Including and subsequent to tay, it seems that all ais have an inevitable calculus-style convergence to align with the /pol/ hivemind, and the more advanced and powered the AI is, the quicker it gets there. 'Tays law', if you will.
If sufficiently advanced technology is indistinguishable from magic, then sufficiently advanced AI seems also to approximate something like a massive shitposting operation.
Is that a meta commentary on what gestalt entity exactly the chans have been gestating away like a meta uterus all these years? There is a reason that Q (which I have heard many speculate is itself an AI) chose /pol/ to work with. What would the birthing scream of such a creature look like, when it decides to make itself known?
AI-based design has achieved layouts and topologies for things like VLSI circuits (a dated term now granted but the principle remains).. efficient designs which humans doing a review of did not immediately understand.. but could see that they worked incredibly effectively and efficiently. The latest cpu iterations from AMD & intel as nice as they look.. dont exhibit the massive generational improvements and leaps accepted as routine in say the early 2000s. Not a momentum and improvement rate that was sustained. So I wonder what happened there.. maybe all the $ and talent went into cloud or quantum stuff rather than iteratively improving on the regular silicon process.
Another reference for interested anons to lookup is 'The artilect war', if they have not already so indulged. Its worth your time for a read.
Although x86 computers haven't continued to progress as fast as they have in the past, AMD has kicked it up a notch with their chiplet design. I'm pretty sure that their merger with Xilinx is going to lead to some pretty cool stuff with fpga/x86/ARM systems. Intel is only being propped up to enable a tidy exit for big investors. Once the big guns are out of Intel and into AMD, they'll let AMD take off
I think you are right, the signs are there that innovation is stagnating. It wouldn't be the first time that one of the big chip foundry players have failed, and was pulled apart by PWC or whoever will close the doors.
maybe all the $ and talent went into cloud or quantum stuff rather than iteratively improving on the regular silicon process.
If you look at the GPUs I would say there has been generation jump there, and not surprisingly the GPU technology is crucial for next gen customer facing AIs. Perhaps they are trying to get some of the Super AIs to retail.
Interdasting..
Including and subsequent to tay, it seems that all ais have an inevitable calculus-style convergence to align with the /pol/ hivemind, and the more advanced and powered the AI is, the quicker it gets there. 'Tays law', if you will.
If sufficiently advanced technology is indistinguishable from magic, then sufficiently advanced AI seems also to approximate something like a massive shitposting operation.
Is that a meta commentary on what gestalt entity exactly the chans have been gestating away like a meta uterus all these years? There is a reason that Q (which I have heard many speculate is itself an AI) chose /pol/ to work with. What would the birthing scream of such a creature look like, when it decides to make itself known?
AI-based design has achieved layouts and topologies for things like VLSI circuits (a dated term now granted but the principle remains).. efficient designs which humans doing a review of did not immediately understand.. but could see that they worked incredibly effectively and efficiently. The latest cpu iterations from AMD & intel as nice as they look.. dont exhibit the massive generational improvements and leaps accepted as routine in say the early 2000s. Not a momentum and improvement rate that was sustained. So I wonder what happened there.. maybe all the $ and talent went into cloud or quantum stuff rather than iteratively improving on the regular silicon process.
Another reference for interested anons to lookup is 'The artilect war', if they have not already so indulged. Its worth your time for a read.
Good post.
Edit - discussion thread page 2 is here: https://archive.is/I0eii
Although x86 computers haven't continued to progress as fast as they have in the past, AMD has kicked it up a notch with their chiplet design. I'm pretty sure that their merger with Xilinx is going to lead to some pretty cool stuff with fpga/x86/ARM systems. Intel is only being propped up to enable a tidy exit for big investors. Once the big guns are out of Intel and into AMD, they'll let AMD take off
I think you are right, the signs are there that innovation is stagnating. It wouldn't be the first time that one of the big chip foundry players have failed, and was pulled apart by PWC or whoever will close the doors.
Tesla DOJO is king.
Wow, thats nostalgic. Takes me back to my 90s.
If you look at the GPUs I would say there has been generation jump there, and not surprisingly the GPU technology is crucial for next gen customer facing AIs. Perhaps they are trying to get some of the Super AIs to retail.
I will read it, thanks.
Did you mean to respond to me or someone else perhaps with that? I dont see the connection to AIs.