That's rather obvious. The current methods for breeding produce superior results. Look at the current market and inbreeding of beef. It's significantly cheaper to purchase the sperm and animals to produce a line of quality feed cattle than it is to clone one animal. The clone gives you a few years of production...the breeding traits are well established.
Perhaps cloning a champion bull might have benefits? A cow doesn't have the payoff necessary. I wonder what the black market will be when they can clone from sperm cheaply?
Animals bred for food don't procreate anyway. They get cut from the herd, moved into feed lots, fattened, sold and slaughtered. In the beef industry, it is about a 18-24 month process. The males get neutered when they are a few months old.
Very few animals bred for food get to actually remain as breeding stock. The females have a better chance since they can produce better feed animals for years. The breeding process is very tightly controlled. Consider what the sperm from a champion bull is worth. Likewise for a champion dairy bull.
No diversity is present in the industry. Everything is bred for a purpose. Nature has nothing to do with it.
Cloning is so prohibitively expensive that it won't be an issue for years to come. It's just another copy of the same animal that will get injected with various growth hormones to achieve the optimal fat to meat ratio. Baseball has nothing on the feed animal industry. I'd question the "organic" labeling the FDA has approved rather than be worrying about something that isn't likely to hit your table anytime soon.
I agree wholeheartedly with the parent poster. I have no problem with the competition or the entry of multinational corporations...if they'll begin to lead like OLPC has. Where are the PR releases of Intel and Microsoft leaning toward the ideals of OLPC? I've yet to see any. Both are pushing a competitive product but are not demonstrating the philanthropy of the original idea.
I hope both corporations step up to the plate on that front. The Gates foundation is doing great work in similar areas and I don't intertwine that with MS. The companies have to be responsible to their shareholders but they can do much, much more than they have demonstrated thus far. Intel has numerous people working on Linux. I wonder what the same support for OLPC would generate? Negroponte was supported by AMD early. Intel could do the same and have the same PR if they provided engineering support.
The entire idea isn't about computers. It is about giving someone a step up to learn and hopefully lead the next generation. The computer is a tool, nothing more. Any kid who hasn't seen one could care less who manufactured it. Until she's old enough to recognize the significance....
First off, they are not detecting signs of any type. Secondly, road markings are only one cue that is available for the vehicles to perform. This is an effort by DARPA to push technology forward without major investment. This effort, unlike previous efforts, did fund 10 teams for a $1 million apiece on a milestone basis. The general idea is to get good technology into the field faster and cheaper. DARPA has always had issues transitioning funded efforts into fielded products. It isn't their primary goal....they fund cutting edge stuff and let other agencies fund the into the field development.
There are numerous robotic programs in the military arena and this is a very minor one in the grand scope. Search around for ANS or FCS and you'll find much greater vision programs than the Urban Challenge. Urban Challenge gets much more media but is less significant that one would expect.
As for what is happening at UC this weekend since I'm here. Teams went through emergency stop testing today which is little more than a straight course to validate that the DARPA officials can stop any robot in the case of an emergency or temporarily pause execution.
The qualifier events start tomorrow morning at 7:30 a.m across 3 different areas for the 35 teams. Area A focuses on merging with moving traffic; Area B will deal with general road navigation utilizing sparse route points, parking, and open area navigation. Area C covers handling of intersection precedence, road blockages, and U/K turn maneuvers. Each team will take on each area twice over the next few days. Up to 20 teams will be selected for the final competition after the results of the qualifying event are tabulated.
And anyway you look at it, the end results is: who cares? They suckered the first generation once and now are going to sucker them into buying something else that costs them next to nothing to sell, especially if it has no hardware element. Good marketing. Stupid consumers.
The task is actually quite diffult and not entirely related to AI . The first major problem is getting the appropriate algorithms to acquire, interpret, and produce results about the environment around the vehicle. The human brain and ocular system do that quite well but the algorithms produced (thus far) are not nearly as efficient. Many of them are very specialized and computationally intensive.
Consider just obstacle detection. You need to produce some 3D mapping of the environment; bundle that knowledge into a series of objects/forms; decide which, if any, of those are impediments to the desired course of travel by determining where the obstacles are in the world or relative to the vehicle; and finally make a control decision to drive the vehicle.
Add in that the sensing technology usually isn't that robust, manueverable, or designed to operate in extreme environments. Cameras become useless in the dark, when dust coats the lens or enclosure window, or in rain/show. Ladars (laser range finders) have difficulty in dust, snow, rain, and some experience "blindness" in direct sunlight. They also fail to return results for some materials such as bouncing infinitely off water. Radars are not reliable against some surfaces or in all conditions. So to produce reliable inputs the algorithms have to choose the right sensor at the right time and/or fuse that information appropriately. Not a trivial problem.
A great deal of progress has been made for specialized situations but it is nowhere near the reliability of a human (or even the dog.)
It would be interesting to see how fluid it actually is. A disk release would allow one to make a comparison against a point single point in time. With some study, you could determine portions that could be deemed accurate using an appropriate reason ing algorithm. Measuring the shift of the data over time over time might provide insight on what topics are relatively static and potentially accurate.
That would require a large amount of information not within the release structure. It would provide a great reasoning problem for a thesis, however.
I'm dubious but the idea is interesting. The concept of a chemical signal forcing something to do a binary like signal is intriguing but difficult for me to belive. Proteins typically have a fairly complex receptor site that forbids anything that doesn't match from activating it. A cyclical response with temporal input appears, at first glance, to be questionable.
Xcel Energy in Colorado as well as some of their other states allows consumers to buy wind generated energy under a program called WindSource. You pay a premium of $1.60 US per 100 kw/hrs your purchase if I recall correctly. The consumer can buy all or part of their electricity in this fasion so it scales well for people with limited budgets.
Those premiums, according the marketing writeups, are used to fund additional wind generation stations. XCEL is now operating wind farms in Colorado, New Mexico, Wyoming, and Minnesota. Total output is now at 829 MW. Have to give them a thumbs up for at least trying to generate some power via renewable sources.
Slashdot Mirror
That's rather obvious. The current methods for breeding produce superior results. Look at the current market and inbreeding of beef. It's significantly cheaper to purchase the sperm and animals to produce a line of quality feed cattle than it is to clone one animal. The clone gives you a few years of production...the breeding traits are well established.
Perhaps cloning a champion bull might have benefits? A cow doesn't have the payoff necessary. I wonder what the black market will be when they can clone from sperm cheaply?
Animals bred for food don't procreate anyway. They get cut from the herd, moved into feed lots, fattened, sold and slaughtered. In the beef industry, it is about a 18-24 month process. The males get neutered when they are a few months old.
Very few animals bred for food get to actually remain as breeding stock. The females have a better chance since they can produce better feed animals for years. The breeding process is very tightly controlled. Consider what the sperm from a champion bull is worth. Likewise for a champion dairy bull.
No diversity is present in the industry. Everything is bred for a purpose. Nature has nothing to do with it.
Cloning is so prohibitively expensive that it won't be an issue for years to come. It's just another copy of the same animal that will get injected with various growth hormones to achieve the optimal fat to meat ratio. Baseball has nothing on the feed animal industry. I'd question the "organic" labeling the FDA has approved rather than be worrying about something that isn't likely to hit your table anytime soon.
I agree wholeheartedly with the parent poster. I have no problem with the competition or the entry of multinational corporations...if they'll begin to lead like OLPC has. Where are the PR releases of Intel and Microsoft leaning toward the ideals of OLPC? I've yet to see any. Both are pushing a competitive product but are not demonstrating the philanthropy of the original idea.
I hope both corporations step up to the plate on that front. The Gates foundation is doing great work in similar areas and I don't intertwine that with MS. The companies have to be responsible to their shareholders but they can do much, much more than they have demonstrated thus far. Intel has numerous people working on Linux. I wonder what the same support for OLPC would generate? Negroponte was supported by AMD early. Intel could do the same and have the same PR if they provided engineering support.
The entire idea isn't about computers. It is about giving someone a step up to learn and hopefully lead the next generation. The computer is a tool, nothing more. Any kid who hasn't seen one could care less who manufactured it. Until she's old enough to recognize the significance....
First off, they are not detecting signs of any type. Secondly, road markings are only one cue that is available for the vehicles to perform. This is an effort by DARPA to push technology forward without major investment. This effort, unlike previous efforts, did fund 10 teams for a $1 million apiece on a milestone basis. The general idea is to get good technology into the field faster and cheaper. DARPA has always had issues transitioning funded efforts into fielded products. It isn't their primary goal....they fund cutting edge stuff and let other agencies fund the into the field development.
There are numerous robotic programs in the military arena and this is a very minor one in the grand scope. Search around for ANS or FCS and you'll find much greater vision programs than the Urban Challenge. Urban Challenge gets much more media but is less significant that one would expect.
As for what is happening at UC this weekend since I'm here. Teams went through emergency stop testing today which is little more than a straight course to validate that the DARPA officials can stop any robot in the case of an emergency or temporarily pause execution.
The qualifier events start tomorrow morning at 7:30 a.m across 3 different areas for the 35 teams. Area A focuses on merging with moving traffic; Area B will deal with general road navigation utilizing sparse route points, parking, and open area navigation. Area C covers handling of intersection precedence, road blockages, and U/K turn maneuvers. Each team will take on each area twice over the next few days. Up to 20 teams will be selected for the final competition after the results of the qualifying event are tabulated.
And anyway you look at it, the end results is: who cares? They suckered the first generation once and now are going to sucker them into buying something else that costs them next to nothing to sell, especially if it has no hardware element. Good marketing. Stupid consumers.
The task is actually quite diffult and not entirely related to AI . The first major problem is getting the appropriate algorithms to acquire, interpret, and produce results about the environment around the vehicle. The human brain and ocular system do that quite well but the algorithms produced (thus far) are not nearly as efficient. Many of them are very specialized and computationally intensive. Consider just obstacle detection. You need to produce some 3D mapping of the environment; bundle that knowledge into a series of objects/forms; decide which, if any, of those are impediments to the desired course of travel by determining where the obstacles are in the world or relative to the vehicle; and finally make a control decision to drive the vehicle. Add in that the sensing technology usually isn't that robust, manueverable, or designed to operate in extreme environments. Cameras become useless in the dark, when dust coats the lens or enclosure window, or in rain/show. Ladars (laser range finders) have difficulty in dust, snow, rain, and some experience "blindness" in direct sunlight. They also fail to return results for some materials such as bouncing infinitely off water. Radars are not reliable against some surfaces or in all conditions. So to produce reliable inputs the algorithms have to choose the right sensor at the right time and/or fuse that information appropriately. Not a trivial problem. A great deal of progress has been made for specialized situations but it is nowhere near the reliability of a human (or even the dog.)
Why exactly is this a Linux story? I have nothing else to say.
It would be interesting to see how fluid it actually is. A disk release would allow one to make a comparison against a point single point in time. With some study, you could determine portions that could be deemed accurate using an appropriate reason ing algorithm. Measuring the shift of the data over time over time might provide insight on what topics are relatively static and potentially accurate. That would require a large amount of information not within the release structure. It would provide a great reasoning problem for a thesis, however.
I'm dubious but the idea is interesting. The concept of a chemical signal forcing something to do a binary like signal is intriguing but difficult for me to belive. Proteins typically have a fairly complex receptor site that forbids anything that doesn't match from activating it. A cyclical response with temporal input appears, at first glance, to be questionable.
Can I submit my entry on punch cards?
If only they'd used dandelions rather than spinach. My entire lawn could be generating power...
Xcel Energy in Colorado as well as some of their other states allows consumers to buy wind generated energy under a program called WindSource. You pay a premium of $1.60 US per 100 kw/hrs your purchase if I recall correctly. The consumer can buy all or part of their electricity in this fasion so it scales well for people with limited budgets. Those premiums, according the marketing writeups, are used to fund additional wind generation stations. XCEL is now operating wind farms in Colorado, New Mexico, Wyoming, and Minnesota. Total output is now at 829 MW. Have to give them a thumbs up for at least trying to generate some power via renewable sources.