Our memories are not actually very good at all. Not that I'm saying data storage is vastly better, because it degrades and becomes redundant as well.

Agree our memories may be fallible, and that data storage has it's problems, but let's just think this through a bit:

- say my mate Bob said to me "let's meet up for a beer at Sammy's on the 25th, about 5".
- "sure" I say, and then forget.

Later on I decide to go back to my fancy transhuman data bank full of memories and I'm thinking, when was I meeting Bob for a beer? Now, what would be needed to get back my original memory (or video recording of the event):

- Your data bank would need an extremely good search function to get through all the memories of mundane existence to retrieve the right piece of information. Maybe I don't even remember when exactly we had that conversation, so I can't just scroll back to Thursday and play it back

- You'd need the ability to 'tag' someone in your memory bank, to say 'this is Bob' and then some really good facial recognition software so that it wouldn't get confused between people, as well as some great semantic language search function, so then you could search for 'time I met with Bob and talked about meeting for beer'

- in all, trying to retrieve the appropriate "memory" would likely be such an ordeal you'd just call (or text/Skype, whatever in the future) to ask him in person again.

...or.. you could just use a little more brain power to rely on our already-highly-evolved memory and facial recognition systems in our biological brains, which have been subject to millions of years of natural selection, and 'remember'!

Personally, I'll bet on evolution thanks.

"The reality, usually completely overlooked by proponents of transhumanist fantasies, is that the evolution of the next predominant cognitive life-form on this planet is well under way."

I doubt the author has overlooked that. Note the quote I responded to in my earlier post. Anthropocentricity in such debates is, by definition, inevitable.

Even so, I agree that whether we need to be fearful of some future intelligence 'submerging the human consciousness', is somewhat moot.

I'm not sure that I care -- unless I was made a slave ;-).

The possibility of the Internet transcending its current form and emerging into a conscious AI makes good fiction. The severest limitations to this as a possibility lay in the random architecture of the web the inability of the web to interpret its sensory inputs and the paucity of connections. The only way wet brains cando what they do is because ey are designed and pruned and configured by ruthless evolutionary pressure. The human mind is created from a modular architecture that is not a feature of the Internet.
Any emergence would be self realising, not driven bottom up from architecture.

But, Yuri, as I am sure you would agree, in any serious discussion anthropocentricity must not be allowed to preclude objectivity.

Also, I have not suggested that the subordination of our species is something of we should necessarily be fearful.

Indeed in chapter 16 of my first book "Unusual Perspectives" I paint a picture of a very rosy future that might possibly be ours if only we are smart enough to become successful symbionts of the new cognitive entity.

That is probably over-optimistic, though!

Given the track record of our species for violence and hostility, extinction may be a more likely outcome.

But it is generally more fruitful to aspire to the positive, n'est pas?

Au Contraire, Seamus, the relentless evolution of technology that has taken place in the collective imagination of humankind over the last 2,5 million years has reached the stage where it will have soon honed to perfection the transition of what is now the Internet into the next imaginative phase of an overall evolutionary process.

An evolutionary continuum that can be traced at least as far back as the formation of the "construction kit" of chemical elements in stars and supernovae.

Contrary to popular belief, although there is a strong stochastic component in its assembly (as with biological entities), there is no way that it can properly be described as a random structure.
Which, I suspect, if you take you time to think it through carefully, would have to admit.

Peter, I don't hold up too much hope for a civil human future, but yes, we should aspire to utopia.

Perhaps the logic of AI will decide that it would be a better place if we all got on together and solved a few problems. On the other hand, Skynet may not give a damn about animal life. Machines R Us?

BTW, Kurzweil seems upbeat about human futures, in the short term at least.

Yes the Internet is structured, but not not in a way immediately conducive to self organisation or, I suspect, sentience. It is wild speculation and good fiction at this stage. Applying teleology will not help.

Not immediately, no, but notwithstanding the accuracy (or not) of Moore's Law, one would have to believe that some cataclysmic event would be required to prevent a future hundreds of years from now that was not inconceivably altered in terms of human-machine interaction, and perhaps consciousness.

I'm no teleologist, but one doesn't have to be.

when the singularity happens (coming your way soonish - say 2050) - all bets are off ...... we might rightly be considered a virus as agent Smith recounted to Morpheus during his interrogation.

Yes, I know . . . I was attempting to conceive of an internet evolved over 200 years and not just 30 -- considering the pace of technology change. Having fun with this.

I agree that a 'sentient internet' is a tall order considering that the primate brain is around 50 million years evolved, Homo (from habilis) around 2 million.

A human brain may have 100 billion neurons and 100 trillion synapses and a complex of biochemical neurotransmitters and hormones for communication and function, as I'm sure you know.

How an electromechanical construct might become 'sentient' and capable of emotional decision making without that complex of neurotransmitters (and hormones) that drive brain function is, of course, a vexed question.

For example, could a sentient internet or robot (or transhuman) become aggressive without testosterone, or happy or sad without serotonin or dopamine?

Or would the sentient robotic be completely logical, Spock-like, and deliver decisions solely on information and practical imperatives, or instruction, whatever that may mean?

A few further thoughts. Cook in 1770 in that small, wooden boat could not have conceived of an iPhone or the internet - not in his wildest dreams. Unknown unknowns.

Yet now the pace of change is exponential compared to Cook's era.

What might we expect in another 250 years of increasing sophistication of biotechnology, including possible bio-quantum computers, nanotechnology and gene manipulation all melding into one biotech paradigm? We must be less than 50 years away from some form of reliable quantum computer, conservatively.

The idea of humans and transhumans might be irrelevant as a gradual and perhaps subtle integration of biology and semiconductor technology proceeds. And the AI may not be 'artificial' at all.

The 'missing link' in the artificial brain of the self-aware network or robot - the biochemical synergists that provide (emotional) sentience - might easily be simulated with electrochemical constructs, if not by integrated genes and DNA, or perhaps a fusion.

The transhuman might eventuate from the top down rather than the bottom up. That is, a robot could be infused with biological constructs rather than humans receiving IC augmentation.

Thinking about your comment, and its a very interesting subversion, I don't know that biological mediators of cognition could ever be grafted onto a bionic substrate. Eg. Hormones/neuro transmitters would never articulate with electrical systems, but I suspect the answer is that Artificial Sentience might apply arithmetic analogues of neurotransmitters to simulate emotion.

"Artificial Sentience might apply arithmetic analogues of neurotransmitters to simulate emotion."

Yes, that occurred to me as well. But when you say "would never articulate with electrical systems", you will be aware that the human body makes good use of bio-electrical systems - sinus node in the heart, CNS for example.

Pure speculation of course, but I'm thinking more of 'seamless integration' in the longer term rather than 'grafted on'. But as you say, an arithmetic approach may be more applicable, depending on the bio-sophistication of the entity. If we can augment humans with semiconductor enhancements, why not augment robots with biological enhancements? Tougher call perhaps, but worth consideration.

There's an interesting discussion on TED Conversations that discusses biological and physical electrical systems similarities. Some useful comments.

http://www.ted.com/conversations/9089/our_bodies_are_amazing_nano_mi.html

Yep,
I understand how the synapse works and neural transmission occurs. It is electrochemical, not electrical. Whilst an electrical stimulus will provoke neural depolarisation I'm not so sure that that bathing a electrical circuit with a fluid containing hormones will provoke any response... Or if it did you would need an analogue of cellular receptors... Why bother when you could use an algorithm to provoke the same response?
There's a deeper philosophical problem of whether you can actually program a self actualising emotion-producer. What would you do if the only way you could feel angry was to press a button on your chest marked 'angry'? Would you press it?

Yes, cell receptors (gateway proteins for cellular absorption of biochemicals) would be required for activation of hormones etc. Clearly, sloshing droid or network circuits with chemicals is not going to work, and simulating receptors in code doesn't sound like a good idea either!

But this goes to the core of the issues of sentience and self-awareness. For example, the 'fight and flight' hormones, noradrenaline, adrenaline, and then cortisol are inherent in animal 'fitness' (evolutionary fitness), for without them we would not survive threats and pass on our genes. The complete gamut of neurotransmitters and hormones are involved in feelings and 'sentience' if you like, even though the philosophical definition of sentience is not without disagreement. The vast evolutionary scale of such human animal adaptations seems too daunting to attempt to re-create in an algorithm for a silicon-metal entity.

So, could an AI - robot or network - actually be sentient? Could you program a droid to care about its own welfare, or to exhibit altruism or empathy, in any meaningful way? Self-awareness is another, separate cognitive principle. Would a network AI recognise itself in a mirror ;-). (Okay, had to look it up, but these animals recognise themselves in a mirror: humans, great apes, elephants, dolphins and orcas, and magpies.) Even without passing the mirror test, I'm suggesting that AIs that are somewhat self-aware, yet lack sentience might be dangerous indeed if they were self-replicating.

However, perhaps we could look beyond the silicon-metal construct (which may only be a phase, or one branch of AI evolution) to the possibility of biodroids and bio-networks. We might then see robots and networks built of organic materials and utilising gene manipulation, nanotechnology and bio-computing to create organic AIs. With an organic AI, manipulated to code for protein synthesis in RNA, many things start to seem possible. Whether such entities might become self-aware and sentient and decide to do their own thing is worth contemplation (or not!).

Regarding your last comment about the "angry button", one would have to ask how such an entity would know it wanted to be angry? Even if it could perceive a threat, such as a weapon aimed in its direction, 'emotions' (fear, anger?) might not be the best basis for a purely defensive response; a logical response may be better and faster. Kill the weapon pointer. But therein lies the problem. Such a response is only half-smart because it probably can't evaluate the complexity of the situation, the possibility of collateral harm, or the intent of the pointer -- and why that may matter. This is not sentience, in my view.

So ultimately, we probably agree up to a point -- and that point is that silicon-metal AIs are not going to be all that sentient, and probably no match for humans -- except in chess! But let's be vigilant just the same ;-).

As you point out, Yuri, we have to be rather careful in our use of the word "sentience".

For the empiricist, however, its meaning can be resolved into two components:

Firstly, that of possession by an entity of sensory input. We would generally tend to extend that definition to include within this category information input of information that pertains to that entity as a prerequisite. This, of course, equates to consciousness, awareness of self.

At a very low level your web-cam with your computer system in its field of view is conscious and within this context, sentient.

Similarly, a bacterium can, in this way, be regarded as sentient. But perhaps we can draw the line at a thermostat? :>)

The second, quite different component, is that of emotion.

Emotions being endogenous algorithms which serve to modulate behavior in such ways as to optimize the probability of replication of the genes expressed by the organism.

These algorithms have evolved to match the environmental niche of a species and can be implemented in various ways and at different levels of sophistication.

In ourselves and other vertebrates the neural responses largely originating from the amygdala are greatly enhanced by the mediation of hormones.

It is more difficult to identify your computer/webcam example with emotion, although we should perhaps bear in mind that it will almost certainly have a proprioreceptor that, if it overheats, will launch an algorithm that modulates its behavior in such a way that it shuts down!
So be kind to your poor sentient computer :>) Although it does not replicate, the idea complexes of which it is an expression certainly do.

I am pleased to see that you are using "network" as well as the android/robot/transhuman notions of yesteryear in your cogitations.

One of the concepts which leads you, along with most others, into the anthrocentric trap is that of design. It is also the concept that lays at the heart of most religions. We intuitively assume that individuals of our secies "design" things. But only in a very trivial everyday sense is this seemingly obvious notion valid.

It can be argued, with strong evidential support, that we do not, invent or create artifacts or systems but that , rather, these are more properly viewed as having evolved within the collective imagination of our species.

To quickly put this counter-intuitive view into focus, would you not agree that the following statement has a sound basis?

We would have geometry without Euclid, calculus without Newton or Liebnitz, the camera without Johann Zahn, the cathode ray tube without JJ Thomson, relativity (and quantum mechanics) without Einstein, the digital computer without Turin, the Internet without Vinton Cerf.

The list can. of course be extended indefinitely.

Nature rules! Not humankind. And that my friends, is the essence of my proposition that the present Internet is the embryonic form of what will soon become the predominant cognitive entity on this planet.

Incidentally, I note that you still speak of silicon/metal devices implementing your future devices. In actuality, the many allotropes of carbon are far better candidates for any projected advances in information processing. Diamond looking particularly suitable for quantum devices.
Certainly for Netty, whose daughters will, as suggested in chapters 15 & 17 of "Unusual Perspectives", be far better equipped for extra-terrestrial activities with yet another set of the special properties of carbon.

Have a nice day.

Peter, anything that you or I write is anthropocentric, even if that is not your intention, because to suggest otherwise is oxymoronic. And I would suggest that the religiosity of the Singularity movement is indeed anthropocentric.

To suggest that the future of technology advancement in which AI's do not achieve sentience or even self-awareness is not necessarily anthropocentric. It may just be good analysis! Your position is a non-sequitur.

(Personally, I am not a humanist. I see Homo as just another genus out of control, driven by an accidental big brain to wreak havoc on a fragile planet.)

Even so, I'm not ruling anything out over large timescales, if you read my posts. But the Singularity in 2050? Give us a break!

Sure, I've seen reports of the work on carbon transistors and other approaches as well, but it's not a done deal, and silicon is the standard at the moment. Lets not get ahead of ourselves.

BTW, your webcam example doesn't do it for me for sentience!

I would suggest Tim that to upgrade such a system a jackplug or nanopill software would be a more elegant design with the internal hardware constructed in such a way to eliminate the need for major surgery - think Matrix.

Yes, but I still think that that technology would become quickly redundant as well.

I was watching a panel discussion with Neil DeGrasse Tyson were he was asked about extra-terrestrials. Part of that was him talking about technology and how amazing it is. His example was that an iPhone from today taken back to the 90's would be mind blowing, possibly leading to witch burnings, because of the massive changes in such a short period of time. Just think, we've gone from 5.25" floppy disks to 3.75", to CDs, to DVDs, to pen drives, to cloud storage in about 20yrs, that's only a fraction of our lifespans.

So upgrades would still be problematic, because you are going to have compatibility issues at some point.

So the kid that is born with no arms and legs is perfect?

So the kid born with their organs the wrong way around is perfect?

So the kid born with the heart defect is perfect?

You might want to reassess your belief structure there Ian.

That argument might have substance except that transhumanists couldn't care less about the child with birth defects, they are only interested in extending and augmenting their own selfish lives. Indeed, if they ever get near their goal, the second part of the agenda will come into focus. It has been mentioned in another comment already, the great cull of humanity. So who do you think they would choose to cull and who would they choose to allow to live (who's playing God now?)???

These defects you mention Tim are most likely caused by something manufactured. In other words, they were caused by humans. If they were not I would be very suprised. I have not suggested that people with medical problems should not be given the best possible help available but that is a totally different thing from an intellectual circle-jerk planning how to transform themselves into 'Gods'.