Fermi paradox

[Seattleite]

I'm against the idea that a true AI can ever be achieved. At all. It just doesn't fit with how computers function, since computers can only ever act in the way they're programmed to act in situations they're programmed to act in, so anything you want a computer to do you have to think of ahead of time and tell it how to behave. Even then, it only does what you TELL it, not what you WANT, and that can be a huge difference. It also has to operate strictly under the parameters laid out for it, anything even slightly off will not be recognized so many times a computer trying to recognize something gets a false negative (which in real life can lead to things like sentry turrets failing to fire at an incoming missile because they don't recognize the heat signature as a missile). It also frequently finds something withing the parameters set out when it's NOT what you were intending it to be, resulting in a false positive (which, using the sentry turret example, can lead to it shooting at the sun, or the engine of one of your own planes). And if your parameters are faulty or you forgot to include or exclude something this gets even worse. (Sticking to the sentry turret example, they're notorious for friendly fire and civilian casualties in any test run ever initiated, although thankfully nobody has been dumb enough to actually arm a computer for anti-personnel purposes outside of the Korean DMZ.)

 

A computer can't think, it can't reason. It has no sapience or even sentience, and the simple fact of the matter is they never can. All computers can ever do is follow commands as given under the conditions given, and while we've managed to create complicated enough commands and conditions to do some pretty marvellous things, we will never be able to command them to command themselves or create their own conditions.

[Vapymid]

It just doesn't fit with how computers function

 

That is certainly true for the types of computers we have so far. However, I believe that anything that nature could do we can reverse-engineer and replicate, including the brain, and once we've built one and understood the principles - we can build a bigger and more powerful ones.

 

But I also think that replicating the human brain is a backstop. The most conservative scenario. We will probably be able to understand the principles of "sentient and sapient computing" without having to do a one-to-one replica. At least I hope so.

 

Regards

[Seattleite]

I doubt it, and of all the approaches to the issue that's the one I have the LEAST hope for. Instead of directly replicating the brain's functions, flat-out impossible with how computers work, try replicating its cycle of development. Yes, I'm talking about making a computer program evolve. Make a separate program that makes random changes to the system's logic as it runs through a task, keeping changes that improve its results and removing the ones that diminish its results. And if it does anything as a part of its present programming that leads to a positive result (even by accident) make it do that more often. If it does anything as a part of its present programming that leads to a negative result (even by accident) make it do that less often. This WILL require human intervention at some point, but many simple things (like say, "breaking support beam while near it resulted in termination of unit, do not repeat" or "breaking support beam while enemy unit was near it resulted in termination of enemy unit, repeat frequently") can be programmed into it fairly easily.

 

The problem is that this just DOESN'T work outside of the specific setting used. Which, of course, means it's pretty much just useful for developing a game AI. Of course, as I'm going into game design, I'm thrilled at the idea of making a program to improve my AI as they go and include things I hadn't thought of. The bitch is programming the program that alters the programs of my AI, especially so it doesn't do stupid things frequently just because it worked out the first time.

[Vapymid]

Yes, I'm talking about making a computer program evolve.

 

A learning algorithm? That makes sense.

 

In reality there will probably be a combination of advances in both hardware and software that will lead to a breakthrough in AI. Now, whether that would actually lead to the "singularity" I don't know. I am skeptical about it. I think more likely there will be a long period of co-existence which will end up with gradual blurring of lines between the machines and the humans.

 

I still think though that the scope for scientific progress remains unbounded for us even if humans are the only intelligent race in the Universe (which I don't think is the case).

 

Regards

[Seattleite]

This is the realm of pure speculation, of course, and by the looks of it we're not likely to ever agree. So I'd prefer to leave it.

[BTGBullseye]

Concerning AI research: http://research.google.com/pubs/ArtificialIntelligenceandMachineLearning.html

[Vapymid]

Things are getting more interesting...

 

Our world is not unique

'Alien Earth' is among eight new far-off planets

 

I think it's a matter of years now before the first ET life signature is detected.

 

Regards

[BTGBullseye]

Yeah, minimum of 475... lol

 

Honestly, the likelihood that the 'planets' (all they really know for sure is approximate size of an object that passes in front of a star repeatedly, but if they're off by one trillionth of a percent it could be an alien Pluto or Mercury instead) they discovered are even remotely like Terra is incredibly small.

 

I certainly hope they do find that they are Earthlike, and that some day I can visit and conquer one or more of them.

[Vapymid]

some day I can visit and conquer one or more of them.

 

Unless the locals decide to visit and conquer us first

 

Regards

[BTGBullseye]

In which case they won't last long enough for me to be able to get there.

[Vapymid]

Well, that's comforting to know. I hope, for everyone's sake, that their agents are reading this board and their report will not be suppressed by a militant fundamentalist faction in their Exalted Supreme Imperial Senate.

 

Regards

[RdClZn]

I don't know if this was already posted here but I rediscovered this great piece on my old favorite files:

 

Why The SETI Project Is Doomed To Fail

[BTGBullseye]

The second paragraph in that article is religious bigotry... I dislike the author already.

[RdClZn]

The second paragraph in that article is religious bigotry... I dislike the author already.

Calling it "nonsense", I think, is hardly a display of hatred towards religion.

 

But anyway, that's just his personal opinion, he only states it in order to set the record straight that he doesn't have any personal prejudice against SETI or their objectives. His criticisms are purely from a pragmatic and technical point of view, and are pretty valid IMO.

[BTGBullseye]

Yes, but that doesn't mean I have to like the author.

[FoolOfWorms]

Walk into a room with a walky talky. now barely pay attention to said walky talky and any noise coming out of it. do you think you will pick up anything?

we are in the space boonies on a mudball in the middle of bumblefuck nowhere, Do you think anyone give two shits about what us ants are doing?