I don't know where you live, but in my town, there are plenty of roads with parallel parked cars next to a sidewalk with kids playing on it, who could decide at any point in time to run out into the street from between parked cars to chase a ball. Driving slow enough to guarantee that you can always stop is not an option. People just drive the speed limit, and assume that the kids will watch out.
"those two people on the sidewalk are.an adult who is holding a child's hand, therefore the child is not likely to run out", or "the adult is not holding the child's hand and he is running everywhere, he may run onto the road".
It wouldn't surprise me if machine learning could get a better score at predicting such behavior, by the way.
Rocket fuel is cheap, though. I think a Falcon 9 load of fuel is about $200,000 mostly for the highly refined kerosine. Future engines will use methane which is quite a bit cheaper, since it's much easier to purify. Given that people pay millions just for the satellite hardware, you hit a point of diminishing returns of launch cost reduction.
Also, most satellites are launched in LEO, and a space elevator is only marginally useful for that. You still need a good boost to get to orbital speed.
And of course, rockets are more flexible. You can make a bigger rocket for much lower cost than you can build a bigger space elevator. You can also use the same rocket for polar orbits.
Plus you have to add cost for maintenance of the space elevator. You're not just paying for electricity.
The problem is that rockets cost tend to be heavy on 'marginal' costs, i.e., shooting twice as many rockets costs twice as much more
There is still a lot of cost reduction that can be done to the rocket design. Right now, even the F9 rocket wastes the upper stage. If the BFR can do reusable upper stages (and the fairing), the marginal costs will come down a lot. Ultimately, if you could just refuel the rocket and launch it again, the marginal costs could be lower than the space elevator, especially for LEO.
Yes, I was aware of that. However, it's a decent indication that the requirements are already uncomfortably close to the margin, especially considering that many real world practical problems needs to be factored in that would add more weight to the tether or that would degrade its strength.
If we could land the rocket booster and reuse it (crazy idea, I know) then you should compare the price of electricity with the price of the booster's fuel.
The price of the boosters should then be compared to the construction costs of the space elevator.
a 1cm cable of the carbon nanofibre in TFS/TFA would be capable of supporting a 5000Km length of itself at 1g.
According to Wikipedia article on space elevators:
An untapered space elevator cable would need a material capable of sustaining a length of 4,960 kilometers (3,080 mi) of its own weight at sea level to reach a geostationary altitude of 35,786 km (22,236 mi) without yielding.
Like I said, it's barely strong enough to hold its own weight.
Isn't the demonstrated fact that if you eat organic = you are likely to have that outcome enough?
That hasn't been demonstrated, though. Maybe it you did a study, you could find that people living in zipcodes that start with "9" have better school grades. Based on that study, would you argue that we should renumber zip codes so that they all start with "9" ?
Higher organic food scores were positively associated with female sex, high occupational status or monthly income per household unit, postsecondary graduate educational level, physical activity, and former smoking status
And yet, in their conclusion they write: "promoting organic food consumption in the general population could be a promising preventive strategy against cancer"
They could also have written instead: "promoting organic food consumption in the general population could be a promising way to increase monthly income per household unit"
Studies based on surveys, like this one, should be viewed with far more skepticism than studies based on controlled experiments.
Studies like this should not even appear in popular press. It should be published in an appropriate journal, and then used as a possible starting point for real research. When a real causal link is confirmed, then it's time to publish an article in the paper.
Turbine on ground turns to raise things up like flagpole on cable.
So, there's an additional cable and a pulley to hoist the load ? The problem is that carbon nanofiber is just barely strong enough to hold it's own weight. You don't have much budget for additional infrastructure mass hanging off the tether.
In a population-based cohort study of 68946 French adults, a significant reduction in the risk of cancer was observed among high consumers of organic food.
reduction: the action or fact of making something smaller or less in amount, degree, or size.
"Making something smaller" implies causality.
Their article continues: "Although the study findings need to be confirmed, promoting organic food consumption in the general population could be a promising preventive strategy against cancer.".
What does "confirming" the "study findings" mean ? The study only found associations. Confirming the association doesn't lead to their conclusion.
So yes it would appear that scientists actually know what they are doing
They realize that there are many confounders, but they seem overly optimistic that they have removed most of the effects. Quoting from their paper "residual confounding resulting from unmeasured factors or inaccuracy in the assessment of some covariates cannot be totally excluded."
They don't even acknowledge that the Cox model is too idealized to be used with any degree of authority. Also, in order to properly control for confounders, you need to have the right data set. If, for example, a healthy lifestyle correlates strongly with consumption of organic food in your study group, then you cannot pick the two parameters apart. You also need groups of people with healthy lifestyle that don't eat organic food, and people that eat organic food but follow a bad overall lifestyle. Those last two groups are usually going to be much smaller, leaving you with very little information (the determinant of the matrix your trying to invert is getting close to zero, which means the errors are going to be big). The more different health parameters you're including in the study, the smaller the different subgroups become, and the bigger the errors.
You might be right but it's at least equally possible that you lack vision about the changes it could make.
At the time that people start collecting the funds and building a space elevator, they must similarly lack vision regarding possible future cost improvements of rockets.
Many studies are not necessarily better if they are all sloppy. The problem in nutrition is that most studies are just observational. The controlled trials are often done on animals, and/or short duration, and/or small, and/or poorly controlled. Ideally, you'd want to have good control, on large groups of humans, for most of their life. That's just not possible.
Sure, most studies try to do it by running the Cox multivariate analysis model, checking residuals, and then claim they are done. This is a standard feature of the statistical package. You can use it even if you don't fully understand the limitations.
The problem is that the Cox model assumes the confounders are: linear, independent, and time-invariant. But we know that for many health-related parameters, none of these three conditions hold. Dose/response usually follows a bathtub curve. Bad things enforce each other. For example, high blood pressure is a risk factor because it stresses the arteries. Smoking is bad because it stiffens them (among other things of course). Smoking combined with high blood pressure is much worse than either factor alone, because the mechanisms that the body uses to protect arteries from high blood pressure are especially sensitive to damage from smoking. And as damage accumulates throughout the life span, you can forget about time invariance.
In addition, most studies don't keep track of all possible confounders. You can't control for things you don't measure. And the confounders that are measured, are often measured with huge uncertainty, such as sending people a questionnaire where they are asked to estimate the amount of french fries they ate last year.
Slashdot Mirror
A self-driving car cannot tell the difference between an animal crossing the road and a small child
State of the art machine learning can already do that.
How would one legally get from one side of the road to the other when in the countryside without a crossing in sight?!
I know Americans that would take their car.
I don't know where you live, but in my town, there are plenty of roads with parallel parked cars next to a sidewalk with kids playing on it, who could decide at any point in time to run out into the street from between parked cars to chase a ball. Driving slow enough to guarantee that you can always stop is not an option. People just drive the speed limit, and assume that the kids will watch out.
"those two people on the sidewalk are.an adult who is holding a child's hand, therefore the child is not likely to run out", or "the adult is not holding the child's hand and he is running everywhere, he may run onto the road".
It wouldn't surprise me if machine learning could get a better score at predicting such behavior, by the way.
Rocket fuel is cheap, though. I think a Falcon 9 load of fuel is about $200,000 mostly for the highly refined kerosine. Future engines will use methane which is quite a bit cheaper, since it's much easier to purify. Given that people pay millions just for the satellite hardware, you hit a point of diminishing returns of launch cost reduction.
Also, most satellites are launched in LEO, and a space elevator is only marginally useful for that. You still need a good boost to get to orbital speed.
And of course, rockets are more flexible. You can make a bigger rocket for much lower cost than you can build a bigger space elevator. You can also use the same rocket for polar orbits.
Plus you have to add cost for maintenance of the space elevator. You're not just paying for electricity.
The problem is that rockets cost tend to be heavy on 'marginal' costs, i.e., shooting twice as many rockets costs twice as much more
There is still a lot of cost reduction that can be done to the rocket design. Right now, even the F9 rocket wastes the upper stage. If the BFR can do reusable upper stages (and the fairing), the marginal costs will come down a lot. Ultimately, if you could just refuel the rocket and launch it again, the marginal costs could be lower than the space elevator, especially for LEO.
See that word "untapered" in your quote?
Yes, I was aware of that. However, it's a decent indication that the requirements are already uncomfortably close to the margin, especially considering that many real world practical problems needs to be factored in that would add more weight to the tether or that would degrade its strength.
After all, there's no way they'll ever improve the on the very first ones produced in the lab.
You can't improve the theoretical strength of the carbon-carbon bond.
If we could land the rocket booster and reuse it (crazy idea, I know) then you should compare the price of electricity with the price of the booster's fuel.
The price of the boosters should then be compared to the construction costs of the space elevator.
a 1cm cable of the carbon nanofibre in TFS/TFA would be capable of supporting a 5000Km length of itself at 1g.
According to Wikipedia article on space elevators:
An untapered space elevator cable would need a material capable of sustaining a length of 4,960 kilometers (3,080 mi) of its own weight at sea level to reach a geostationary altitude of 35,786 km (22,236 mi) without yielding.
Like I said, it's barely strong enough to hold its own weight.
Isn't the demonstrated fact that if you eat organic = you are likely to have that outcome enough?
That hasn't been demonstrated, though. Maybe it you did a study, you could find that people living in zipcodes that start with "9" have better school grades. Based on that study, would you argue that we should renumber zip codes so that they all start with "9" ?
Higher organic food scores were positively associated with female sex, high occupational status or monthly income per household unit, postsecondary graduate educational level, physical activity, and former smoking status
And yet, in their conclusion they write: "promoting organic food consumption in the general population could be a promising preventive strategy against cancer"
They could also have written instead: "promoting organic food consumption in the general population could be a promising way to increase monthly income per household unit"
These are both associations they have found.
Studies based on surveys, like this one, should be viewed with far more skepticism than studies based on controlled experiments.
Studies like this should not even appear in popular press. It should be published in an appropriate journal, and then used as a possible starting point for real research. When a real causal link is confirmed, then it's time to publish an article in the paper.
Turbine on ground turns to raise things up like flagpole on cable.
So, there's an additional cable and a pulley to hoist the load ? The problem is that carbon nanofiber is just barely strong enough to hold it's own weight. You don't have much budget for additional infrastructure mass hanging off the tether.
In a population-based cohort study of 68946 French adults, a significant reduction in the risk of cancer was observed among high consumers of organic food.
reduction: the action or fact of making something smaller or less in amount, degree, or size.
"Making something smaller" implies causality.
Their article continues: "Although the study findings need to be confirmed, promoting organic food consumption in the general population could be a promising preventive strategy against cancer.".
What does "confirming" the "study findings" mean ? The study only found associations. Confirming the association doesn't lead to their conclusion.
So yes it would appear that scientists actually know what they are doing
They realize that there are many confounders, but they seem overly optimistic that they have removed most of the effects. Quoting from their paper "residual confounding resulting from unmeasured factors or inaccuracy in the assessment of some covariates cannot be totally excluded."
They don't even acknowledge that the Cox model is too idealized to be used with any degree of authority. Also, in order to properly control for confounders, you need to have the right data set. If, for example, a healthy lifestyle correlates strongly with consumption of organic food in your study group, then you cannot pick the two parameters apart. You also need groups of people with healthy lifestyle that don't eat organic food, and people that eat organic food but follow a bad overall lifestyle. Those last two groups are usually going to be much smaller, leaving you with very little information (the determinant of the matrix your trying to invert is getting close to zero, which means the errors are going to be big). The more different health parameters you're including in the study, the smaller the different subgroups become, and the bigger the errors.
at lease you protecting yourselves from those caravans of migrants hundreds of miles from your borders.
The Uyghurs would like a subscription to your newsletter.
Give me a theory on the possible cost savings then. Nuclear rockets are not an option.
Full reuse with minimal maintenance and minimal crew requirements.
You might be right but it's at least equally possible that you lack vision about the changes it could make.
At the time that people start collecting the funds and building a space elevator, they must similarly lack vision regarding possible future cost improvements of rockets.
How did African elephants end up in China ? They are non migratory.
Many studies are not necessarily better if they are all sloppy. The problem in nutrition is that most studies are just observational. The controlled trials are often done on animals, and/or short duration, and/or small, and/or poorly controlled. Ideally, you'd want to have good control, on large groups of humans, for most of their life. That's just not possible.
It is impossible to control for confounders.
Sure, most studies try to do it by running the Cox multivariate analysis model, checking residuals, and then claim they are done. This is a standard feature of the statistical package. You can use it even if you don't fully understand the limitations.
The problem is that the Cox model assumes the confounders are: linear, independent, and time-invariant. But we know that for many health-related parameters, none of these three conditions hold. Dose/response usually follows a bathtub curve. Bad things enforce each other. For example, high blood pressure is a risk factor because it stresses the arteries. Smoking is bad because it stiffens them (among other things of course). Smoking combined with high blood pressure is much worse than either factor alone, because the mechanisms that the body uses to protect arteries from high blood pressure are especially sensitive to damage from smoking. And as damage accumulates throughout the life span, you can forget about time invariance.
In addition, most studies don't keep track of all possible confounders. You can't control for things you don't measure. And the confounders that are measured, are often measured with huge uncertainty, such as sending people a questionnaire where they are asked to estimate the amount of french fries they ate last year.
That's why you do all your time logging with UTC.
(insert smiley emoji here).
You mean :-)
Almost no one has a job need for daylight
No, but job schedule determines other things, such as commute and free time.
If you have a specific need for daylight, schedule appropriately.
Often, schedules need to match other schedules, outside of our control. For instance, many people can't just decide to shift their work schedule.
intelligence: the ability to acquire and apply knowledge and skills.
Sounds like a perfectly appropriate word to use in this case.