Current AI is already running into the limits. It can’t find more data as it already consumed everything on the internet and needs more power for growth that is not available on the grid.
I agree, although LLM models are impressive in some ways, we seem to be hitting some pretty serious limitations of that model.
I’m sure better models will be found, and we will probably have occasional technology leaps in the future as we have had in the past. I just doubt they will continue to accelerate as they have done historically.
Personally I’m a bit disappointed that we haven’t developed more since the 70’s. I thought computer based automation would be much faster and better, and standard working hours would have been about halved around year 2000.
So I’m not that impressed, despite there have been some cool developments. But things take time.
PS: In the 70’s fusion power was estimated by our physics teacher to be about 50 years away.
I remember it clearly, and I remember thinking as a teen, who the hell wants to work on something that will take 50 years?!
Now 50 years later, I’m not sure we are even half way there, and instead of cheap plentiful clean energy, we have climate change because we still use fossil fuels.
So again I’m not that impressed with where we are, compared to what I hoped and expected 50 years ago. And the more time passes, the further away a singularity seems to be. As in it’s never going to happen.
It’s been argued too that because of the way funding and grants work, research goes after buzzwords, commercial interesting stuff and mathematical fiction instead of fundamental research.
Current AI is already running into the limits. It can’t find more data as it already consumed everything on the internet and needs more power for growth that is not available on the grid.
I agree, although LLM models are impressive in some ways, we seem to be hitting some pretty serious limitations of that model.
I’m sure better models will be found, and we will probably have occasional technology leaps in the future as we have had in the past. I just doubt they will continue to accelerate as they have done historically.
Personally I’m a bit disappointed that we haven’t developed more since the 70’s. I thought computer based automation would be much faster and better, and standard working hours would have been about halved around year 2000.
So I’m not that impressed, despite there have been some cool developments. But things take time.
PS: In the 70’s fusion power was estimated by our physics teacher to be about 50 years away.
I remember it clearly, and I remember thinking as a teen, who the hell wants to work on something that will take 50 years?!
Now 50 years later, I’m not sure we are even half way there, and instead of cheap plentiful clean energy, we have climate change because we still use fossil fuels.
So again I’m not that impressed with where we are, compared to what I hoped and expected 50 years ago. And the more time passes, the further away a singularity seems to be. As in it’s never going to happen.
Yes, every 10 years or so. That’s been the norm.
You’re not alone. The rate of innovation is indeed going down.
https://www.sciencealert.com/innovation-in-science-is-on-the-decline-and-were-not-sure-why
It’s been argued too that because of the way funding and grants work, research goes after buzzwords, commercial interesting stuff and mathematical fiction instead of fundamental research.