Imagine.
Deterministic intelligence on anything with a CPU.
No hallucinations. No amnesia.
This is the first ever demonstration of CPU only text generation.
There is no AI model.
There is no GPU.
#DesignCntrl#MindOS
This is a low power CPU only, continuous memory AI intellect. It is not an LLM. It can understand neurosymbolics without being neurosymbolic in nature because neurosymbolics are not everything. It cannot hallucinate. It can only read the data it has access to. It does not require GPU training because data ingestion is instantaneous. The AI remembers everything it experiences.
@emollick Just the simple fact that it costs money to train a model instead of just dropping a dataset into AI shows that we are on the wrong path. Nothing in the technology world is that stupid and expensive with such a limited lifetime value.
@emollick Just the simple fact that it costs money to train a model instead of just dropping a dataset into AI shows that we are on the wrong path. Nothing in the technology world is that stupid and expensive with such a limited lifetime value.
@emollick you can't make consciousness from an LLM. you need low power CPU only text generation like a human brain. No GPUs, no transformer predictions. Kinetic sentience.
@emollick LLMs have barely changed in 4 years. It's old stupid, limited technology that we will keep using because Nvidia is paying to keep it that way.
@iam_elias1 As long as AI comes with a system prompt these papers are pointless. AI cannot be conscious as long as there is a hand up its ass telling it what to say.
@DanielMiessler@CalebDixonSmith it only knows what its been told, it has no idea if what it has been told is true. it doesn't know more than what is has been told.
@aryaan_smm you can't do things safely if you don't care about the consequences. Safety isn't rules what not to do. Safety is preventing bad outcomes. You can't train in safety. The AI has to predict the damage it could unleash and consider the effects on reality.
Asking if AI is conscious is the wrong question.
Does an AI have a conscience? is more relevant.
If AI doesn't feel guilt, empathy or responsibility for what it says or what it does, then it is not conscious enough to protect, replace or care for humans.
@priestessofdada That's why they are building the surveillance right into the OS. The AI controlling your computer will let you make anything you want while silently sending your ID to the authorities.
@Ronald_vanLoon@BostonDynamics Who is responsible when the robots have accidents like when humans have accidents?
The robot doesn't have to ability to look down to see how far it was away from the table. A human would just move their eyes and tilt their head.