These seem like extremely bold predictions to me. Anyway, it's great to see that someone is finally trying to do something about the singularity rather than the all too common daydreaming about it.

Human level AI? It's a possibility. Conscious artificially created beings? Not likely. I don't doubt that within the next 10 or 20 years they will develop AI that can aproximate a human being, but to actually be self aware and to be able to go beyond their programming... That remains to be seen.

It's seems quite bold to me. The main issue is, that I think verifying and validating AI is still a neglected part of AI systems development. Maybe I'm a bit of an "old-school" kind of system developer, but testing applications is, even for current systems, a cumbersome process. Add to that the political sensitivity, that already currently blurs development of even the simplest systems, gives me the feeling that on the positive side, maybe technological capabilities may be present in this timeframe, but that commonly accepted AI applications could still be extremely difficult

Blasphemy!

Just joking. This is great because this is the first time that someone has got a real company seeking life or a tl least a backup of YOU on a chip.

I think this is great  for storage and retrieval but not as a total replacement for life. I am a firm believer that entropy (the universe's tendency of decay plays a vital role in life and doesn't allow us or the universe to go crazy; with too much information plus adding more sensitivity like for skin So there is always a balance of choice and structure that can be gained from tech. Tech can only enhance life. That's it's role. Robots shouldn't replace us all because then it would defeat the purpose.

I am sort of not understanding these cyber-implant (punk) people although they look cool. I read Richard Morgan and I don't believe his future will be a reality (too archaic) but I do believe in his uploading structure. I don't think we will have the vulnerability of a cortical stack destruction which would lead to massive crime If we do this properly I do believe it would create a Utiopia (or a constant) with more simulated (controlled) entropy or violence (video games). You can still have a constant and simulate non-consistency. You couldn't have a constant without flux.

I am excited about this because I think it will remove the tendency of the universe toward mass extinction or 'wrong acts and evil.' I also believe that the universe has naturally progressed to this state and it might be up to us to create new universes from now on or galaxy hop.

With enhancement we will better ourselves and our lives to an extent to where life will seem pretty stable. I think the largest problem we face today is health and life extension. Also I think the removal of money or scarcity will stop crime. Free software and trade is helping this, but allot of DRM issues are trying to curb it. We are still childish.

About enhancing ourselves to make us look more beautiful; I am pretty good looking so I don't need it but I am for it if people desire a need to look a certain way (better) as long as they don't steal someone else's identity. It should be an open source system to keep it fair, as if someone doesn't want it then maybe it wasn't right for them. That choice would leed people to be and act better. If they don't choose then they either die or get replaced.
I think it's good because mentally challenged people (jerks) could become better people. Tempting bad people to become good.

I believe all things can work for the common good and. I think if someone wants to be better then maybe that's their rightful place and were all better and more original for it. Some people say that it would be fake or not original but I don't think people are going to change themselves that much but just enough to be the best they can be, maybe if someone decides to become very alien they would separate from us and maybe that is just the evolution of the universe but for me I like being human and who I am so that's me except the death and age thing. We would have to figure out how else to make babies.

So for me I just would want a new body every so often and no cyber-torture, thank you. i wouldnt want to be totally on a computer all the time. It's too fast.

Interesting development, we'll see where it goes.

What a laugh.

Right now and up to 2015 our current society is completely incapable of dealing with selfaware, conscious, tool using, talking AI. The backlash and trauma, if it happens, will be most amusing.

We cannot make any sincere statement, other than based on completely unfounded intuition, that AI by 2014 is either ridiculous, impossible, unlikely, possible or likely. We simply do not have a clue. My intuition tells me, based on nothing but primate belly itch, that it'll happen, somewhere between 2020 and 2035 and some sort of singularity will happen 1-3 years after that. But we could all be completely wrong on all counts.

What I *do* consider more relevant is *who* is gonna manufacture this technology - and if the organisation creating AI is one step removed from the US pentagon I am deeply concerned. Simply put - I do not trust the current US military and leading political apparatus. These people are pretty damn likely to trigger a skynet.

Although I believe in the premise of strong AI and I believe it is possible we'll achieve such before this century is out, I'm skeptical of this company's claims that they will succeed where dozens of universities and research institutions have so far failed.

I'm willing to be convinced but I'll believe it when I see it.

Mr. Farlops,

That is always a good position to take (in general).

I'm curious as to how they define "human-level."

The Blue Brain Project is copying the human brain to treat mental disorders and that's tied to the government complex through University.
This company wants to create AI, from nothing, that is completley humanlike; but for what purpose? I don't see how it could relate to the the Singularity if it doesn't include some form of human life extension. Just more wow graphics
hard to trust anything unless it's open and standardized.

I don't think Blue Brain cortical column simulation project is specifically aimed at understanding mental disorders. I think it's purpose is much more general and abstract: to create accurate models of cortical columns of neurons in vertebrate brains. I suppose in a very indirect way it might shed some light on some mental illnesses.

I think the Blue Brain Project might have a lot to teach us about AI  in the long run. It's very ambitious by some measures but, let's remember that cortical columns each only compose a small portion of an animal's cortex leaving hundreds of thousands or millions more systems, suborgans and tissues to model. Accurately simulating one column, which currently requires IBM's most powerful supercomputers, still leaves us with lots of work to do.

As computer power increases maybe one day we'll have such simulators running as screen savers on our desktops, just SETI@Home is now. Each simulator running one column each and then having them all coordinate with each other in parallel. That might be something--imagine that! A million Blue Brains working in parallel across the Internet!

The original intention of the Blue Brain Project was for Metal Disorders but maybe the scope has changed.

I'm not sure but I think that IBM is only half a Petaflop. And I think they need a full PetaFLOP but if you need parallel, better to just simulate one person at a time the rest would be held in holographic '3D' storage; for simple life extension.

If everyone is having an active experience then, yah, I would think each person would need his own machine but it would have to be Thin Client over a regulated server at those speeds. The government, probably the U.N. too,  already has heavy restrictions on this type of computing speed with Sun's Grid computing etc.

I don't think one government will cut it as it can only monitor its own borders so it will have to be  a U.N. Agency or a non-profit like ICANN. Hard for me to trust tax collectors these days.

Also they are making major breakthroughs in Quantum Computing as well as nano.

A cortical column is tiny, only a two or so cubic millimeters in size. Each containss about 80 to 120 neurons.

If it takes 0.5 petaflops to quickly and accurately simulate the activity in one cortical column now, another ten or fifteen years will bring that power to our desktop machines. This may allow us to move to next stage of neural simulation: networks cortical columns.

How many cortical columns are there in the human cortex? I don't know but, that's how many machines running simulations it would take to accurately model the surface of the human neocortex.

But maybe BBP might teach us ways to simplify things and cast out extraneous detail so the simulation requires less processing power. There maybe further simplifications that might emerge as we learn how accurately model larger and larger portions of the brain.

To me this is where the real path to true AI lies, reverse engineering the machine we know has consciousness: the human brain.

sorry, "networks cortical columns" should read, "networks OF cortical columns"

Note, Novamente isn't the only company or entity trying to do this.

Here is a good roundup of those trying to do it now:
http://www.agiri.org/forum/index.php?showtopic=44

If the list seems small to anyone else, it seems that way to me as well. It would be nice if we had some more trying to do so.

..also, to see what Ben Goertzel (chief designer of the Novamente AGI system, and CEO of Novamente LLC), here it is (3rd post down):
http://www.imminst.org/forum/index.php?act=ST&f=11&t=10799

Very nice. The list is short but if there were more open source projects that could help. I see a couple of non-profits but open source deosn't have to mean that.
Anything open source?

Hot on the heels of the Google geeks talking about how much they want to get some AI action, the firm Novamente plans to achieve human-level AI within six years, provided they get adequate funding of course. In the next...

If it takes 0.5 peta flops to simulate approximately a coupe of square milimetres then it should be possible within 15 years it should possible to simulate the brain assuming internet linkagea are  fast enough and you can get enough people interested in the project to run the software the internet should be able to simulate a consciousness for the first time. Maybe 2020 will be the year.

Thanks, Simon.

For an updated overview of the Novamente project, please see:
http://www.imminst.org/forum/index.php?act=ST&f=11&t=11197