Regarding current ai tools augmenting human thinking, I wonder if that will continue given the agentic and autonomous reasoning capabilities being designed into the next set of frontier models.
I am not sure I believe human level AI will be augmenting for us as we would like to imagine.
There are a host of congnitive biases that promote AI being used to off land cognitively demanding tasks (automation bias, authority bias, system 1 thinking) that I think will be very hard to fight against.
I don't think AI is very good at automating human thought. Augmentation is way more promising than automation. All the use expert people do with AI is not to get rid of their more precious thoughts but to offload the boring/uninspiring tasks.
I agree the current AI models are not good at automating human thought, but they are getting pretty close imo, for cognitive valuable human thought. Although, High schoolers already seem to be using it to automate their work and teachers are grasping for ways to prevent from being fooled by it. I don't think limiting to how expert people use AI is a good assessment of how AI will impact society.
I agree augmentation is way more promising than automation, especially for knowledge workers. However, the market incentives at play and the desire for human level AI models seem to skew automation over augmentation, from the corporate side of things.
from Mustafa Suleyman, head of Microsoft AI "The tech CEO’s answer was this: “I think in the long term—over many decades—we have to think very hard about how we integrate these tools because, left completely to the market…these are fundamentally labor replacing tools.” https://gizmodo.com/deepmind-founder-ai-davos-mustafa-suleyman-openai-jobs-1851176340#
The economist perspective of AGI impact on human labor markets do not look ideal either.
Sounds like a very interesting panel session. Looking forward to watching it!
Regarding current ai tools augmenting human thinking, I wonder if that will continue given the agentic and autonomous reasoning capabilities being designed into the next set of frontier models.
Why do you think they will stop helping humans think?
I am not sure I believe human level AI will be augmenting for us as we would like to imagine.
There are a host of congnitive biases that promote AI being used to off land cognitively demanding tasks (automation bias, authority bias, system 1 thinking) that I think will be very hard to fight against.
I don't think AI is very good at automating human thought. Augmentation is way more promising than automation. All the use expert people do with AI is not to get rid of their more precious thoughts but to offload the boring/uninspiring tasks.
I agree the current AI models are not good at automating human thought, but they are getting pretty close imo, for cognitive valuable human thought. Although, High schoolers already seem to be using it to automate their work and teachers are grasping for ways to prevent from being fooled by it. I don't think limiting to how expert people use AI is a good assessment of how AI will impact society.
I agree augmentation is way more promising than automation, especially for knowledge workers. However, the market incentives at play and the desire for human level AI models seem to skew automation over augmentation, from the corporate side of things.
from Mustafa Suleyman, head of Microsoft AI "The tech CEO’s answer was this: “I think in the long term—over many decades—we have to think very hard about how we integrate these tools because, left completely to the market…these are fundamentally labor replacing tools.” https://gizmodo.com/deepmind-founder-ai-davos-mustafa-suleyman-openai-jobs-1851176340#
The economist perspective of AGI impact on human labor markets do not look ideal either.
https://www.nber.org/system/files/working_papers/w32980/w32980.pdf