Seems like a loaded word. What you want to do is start off with a huge
universe of hypotheses, and winnow things down from there. The better
we are at excluding things, the smaller the area in which the truth
can survive exists. So, let's say we start with a billion possible
hypotheses, but we find the falsification criteria for all of them
except for half a dozen. Assuming the last six hypotheses don't share
the same falsification criteria, it is an open question as to which
one is "better" - it could be that we can never narrow things down any
further than those six hypotheses, for either practical or theoretical
reasons.
The phase better is a substitution for the word correct. It's
highly unlikely that any of those final six are correct. They are
just better than the ones that we know are wrong.
Also it is not trivial to select a huge universe of hypotheses to
winnow. There are always an infinite number of hypotheses that will
fit any data. What makes science amazing is that we somehow pick
hypotheses that not only fit current data but also predict new
surprising results. Special relativity was strongly motivated by
Maxwell's equations implying that light has a constant speed.
Yeah, I'm mixing metaphors here - although, fun fact, you can get
random sets from deterministic processes. Heck, you can argue that all
random values are in fact deterministic, but simply unpredictable a
priori:)
Right, many people believe quantum mechanics really does describe
a random universe, but who knows. Personally I find it hard to
believe that all the information in the universe was deterministically
contained at the start. Randomness easily solve this problem.
The fact that we can make these predictions based on the wide
variation between day and night, and summer and winter, doesn't imply
that we have any predictive capacity year to year, or decade to
decade, or century to century.
I said 2100 which is around a century in the future. We have
predictive capacity on climate from century to century and we have
historical records going back millions of years to back it up.
This is mostly a factor of what kind of swings we're talking about -
we get massive swings of temperature due to seasons and day/night. The
observed 1.0C/century since about 1850 simply isn't anywhere near
significant enough for us to be very confident.
I can have two random variables that have very similar means and
huge variances. A few measurements of these variables are going to be
all over the place. However, with enough measurements I can say with
high confidence that their means are significantly different.
The statistics are clear (assuming you believe the measurements.)
Temperatures are rising. Why and what impact it will have I can't
really say except by appealing to experts.
Worse than that, our typical "confidence" when it comes to "climate"
tends to revolve around the average global temperature - something
nobody actually experiences. Because our spatial resolution of our
"climate" predictions aren't functionally useful, it's difficult to
divine the impacts - you can have +2.0C global average temperature,
and different distributions of that could mean that humans experience
+10.0C, or even -10.0C. I.e., global average temperature cannot
possibly give you actionable information because distributions count.
That's an interesting point. I don't know much about the models,
so I can't say if they can be improved in this sense. However,
stability is a good thing. Hotter or colder is different and will
cause problems for things like farming. Also more energy in a system
will probably cause more violent weather which again causes problems,
and average temperatures will effect sea level...
Well, specifically they have little value because without th
Well, I'm having fun with this so I'll go another round. (Warning at
some point slashdot locks down the comment system.)
More importantly, you need to make the logical argument that the lack
of these falsification criteria exclude other hypotheses, including
the null.
I guess your talk of null hypothesis and other terminology makes me
think of statistics. And there is an argument to be made about
repeatability in science...
If five hundred different hypotheses could still be valid
with the specified falsification criteria unfound, you haven't made a
convincing argument that your hypothesis is correct. To do that, you
want to show that not only is your hypothesis as of yet unfalsified
(after trying really hard to falsify it), you want to show that other
competing hypotheses are excluded.
Not to quibble, but I assume you mean you want to show your hypothesis
is "better" than the competing hypotheses.
Stochastic means that small variations in the input can have huge
impacts on the output. The butterfly effect creating hurricanes and
that sort of thing.
I think you mean chaotic. The word stochastic is an adjective that
describes something that was randomly determined. One of the
interesting things about chaos theory is that traditionally deals with
deterministic systems. While I'm not a mathematician, I'm sure there
are connections between chaos and randomness. I guess a pseudo random
number is a chaotic system that is used to model a stochastic
system...
The problem with modeling stochastic phenomena is that unless you have
perfect input (which we never do), you will *never* accurately predict
the future in any sort of way, even if your model has every physical
property of matter set correctly. A single grain of sand in the wrong
place, a mismatch five thousand digits after the decimal point in an
input parameter, can cause your simulation to diverge from reality.
This gets into the whole weather versus climate. While this is one
reason it is difficult to predict the weather, I can still say with
high confidence that it will be colder in NYC on Dec 25 2100 then Jun
25 2100. (What confidence means in this context is something you
could question.)
As for a chaotic system being predictable, it can be qualitatively
predictable. There are things like attractors, and I presume one
could build systems with inputs to move in the phase space from
attractor to attractor. (Not sure why an engineer would want to do
this.) There are very credible models that the brain is a chaotic
system which is a sophisticated machine that can be controlled by at
least the owner. (Think about that:))
Astrology is the study of how the universe works. Yet, I think we
would agree that it isn't science. What criteria would you use to
exclude astrology from the realm of "science"?
I'm not quite sure I would 100% agree that Astrology involves
studying how the universe works, but I wouldn't fight someone over the
claim. I would just claim it's bad science. Science is on a quality
spectrum.
Now I don't want to harp on semantics. You kind of asked for my
definition of science, and I gave it. I think it's a pretty standard
definition, but of course like most words, science has many
definitions. You want to define it in terms of a precise meaning of
the scientific method. That's fine, but you also seem to want to
imply things that don't fit your definition have little value. No, you
need to argue the specifics of why they have little value.
Of course not - there is a demarcation, which Karl Popper worked out -
falsifiability. If you include that in your argument/persuasion,
you're playing the science game. If you don't include that in your
argument/persuasion, you may be making really cool arguments, and you
might be very persuasive, but yo
Well the results are better when you factor in how the cost of electricity is rising. I assumed 1.02% is reasonable, but someone here commented it's more like 1.08%. That would mean that a system that generated $100 of electricity today is generating $1000 after 30 years. Very unmortgage like.
Well, the model may be falsifiable, but the model itself is not a
necessary and sufficient hypothesis. That is, even if a model is not
falsified, it simply does not exclude the null hypothesis.
I'm not quite following your terminology. Are you trying to ground
things in statistics where AGW can't do repeated experiments.
Heck, the *exact* same model, given even slightly different inputs,
contradicts itself - a feature of the stochastic nature of weather and
climate.
Stochastic does not mean everything is unpredictable. It's colder
every winter for as long as I can remember:)
Again, it looks "sciencey" but it isn't following the scientific method.
The scientific method does not define science. I would claim it's a
convenient tool to do science. If someone argues your hypothesis is
wrong, you can explain how you tried to follow the scientific method
to justify some of your arguments.
Maybe if we used terms in the same way it would be easier:)
Sure, I would claim science is the study of how the universe works.
Convincing arguments using many tools is persuasion. Empirical
evidence shows that there are lots of religions that are prima facie
non-scientific, but very persuasive.
It seems you are trying to do guilt by association. Science involves
coming up with arguments to show your hypothesis is right. Society
has a long history of evaluating these arguments to determine if they
are true. Even for something as solid as math (which is not science),
old results have been discredited because people found flaws in the
arguments. Just because people use tricks to argue for things that
aren't true doesn't mean all arguing/persuasion is "unscientific".
"Science" is really something we have to agree on before we can have a
scientific discussion, though. Karl Popper and his work on
falsifiability and the demarcation problem represents an important bit
of common ground - and without that common ground, it's *really* hard
to be persuasive to someone who understands the scientific method from
a first principles philosophical point of view:)
I understand the philosophy a bit and spent some time looking into
that perspective, but I didn't find much utility...
The whole point of the scientific method was, as Feynman put it, "a
belief in the ignorance of experts". Whereas before, the only people
allowed to comment on the world were the hallowed authorities, the
scientific method, and the process of falsifiability, democratized the
pursuit of knowledge by setting a table of rules that had to be *more*
than simply persuasive. In the scientific method, you are supposed to
challenge your own ideas - and in fact, challenge them with incredible
fervor to demonstrate their strength. This is the polar opposite of
looking for corroborating evidence, and our inherent human tendency to
confirmation bias.
Another way to interpret (some of this) is to say that the scientific
method is a simple way to make a convincing scientific argument. It's
a useful template or pattern that simplifies evaluation. That's
great, but that doesn't mean it works for every problem in the
universe. Perhaps one needs a different template or even an ad-hoc
argument for some issues.
So, once someone understands the scientific method, thoroughly from
first principles, every "convincing argument" of AGW simply doesn't
seem quite as persuasive. Correlation is not causation, and until AGW
can be stated as a necessary and sufficient falsifiable hypothesis,
it's literally not science.
Just because it doesn't fit your definition of science doesn't mean it
doesn't have value.
Do you know of any good tools to do this calculation? I've been using the on-line mortgage tools. (Google has one pop up when searching for mortgage calculator.) This allows one to depreciate the investment (maybe set the value to 0 after 30 years.) What's missing is a way to inflate the payments. (Which actually makes a lot of sense for a real mortgage.) Do I need to write my own tool? I looking at ground based heat pumps and solar roofs.
In hindsight, I should have been more specific and asked what you
think about AGW that is not science.
However, to keep things interesting, some might claim that climate
models predicting 50 years into the future are not scientific. While
they are somewhat falsifiable, the empirical results will be limited
in generalizability and statistically weak.
While I don't know much about climate science, on principle, I
would disagree. One can make meaningful predictions based on
combining well verified components and using independent historical
evidence to justify the models. Of course this has problems, but so
does all science. One can disagree and focus on a philosophical
definition of science, but real "science" is about making a convincing
argument using the many tools at our disposal.
I assume you are fine with that, and you seem to be focused on a
convincing argument to connect increases in CO2 to human activity. I
already think there's enough evidence for that claim, but it's fine to
strength the claim. Good luck.
1) a list of observations, which if observed, mean a hypothesis is false;
2) a logical argument that the lack of those falsifications means that
a hypothesis must be favored over all others (including the null).
This is a big topic, perhaps you can ground things with climate
science.
You might want to rephrase that. There ARE provable models, where we
understand the math and conditions well enough to predict the future
with in a known level of error. We have a lot of experimentally proven
models which we can rely on for things like lift/drag for an aircraft
wings using fluid dynamics, precipitation run off volumes from urban
developments, short term weather forecasting, aircraft fuel
consumption and more.
I guess I'm OK with that definition of provable. My instinct is to
reserve provable for math, and your original statement didn't give a
lot of context...
Perhaps you mean "No climate model that has a provable amount of error
used for climate change studies"? I'll agree with that.
That's not what I meant, but it's a reasonable point. Should we throw
out all science that doesn't fit your definition of provable? For such
a strong reaction, one would need to give a much more precise
definition of provable. Do we need to do "controlled" experiments?
How often do they need to be repeated? Is evolution provable? Is
astrophysics provable? How about science that deals with the human
brain? I'm guessing good, useful, and honest science can be done
without reaching your standard of provable by combining a range of
evidence. I would say that this includes current research into
climate.
Obviously we are in uncharted territory, where we have CO2 LEADING temperature in our models.... Which, as the gemologist concludes, means our models are not necessarily wrong, but are also not provably right.
No model is provably right. However the behavior has a solid scientific explanation based on feedback loops and is not controversial. https://skepticalscience.com/c...
However my point is that it's debatable whether they
actually make your brain better. They unquestionably are useful in
forcing your brain to do things it doesn't want to do.
It's difficult to define better. Evolution is slow, and our thinking
choices might be better (which is why we evolved the big brains.) If
need to get somewhere in my car, but I'm tired, the smart move is to
take a good stimulant. We haven't evolved enough in the last 100
years to "know" it's bad to fall asleep at the wheel.
the digital sample is a perfect representation of the original analog waveform
It's probably good enough given the typical noise floor, but I
wouldn't say it's a perfect representation. Definitely overkill for
audio, but I'm sure scientists need lots of bits.
You keep confusing the marginal cost to make a car with profit margin of the entire company. They are different things. If you want to make your argument, you need to say why the marginal cost of making a Tesla is wrong.
I curious as to how many of these level 3 chargers are in the US. Is it comparable to Tesla? I think a good charging network is a big factor in selecting an electric car. Tesla has lots of stations and is already at 120kW and is planning to go to 350kW. Is there a comparable option for other manufactures?
For example in ImageNET, the most famous image recognition dataset,
there are a thousand categories including hundreds of breeds of dogs
and cats. Humans get 5% errors with at least a day of training. Neural
nets get 4% error. When you give the net enough examples, it becomes
better than human. We just need to add more training data.
I assume that testing is based on using the same distribution as
training. (The easiest way to do that is to take a big set of
examples and randomly split it into train and test.) Humans do well
even with changes in the distribution which is much more realistic for
many real world tasks.
There is fascinating research to suggest that video (maybe even two channel video) is an essential part of human vision training.
http://www.sciencemag.org/news...
Competition definitely needs to be part of the decision making process. However, I'm pessimistic there would much competition in the current proposals. For example, for a bridge there might be no realistic alternative. Even when alternatives exist there is little incentive to keep prices down when the sum of the demand is comparable to the sum of the capacity and there are significant barriers to entry to create new capacity. It's not like one company can lower their price and take all the traffic; they would get overloaded. It's more likely the companies would optimize their collective profits by matching price increases creating implicit but legal collusion. Corporations hate real competition.
going forward the customers who use the facilities will pay for their
maintenance and upkeep. Just like any other pay to use piece of
property
Don't forget the profit. Since much of this infrastructure will have
monopoly level protection, they can charge what the market will bear
without real competition. It's a massive giveaway. A real market
needs competition not collusion.
One last thing, I'm sure I haven't changed your mind. I'm mostly just
venting. But to anyone reading this who sees through the systems used
to contain and oppress the working class, any ideas how to convince
this AC before it's too late?
To me it seems hopeless. The AC got 6 up votes, but most of what
he says doesn't make sense. It's an important skill to convince
people one knows what one is talking about, unfortunately this skill
seems independent of actually knowing what one is talking about.
If you automate all the jobs away, then where, precisely, is the money
supposed to come from, that your potential clients would use, to buy
your stuff in the first place? In other words: How would you make
money in the first place?
The people, who own the machines would be fine. They would just
trade what their machines produced. If the machines are complex
enough they wouldn't even trade, they could just get the machines to
make what they want. It would become the ultimate information economy
since you would still need ideas for new things to make. This is what
would have value. So I guess one could still get by being a worker,
but I don't see a lot of jobs. As usual, it's best to have capital in
a capitalist economy.
Everybody would be fine with full automation, if the wealth that that
causes would go to those, who actually worked to generate it.
I don't understand this. The whole point is that it will be very
hard for everyone to contribute. The people who could not contribute
in that society would need to form their own economic society, but
it's not clear how they would "pay" for things like resources and land
to run it. It's a bit like the movie Elysium.
Back to your point, I think some simple mathematical models can
show how wealth concentration is inevitable even in a presumably fair
society. Freedom has it's price.
There is an optimal rate, not near 100% and not near 0%, that
maximizes revenue. Raising rates above the optimal rate hurts revenue,
reducing them below the optimal rate reduces revenue.
While agree with what you say, I would also so it's uncontroversial
that this optimal rate is a fickle things. It changes over time and
depends on things like the current rate, so probably whatever rate you
pick this year is not optimal next year. However, I think your larger
point stands and it's not worth dwelling on how to optimize since it's
probably not the right thing for the government's tax goal to be be
maximum revenue.
You got right down to the crux of the problem. What are
we trying to define, and how should we define it? We are trying to
define "minimum" here for the purpose of providing taxpayer subsidies
to telcos.
Yes that is the most important question, though I side stepped it.
I really don't know the penalties and rewards for the ISPs not doing
enough "broadband".
The household that bought a 4K monitor can afford to get more than the
basic service.
I would assume anyone who could upgrade would be considered covered
even if they chose a cheaper rate. (I guess there must be some type
of rule on price otherwise an ISP could offer 1Gps at 1 billion just
to satisfy this requirement.)
If we define "broadband" to be 25Mbps then that means
everybody who has 20Mbps service will get upgraded to 25Mbps, on the
taxpayer's dime.
Perhaps I'm too cynical, but I assumed the ISPs want to lower the
minimum. If it's free money for them then they should want a huge
minimum. Are you saying the FCC is going against the major ISP
wishes?
An HD NetFlix stream is 5Mbps. 10Mbps allows two simultaneous HD
streams, or one HD stream plus plenty of headroom for other normal
activities.
I'm not sure it makes sense to use old habits as a reference point.
Netflix 4k needs around 20Mbps, so by that metric 50Mbps would be
about right. (While 4k is not useful on most tvs, it's nice on a big
computer monitor from 2 feet away.)
I would rather that the FCC define it to be 10Mbps, but actually check
that this bandwidth is available consistently during peak usage.
I'd rather they keep 25Mbps and check it. Just enforce the existing
regulations instead of weaken them.
However, the biggest defect in this FCC proposal is the wireless
aspect. Most unlimited wireless plans tap out around 22 GB per month. At 10
Mbps that's about 2 hours of full speed data a month. So you can
watch about 4 hours of your HD Netflix or 1 hour of 4k Netflix.
You go visit any native reserve with no local industry and see how your community will look.
Yes it could be bad, but I would like to see some reasonable experiments to see if it could be viable with decent regulation. At least it has the advantage that it's not too hard to go someplace else.
No, but the price of all goods rise to what the market will bear. When
everyone's money increases by the same amount "what the market will
bear" goes up as well, hence the price rises.
Of course, it's easy to come up with exceptions to your rule, but
I don't think that's the point. I assume you are trying to give an
intuitive argument for something. (I'm not sure what that is.)
My general point is that economic theory has been a big failure.
It has very little predictive value. However, it is obviously
important and worth studying. I just wish we could extract the value
from the garbage.
Money isn't the endgame of what an economy does; it never has been,
rather it is all about creation and allocation of resources.
Very true.
If you just give people money, you just give them money,
and at the end of the day they'll just outbid one another for those
same resources.
You go to the auction to buy groceries? I think the intuitive idea is
that there is a fixed size pie. UBI gives some people a larger slice
of that pie than no UBI.
If nobody builds anymore housing in say SF, then guess what? No amount
of UBI is going to solve the shortage of available housing.
This is not a problem with UBI. This is a problem with housing in
SF. If I was an enterprising business man, to profit off UBI, I would
create a community in an area with low property costs. I would bet
that some of the people with no good place to live would relocate
given that they had no job to tie them to a location.
Note: I'm not saying this community would turn out great. (It
could be a nightmare.) This is why one needs to do real experiments
with UBI instead of trying to apply naive economic theory to it.
Slashdot Mirror
The phase better is a substitution for the word correct. It's highly unlikely that any of those final six are correct. They are just better than the ones that we know are wrong.
Also it is not trivial to select a huge universe of hypotheses to winnow. There are always an infinite number of hypotheses that will fit any data. What makes science amazing is that we somehow pick hypotheses that not only fit current data but also predict new surprising results. Special relativity was strongly motivated by Maxwell's equations implying that light has a constant speed.
Right, many people believe quantum mechanics really does describe a random universe, but who knows. Personally I find it hard to believe that all the information in the universe was deterministically contained at the start. Randomness easily solve this problem.
I said 2100 which is around a century in the future. We have predictive capacity on climate from century to century and we have historical records going back millions of years to back it up.
I can have two random variables that have very similar means and huge variances. A few measurements of these variables are going to be all over the place. However, with enough measurements I can say with high confidence that their means are significantly different.
The statistics are clear (assuming you believe the measurements.) Temperatures are rising. Why and what impact it will have I can't really say except by appealing to experts.
That's an interesting point. I don't know much about the models, so I can't say if they can be improved in this sense. However, stability is a good thing. Hotter or colder is different and will cause problems for things like farming. Also more energy in a system will probably cause more violent weather which again causes problems, and average temperatures will effect sea level...
I guess your talk of null hypothesis and other terminology makes me think of statistics. And there is an argument to be made about repeatability in science...
Not to quibble, but I assume you mean you want to show your hypothesis is "better" than the competing hypotheses.
I think you mean chaotic. The word stochastic is an adjective that describes something that was randomly determined. One of the interesting things about chaos theory is that traditionally deals with deterministic systems. While I'm not a mathematician, I'm sure there are connections between chaos and randomness. I guess a pseudo random number is a chaotic system that is used to model a stochastic system...
This gets into the whole weather versus climate. While this is one reason it is difficult to predict the weather, I can still say with high confidence that it will be colder in NYC on Dec 25 2100 then Jun 25 2100. (What confidence means in this context is something you could question.)
As for a chaotic system being predictable, it can be qualitatively predictable. There are things like attractors, and I presume one could build systems with inputs to move in the phase space from attractor to attractor. (Not sure why an engineer would want to do this.) There are very credible models that the brain is a chaotic system which is a sophisticated machine that can be controlled by at least the owner. (Think about that :))
I'm not quite sure I would 100% agree that Astrology involves studying how the universe works, but I wouldn't fight someone over the claim. I would just claim it's bad science. Science is on a quality spectrum.
Now I don't want to harp on semantics. You kind of asked for my definition of science, and I gave it. I think it's a pretty standard definition, but of course like most words, science has many definitions. You want to define it in terms of a precise meaning of the scientific method. That's fine, but you also seem to want to imply things that don't fit your definition have little value. No, you need to argue the specifics of why they have little value.
Well the results are better when you factor in how the cost of electricity is rising. I assumed 1.02% is reasonable, but someone here commented it's more like 1.08%. That would mean that a system that generated $100 of electricity today is generating $1000 after 30 years. Very unmortgage like.
I'm not quite following your terminology. Are you trying to ground things in statistics where AGW can't do repeated experiments.
Stochastic does not mean everything is unpredictable. It's colder every winter for as long as I can remember :)
The scientific method does not define science. I would claim it's a convenient tool to do science. If someone argues your hypothesis is wrong, you can explain how you tried to follow the scientific method to justify some of your arguments.
Sure, I would claim science is the study of how the universe works.
It seems you are trying to do guilt by association. Science involves coming up with arguments to show your hypothesis is right. Society has a long history of evaluating these arguments to determine if they are true. Even for something as solid as math (which is not science), old results have been discredited because people found flaws in the arguments. Just because people use tricks to argue for things that aren't true doesn't mean all arguing/persuasion is "unscientific".
I understand the philosophy a bit and spent some time looking into that perspective, but I didn't find much utility...
Another way to interpret (some of this) is to say that the scientific method is a simple way to make a convincing scientific argument. It's a useful template or pattern that simplifies evaluation. That's great, but that doesn't mean it works for every problem in the universe. Perhaps one needs a different template or even an ad-hoc argument for some issues.
Just because it doesn't fit your definition of science doesn't mean it doesn't have value.
Do you know of any good tools to do this calculation? I've been using the on-line mortgage tools. (Google has one pop up when searching for mortgage calculator.) This allows one to depreciate the investment (maybe set the value to 0 after 30 years.) What's missing is a way to inflate the payments. (Which actually makes a lot of sense for a real mortgage.) Do I need to write my own tool? I looking at ground based heat pumps and solar roofs.
In hindsight, I should have been more specific and asked what you think about AGW that is not science.
However, to keep things interesting, some might claim that climate models predicting 50 years into the future are not scientific. While they are somewhat falsifiable, the empirical results will be limited in generalizability and statistically weak.
While I don't know much about climate science, on principle, I would disagree. One can make meaningful predictions based on combining well verified components and using independent historical evidence to justify the models. Of course this has problems, but so does all science. One can disagree and focus on a philosophical definition of science, but real "science" is about making a convincing argument using the many tools at our disposal.
I assume you are fine with that, and you seem to be focused on a convincing argument to connect increases in CO2 to human activity. I already think there's enough evidence for that claim, but it's fine to strength the claim. Good luck.
This is a big topic, perhaps you can ground things with climate science.
I guess I'm OK with that definition of provable. My instinct is to reserve provable for math, and your original statement didn't give a lot of context...
That's not what I meant, but it's a reasonable point. Should we throw out all science that doesn't fit your definition of provable? For such a strong reaction, one would need to give a much more precise definition of provable. Do we need to do "controlled" experiments? How often do they need to be repeated? Is evolution provable? Is astrophysics provable? How about science that deals with the human brain? I'm guessing good, useful, and honest science can be done without reaching your standard of provable by combining a range of evidence. I would say that this includes current research into climate.
No model is provably right. However the behavior has a solid scientific explanation based on feedback loops and is not controversial. https://skepticalscience.com/c...
It's difficult to define better. Evolution is slow, and our thinking choices might be better (which is why we evolved the big brains.) If need to get somewhere in my car, but I'm tired, the smart move is to take a good stimulant. We haven't evolved enough in the last 100 years to "know" it's bad to fall asleep at the wheel.
So you're saying Milton Friedman was a communist. https://medium.com/basic-incom...
It's probably good enough given the typical noise floor, but I wouldn't say it's a perfect representation. Definitely overkill for audio, but I'm sure scientists need lots of bits.
You keep confusing the marginal cost to make a car with profit margin of the entire company. They are different things. If you want to make your argument, you need to say why the marginal cost of making a Tesla is wrong.
I curious as to how many of these level 3 chargers are in the US. Is it comparable to Tesla? I think a good charging network is a big factor in selecting an electric car. Tesla has lots of stations and is already at 120kW and is planning to go to 350kW. Is there a comparable option for other manufactures?
I assume that testing is based on using the same distribution as training. (The easiest way to do that is to take a big set of examples and randomly split it into train and test.) Humans do well even with changes in the distribution which is much more realistic for many real world tasks.
There is fascinating research to suggest that video (maybe even two channel video) is an essential part of human vision training. http://www.sciencemag.org/news...
Competition definitely needs to be part of the decision making process. However, I'm pessimistic there would much competition in the current proposals. For example, for a bridge there might be no realistic alternative. Even when alternatives exist there is little incentive to keep prices down when the sum of the demand is comparable to the sum of the capacity and there are significant barriers to entry to create new capacity. It's not like one company can lower their price and take all the traffic; they would get overloaded. It's more likely the companies would optimize their collective profits by matching price increases creating implicit but legal collusion. Corporations hate real competition.
Don't forget the profit. Since much of this infrastructure will have monopoly level protection, they can charge what the market will bear without real competition. It's a massive giveaway. A real market needs competition not collusion.
To me it seems hopeless. The AC got 6 up votes, but most of what he says doesn't make sense. It's an important skill to convince people one knows what one is talking about, unfortunately this skill seems independent of actually knowing what one is talking about.
The people, who own the machines would be fine. They would just trade what their machines produced. If the machines are complex enough they wouldn't even trade, they could just get the machines to make what they want. It would become the ultimate information economy since you would still need ideas for new things to make. This is what would have value. So I guess one could still get by being a worker, but I don't see a lot of jobs. As usual, it's best to have capital in a capitalist economy.
I don't understand this. The whole point is that it will be very hard for everyone to contribute. The people who could not contribute in that society would need to form their own economic society, but it's not clear how they would "pay" for things like resources and land to run it. It's a bit like the movie Elysium.
Back to your point, I think some simple mathematical models can show how wealth concentration is inevitable even in a presumably fair society. Freedom has it's price.
While agree with what you say, I would also so it's uncontroversial that this optimal rate is a fickle things. It changes over time and depends on things like the current rate, so probably whatever rate you pick this year is not optimal next year. However, I think your larger point stands and it's not worth dwelling on how to optimize since it's probably not the right thing for the government's tax goal to be be maximum revenue.
Yes that is the most important question, though I side stepped it. I really don't know the penalties and rewards for the ISPs not doing enough "broadband".
I would assume anyone who could upgrade would be considered covered even if they chose a cheaper rate. (I guess there must be some type of rule on price otherwise an ISP could offer 1Gps at 1 billion just to satisfy this requirement.)
Perhaps I'm too cynical, but I assumed the ISPs want to lower the minimum. If it's free money for them then they should want a huge minimum. Are you saying the FCC is going against the major ISP wishes?
I'm not sure it makes sense to use old habits as a reference point. Netflix 4k needs around 20Mbps, so by that metric 50Mbps would be about right. (While 4k is not useful on most tvs, it's nice on a big computer monitor from 2 feet away.)
I'd rather they keep 25Mbps and check it. Just enforce the existing regulations instead of weaken them.
However, the biggest defect in this FCC proposal is the wireless aspect. Most unlimited wireless plans tap out around 22 GB per month. At 10 Mbps that's about 2 hours of full speed data a month. So you can watch about 4 hours of your HD Netflix or 1 hour of 4k Netflix.
Yes it could be bad, but I would like to see some reasonable experiments to see if it could be viable with decent regulation. At least it has the advantage that it's not too hard to go someplace else.
Of course, it's easy to come up with exceptions to your rule, but I don't think that's the point. I assume you are trying to give an intuitive argument for something. (I'm not sure what that is.)
My general point is that economic theory has been a big failure. It has very little predictive value. However, it is obviously important and worth studying. I just wish we could extract the value from the garbage.
Very true.
You go to the auction to buy groceries? I think the intuitive idea is that there is a fixed size pie. UBI gives some people a larger slice of that pie than no UBI.
This is not a problem with UBI. This is a problem with housing in SF. If I was an enterprising business man, to profit off UBI, I would create a community in an area with low property costs. I would bet that some of the people with no good place to live would relocate given that they had no job to tie them to a location.
Note: I'm not saying this community would turn out great. (It could be a nightmare.) This is why one needs to do real experiments with UBI instead of trying to apply naive economic theory to it.