I have a hypothesis on how computers could gain sentience

[Alex K]

That these types of machines achieve such a state doesn't seem more likely to me than a high end OCR software doing it. Neither has the capacity to, metaphorically, say "Oh sod it, I didn't wanna do this!". Their tasks are too specific and limited. You don't have one unified AI Leviathan living in the Google mainframe figuring out the human condition, it's a bunch of algorithms getting fed numbers and spitting out others. These may be AIs learning patterns in the data they are fed, but they are not set up to learn what they themselves are doing. Methinks.

[Rhondazvous]

While a human can program a computer to use predictions to control other humans, the computer itself, no matter how state of the art it is, does not have a will of its own. A human can fight against bad habits and addictions by sheer will. A computer has to do what it is told.

[Cyberman]

We don't need a computer with antisocial personality disorder.

We might not get the choice.





And turning it off might not even be an option:

[AFTT47]

While a human can program a computer to use predictions to control other humans, the computer itself, no matter how state of the art it is, does not have a will of its own. A human can fight against bad habits and addictions by sheer will. A computer has to do what it is told.

Well, this "sheer will" thing is unexplainable as far as I know. What you're describing is consciousness and we don't have any explanation for it. What makes our brains any different from a computer made of silicon? Complexity? Maybe but we don't have a model to explain that. It's a complete mystery at this point.

Do you understand? If aliens visited us now and gave us the ability to build computers billions of times more powerful than what we have today - even more powerful than the human brain - we still wouldn't be any closer to creating an artificial consciousness than we are now. Because we don't know what consciousness even is. We think, therefor we are but we have no idea how.

[I_am_not_mafia]

Would you agree that the largest internet companies have been learning how to predict, with pin point accuracy, what humans want and desire?

If you agree, then wouldn't it be probable that the computer system running around these predictions, could start realizing that that there was no point in redirecting everything back to the humans. This computer system is able to start manipulating humans using its ability to predict human behavior, by controlling them with the flow of content.

    So I guess I am saying that an algorithm could become so accurate, that it learns that it can control humans with their predictability. Kind of like a quantum level sub-conscious that becomes self aware.

   Crackpot or no? Sorry if I'm missing science. Please correct any errors. Thanks

It's like saying if I put enough weather vanes on my house then I can control the weather.

---

That these types of machines achieve such a state doesn't seem more likely to me than a high end OCR software doing it. Neither has the capacity to, metaphorically, say "Oh sod it, I didn't wanna do this!". Their tasks are too specific and limited. You don't have one unified AI Leviathan living in the Google mainframe figuring out the human condition, it's a bunch of algorithms getting fed numbers and spitting out others. These may be AIs learning patterns in the data they are fed, but they are not set up to learn what they themselves are doing. Methinks.

This.

[BrianSoddingBoru4]

I'm neither a computer guy or a neuroscientist, but I tend to agree with the contention that once a machine 'brain' achieves the same number of synapses - for want of a better word - as a human brain, self-awareness is all but inevitable. Since the number of synapses in a typical human brain might be as high as 1000 trillion, I don't think we have to worry about a Skynet scenario any time soon.

Boru

[Mr Greene]

I'm neither a computer guy or a neuroscientist, but I tend to agree with the contention that once a machine 'brain' achieves the same number of synapses - for want of a better word - as a human brain, self-awareness is all but inevitable.  Since the number of synapses in a typical human brain might be as high as 1000 trillion, I don't think we have to worry about a Skynet scenario any time soon.

Boru

The bulk of which are concerned with maintaining homeostasis.

[BrianSoddingBoru4]

I'm neither a computer guy or a neuroscientist, but I tend to agree with the contention that once a machine 'brain' achieves the same number of synapses - for want of a better word - as a human brain, self-awareness is all but inevitable.  Since the number of synapses in a typical human brain might be as high as 1000 trillion, I don't think we have to worry about a Skynet scenario any time soon.

Boru

The bulk of which are concerned with maintaining homeostasis.

Who synapses choose to have sex with is their own business. Don't hate.

Boru

[masterofpuppets]

On the most fundamental level, a CPU and human brain essentially function in the same way: electrical signals. A CPU's transistor is analogous to a neuron in the brain. The earliest evolved brains were essentially a cluster of nerve cells which then grew in number and size over millions of years, and as this happened sentience and then eventually consciousness gradually emerged.

Our CPUs have been doubling in transistor count every two years for decades now, although this has slowed down in recent years and may eventually stop at 5-nanometre nodes (although Intel has stated that CPUs may reach 100 billion transistors in 2026, equivalent to the number of neurons in the brain[1]). It's certainly possible that CPUs could eventually "think for themselves" and begin making decisions without human input, although whether or not that qualifies as sentience is unknown, as is the case with "primitive" life.

[1] http://www.pocket-lint.com/news/126289-i...in-a-brain

[Mr Greene]

Neurons using electrical signals is more an analogy than an actual fact.
They are also rather more complex than simply on/off switches. 
There are 27 different neurotransmitters and over 50 neuroactive peptides:

Agmatine
α2-Adrenergic receptors, imidazoline receptors
NMDA receptors

Aspartate
Asp
–
NMDA receptors

Glutamate
Glu
Metabotropic glutamate receptors
NMDA receptors, kainate receptors, AMPARs

Gamma-aminobutyric acid
GABA
GABA receptors
GABAA receptors, GABA-ρ receptors

Glycine
Gly
–
NMDA receptors, glycine receptors

-serine
Ser
–
NMDA receptors

Acetylcholine
Ach
Muscarinic acetylcholine receptors
Nicotinic acetylcholine receptors

Dopamine
DA
Dopamine receptors, trace amine-associated receptor 1[12][13]
–

Norepinephrine (noradrenaline)
NE, NAd
Adrenergic receptors
–

Epinephrine (adrenaline)
Epi, Ad
Adrenergic receptors
–

Serotonin (5-hydroxytryptamine)
5-HT
Serotonin receptors (all except 5-HT3)
5-HT

Histamine
H
Histamine receptors
–

Phenethylamine
PEA
Human trace amine-associated receptors: hTAAR1, hTAAR2
–

-methylphenethylamine
NMPEA
hTAAR1
–

Tyramine
TYR
hTAAR1, hTAAR2
–

Octopamine
Oct
hTAAR1
–
 
Synephrine
Syn
hTAAR1
–

Tryptamine
hTAAR1, various serotonin receptors
–

-methyltryptamine
NMT
hTAAR1, various serotonin receptors
–
Anandamide
AEA
Cannabinoid receptors
–

2-Arachidonoylglycerol
2-AG
Cannabinoid receptors
–

2-Arachidonyl glyceryl ether
2-AGE
Cannabinoid receptors
–

-Arachidonoyl dopamine
NADA
Cannabinoid receptors
TRPV1

Virodhamine
Cannabinoid receptors
–

Adenosine
Ado
Adenosine receptors
–

Adenosine triphosphate
ATP
P2Y receptors
P2X receptors

Galanin
GALR1, GALR2, GALR3
–

Galanin-like peptide
GALR1, GALR2, GALR3
–

Gastrin
Cholecystokinin B receptor
–

Cholecystokinin
CCK
Cholecystokinin receptors
–

Vasopressin
AVP
Vasopressin receptors
–

Oxytocin
OT
Oxytocin receptor
–

Neurophysin I
–
–

Neurophysin II
–
–

Neuropeptide Y
NY
Neuropeptide Y receptors
–

Pancreatic polypeptide
PP
–
–

Peptide YY
PYY
–
–

Enkephalin
δ-Opioid receptor
–

Dynorphin
κ-Opioid receptor
–

Endorphin
μ-Opioid receptors
–

Endomorphin
μ-Opioid receptors
–

Orexin A
OX-A
Orexin receptors
–

Orexin B
OX-B
Orexin receptors
–

Secretin
Secretin receptor
–

Motilin
Motilin receptor
–

Glucagon
Glucagon receptor
–

Vasoactive intestinal peptide
VIP
Vasoactive intestinal peptide receptors
–

Growth hormone–releasing hormone
GHRH
Growth hormone–releasing hormone receptor
–

Somatostatin
Somatostatin receptors
–

Neurokinin A
–
–

Neurokinin B
–
–

Substance P
–
–

Neuropeptide K
–
–

Adrenocorticotropic hormone
ACTH
ACTH receptor
–

-Acetylaspartylglutamate
NAAG
Metabotropic glutamate receptor 3 (mGluR3)
–

Cocaine- and amphetamine-regulated transcript
CART
Unknown Gi/Go-coupled receptor[14]
–

Bombesin
–
–

Gastrin releasing peptide
GRP
–
–

Kisspeptin
GPR54
–

Nitric oxide
NO
Soluble 
–


CO
–

Heme bound to 

H2S
–
–