis, I am constantly stunned by the effectiveness of modern attention based NLP models. They actually work for real world tasks, and you don't need to be an expert in them to apply them. Not only that, the newer ones are drastically better than the ones only two years ago. These are the models released in public. The sheer amount of…
The question being asked isn't "are language models really good?", though. It's "are language models (moving towards being) conscious?".
Re: Artificial neural networks are making strides towards consciousness?
#10Queue the hacker news haters. As someone that works with neural networks on a daily basis, I am constantly stunned by the effectiveness of modern attention based NLP models. They actually work for real world tasks, and you don't need to be an expert in them to apply them. Not only that, the newer ones are drastically better than the ones only two years ago. These are the models released in public. The sheer amount of…
No one is arguing that these models are not useful, but claiming they are somehow conscious or even remotely equivalent to human reasoning is just a bunch of BS. Especially for something like consciousness, where humans themselves can't even agree on definition!