Random Thoughts Thread

[Montmorency]

If neurons act as a neural network, then a computer software designed for certain tasks and using select characteristics of some neurons will not behave as a "real" neural network.

Lightbulb technology will never be "as good as" a sun when it comes to generating light, both because that's an inherent technological constraint and because the design of a lightbulb, being made for specific purposes, is very far from the design of a star.

[Husar]

If neurons act as a neural network, then a computer software designed for certain tasks and using select characteristics of some neurons will not behave as a "real" neural network.

But now you entered the constraints of using select characteristics and being designed for specific tasks.
My point was that these things are still in development, it seems strange to assume that they cannot one day do more.
And it doesn't just have to be software, apparently some chips are designed as neural networks: http://www.research.ibm.com/cognitiv...ic-chips.shtml

Lightbulb technology will never be "as good as" a sun when it comes to generating light, both because that's an inherent technological constraint and because the design of a lightbulb, being made for specific purposes, is very far from the design of a star.

That is a very strange way to combine these arguments. The physical limitations are relatively obvious, but to say a light bulb is created for specific purposes that differ from those of a star is a bit odd given that the purpose of a light bulb is to replace the light of a star in places or at time where the light of a star is not readily available. Of course the design is different, having a star with the smallest working mass on our planet would be a bit much and also a bit hote, in fact one would rather say having our planet on such a star rather than the other way around.
The light bulb still allows you to see things the same way the star does, so with that purpose in mind, they both provide the same result.

If you are refering to the idea that pouring alcohol over a brain vs pouring it over a computer chip on which a neural network is running yields different results, well, whom are you quoting, Captain Obvious?

[Montmorency]

it seems strange to assume that they cannot one day do more.

What they will be able to do in the future depends on how are present assumptions and what we think we are trying to accomplish are revised. The important thing is that 'more of' what we're doing now will not push boundaries.

they both provide the same result.

And that's exactly what current approaches can manage - and that's fine. But let's not pretend we're building stars when we're just making plain-old lightbulbs.

[Husar]

What they will be able to do in the future depends on how are present assumptions and what we think we are trying to accomplish are revised. The important thing is that 'more of' what we're doing now will not push boundaries.

I'm not so sure, but I'm also not deep in the topic (yet).

And that's exactly what current approaches can manage - and that's fine. But let's not pretend we're building stars when we're just making plain-old lightbulbs.

That's fine, but the comparison with astrology sounds a lot meaner than that.

[Hooahguy]

Is a promotion without a pay raise (or any pay at all) really still a promotion?

EDIT: not talking about my recent donning of the red robes.

[drone]

Is a promotion without a pay raise (or any pay at all) really still a promotion?

Depends. Can you order more people around?

[Hooahguy]

Depends. Can you order more people around?

Yes, I will be working on a pilot program being the liaison between the parent organization and local groups. And somehow Im still an unpaid intern.

[Montmorency]

That's fine, but the comparison with astrology sounds a lot meaner than that.

No, it's spot-on. Echoed here.

[Husar]

No, it's spot-on. Echoed here.

But Astrology is more like a placebo while neural networks actually work. I wasn't aware that people expect them to do our research for us anytime soon. All the mentions I've seen say that such jobs have a very low chance (aka "maybe way in the future") of getting replaced by machines. What I haven't seen until you came along were people who seem to think it's inherently impossibe. Especially after that other theory you had that basically saw the brain as a (rather complicated) input-output-machine. And who says that our future computers can't be partially biological and contain vat-grown neural networks? Some reports of people connecting animal brains to computers already exist.


@Hooahguy: Depends on whether you get something like prestige instead of money, or what drone said, but it can be, I guess.

[a completely inoffensive name]

Lightbulb technology will never be "as good as" a sun when it comes to generating light, both because that's an inherent technological constraint and because the design of a lightbulb, being made for specific purposes, is very far from the design of a star.

Guys, please look up the facts before making statements like these. Wikipedia gives a solar luminosity figure of 98 lumens per watt. I can buy a Cree LED 100W equivalent light bulb that out puts 1,650 lumens at 16.5 Watts. We are already there and the technology is still improving.

[Montmorency]

Guys, please look up the facts before making statements like these. Wikipedia gives a solar luminosity figure of 98 lumens per watt. I can buy a Cree LED 100W equivalent light bulb that out puts 1,650 lumens at 16.5 Watts. We are already there and the technology is still improving.

What do you think raising the subject of energy efficiency relates to my post?

[a completely inoffensive name]

What do you think raising the subject of energy efficiency relates to my post?

First half of your statement was that lightbulb technology will never be "as good as" a sun when it comes to generating light because of "inherent technological constraint".

My point is that we already have lightbulbs that are better at generating light than the sun.

[Montmorency]

My point is that we already have lightbulbs that are better at generating light than the sun.

Why do you say that? Your citing of the metric offers no context to what we make of the sun - and note that with respect to the sun the wattage component of the efficacy apparently refers to the radiated heat rather than direct source electrical consumption or conversion.

No, what you have cited goes toward evaluating some lamps against others in generation of light. Don't misuse metrics in response to a criticism of the misuse of metrics.

[a completely inoffensive name]

Why do you say that? Your citing of the metric offers no context to what we make of the sun - and note that with respect to the sun the wattage component of the efficacy apparently refers to the radiated heat rather than direct source electrical consumption or conversion.

No, what you have cited goes toward evaluating some lamps against others in generation of light. Don't misuse metrics in response to a criticism of the misuse of metrics.

Personally, I think you are trying to avoid admitting that your original comment was sloppy in its wording. This isn't the backroom though, so I don't want to push this thread into a more combative discussion.

[Montmorency]

Personally, I think you are trying to avoid admitting that your original comment was sloppy in its wording.

I can own to that.