An interesting read…
-
An interesting read…
Only What Is Alive Can Be Conscious
https://www.noemamag.com/only-what-is-alive-can-be-conscious
#brain #neuroscience #science #biology #tech #technology #BigTech #AI #ArtificialIntelligence #LLM #LLMs #MachineLearning #GenAI #generativeAI #AISlop #Fuck_AI #Meta #Google #NVIDIA #gemini #OpenAI #ChatGPT #anthropic #claude
-
An interesting read…
Only What Is Alive Can Be Conscious
https://www.noemamag.com/only-what-is-alive-can-be-conscious
#brain #neuroscience #science #biology #tech #technology #BigTech #AI #ArtificialIntelligence #LLM #LLMs #MachineLearning #GenAI #generativeAI #AISlop #Fuck_AI #Meta #Google #NVIDIA #gemini #OpenAI #ChatGPT #anthropic #claude
This might also be of interest to @albertcardona and @troy_s
️
-
This might also be of interest to @albertcardona and @troy_s
️Thanks, indeed, a nice summary of key takeaways from Anil Seth's essay https://www.noemamag.com/the-mythology-of-conscious-ai/ There's a lot to agree with there, including all the assumptions and bad metaphors that those who equate brains with computers repeatedly fall into.
A few points that come to mind:
* The software/hardware separation isn't as clean in computers either, from the CPU microcode to the OS that it runs. In humans, I don't have an analogy for microcode, but languages change the way we perceive the world, and those of us who speak multiple languages do feel sometimes that we express a different character when speaking in a different language.
* On time, computers operating under real-time constraints are a lot closer to what brains do. The optimal outcome is not the best possible one or a designated one, but rather, that which can do the job in the time and resources allotted.
* Operating systems, and broadly most software, operates on an event dispatch loop, whereby the system awaits an event and then reacts, followed by a restart of the loop body. Having the loop not stop is a key characteristic of our cerebral cortices, where most inputs to any one neuron arrive from other cortical neurons, with as little as about 6% of inputs to some select neurons (in layer 4) arriving from external sensory stimuli via the thalamus (https://onlinelibrary.wiley.com/doi/full/10.1002/cne.22133 ) . A recurrent neural network is not that different, conceptually, bridging the huge distance of self-driven input seeking.
* Biological systems are entropically favored https://arxiv.org/abs/1908.08374
* The latest technology is never a good metaphor for the brain, as Matthew Cobb's book details over and over: https://profilebooks.com/work/the-idea-of-the-brain/
-
Thanks, indeed, a nice summary of key takeaways from Anil Seth's essay https://www.noemamag.com/the-mythology-of-conscious-ai/ There's a lot to agree with there, including all the assumptions and bad metaphors that those who equate brains with computers repeatedly fall into.
A few points that come to mind:
* The software/hardware separation isn't as clean in computers either, from the CPU microcode to the OS that it runs. In humans, I don't have an analogy for microcode, but languages change the way we perceive the world, and those of us who speak multiple languages do feel sometimes that we express a different character when speaking in a different language.
* On time, computers operating under real-time constraints are a lot closer to what brains do. The optimal outcome is not the best possible one or a designated one, but rather, that which can do the job in the time and resources allotted.
* Operating systems, and broadly most software, operates on an event dispatch loop, whereby the system awaits an event and then reacts, followed by a restart of the loop body. Having the loop not stop is a key characteristic of our cerebral cortices, where most inputs to any one neuron arrive from other cortical neurons, with as little as about 6% of inputs to some select neurons (in layer 4) arriving from external sensory stimuli via the thalamus (https://onlinelibrary.wiley.com/doi/full/10.1002/cne.22133 ) . A recurrent neural network is not that different, conceptually, bridging the huge distance of self-driven input seeking.
* Biological systems are entropically favored https://arxiv.org/abs/1908.08374
* The latest technology is never a good metaphor for the brain, as Matthew Cobb's book details over and over: https://profilebooks.com/work/the-idea-of-the-brain/
@albertcardona @troy_s Very interesting, thanks! I've got a lot to read this evening.
-
G gambolputte@expressional.social shared this topic
