> Tesla had to do a massive recall of basically every car they sold in China

By that logic Microsoft is recalling Windows many times a month and my Linux is being recalled daily.

What Tesla did was just a software update that added new features. In the context of the contemporary car industry, 'recall' should be split into more precise sub-terms.

Epigenetics = Runtime Configuration Settings

Not at all.

Astronomers were mostly concerned with the planned sats in 1000km orbits, but those would be deployed at 550km as well, so that is no longer an issue.

The Starlink sats will also feature sun-shades starting with launch 9 (in addition to other measures) and that will reduce their albedo greatly as well.

Despite the noise, astronomers do not seem to be too worried (this is just one of the articles). Here is a quote for example:

> "For a 100-second exposure time at twilight, we would lose 0.3% of the exposures. So, that means for every 1,000 exposures we take, three of them would be ruined by a satellite. And that number goes down as the Sun goes down,"

Some observations like long exposures on wide fields of view would be affected, but there is work going on to mitigate that too.

Well... Let's check the news to see if this is really the case:

I guess there will be solar production jobs in the US after all. There will be many more related to solar installation and servicing, of course.

There are some issues with these calculations.
1) The per pound price:
- The prices you used are per kilogram, not per pound
- The prices do not take into account the first stage reusability that will presumably become standard by the time the sats are launched
- If we use the Falcon Heavy costs with reusability (e.g. from here) we get $50mil/(0.7*119930) = about $600 per pound.
- The $600 price per pound includes the SpaceX profit margin. If that is not taken into account the price would be even lower.

2) SpaceX will first launch only 1600 sats to make the system operational. From then on the future expansion can be funded by the operational profits.

Given the above calculations, the launch prices for getting the system to work will be 1600*850*600 = $816 million
That is well within the SpaceX financial capabilities.

Now, the above assumes that the FH launches of the sats would be mass limited, rather than volume limited. I suspect that in reality they would be volume limited, however, thus the price would be higher. In any case it would be much lower than your original estimate.

There are several data points available with which one can make at least approximate calculation:
- The price of a Falcon 9 launch is $62mil (was expendable up to now)
- Gwynne Shotwell said that if first stage can be restored for $2mil, then the reflight would cost $40mil
- The cost of the first stage is known to be around 70% of that of the rocket
- The cost of the rocket is only a part of the F9 launch cost. Let's say that this fraction is R.
- Given that a reflight uses again everything else other than the first stage, from the above one can calculate:

Reflight ($40mil) = Stage 1 restore ($2mil) + everything other than Stage 1 ( $62mil*(1-0.7*R) )

Taking into account the uncertainties, this gives you that the cost of rocket R is between 55% and 60% of the full launch cost.
Having these numbers one can make a ballpark estimate for the cost of FH and the FH reflight. Those are the numbers listed in my previous message.

I have not looked at the patent application, but based on the description, it appears to me that Apple is trying to patent something similar to what WiTricity is doing, rather than what Tesla has demonstrated. From the WiTricity wiki:

Ok. I am not going to start refuting the details as this is not the forum for that. Let me just summarize what you are saying: Austrian economics predicts that inflation will rise eventually.
You see, this is not a testable prediction. It can be said for _anything_ that will happen eventually and that would be almost certainly correct. The question is what would happen now and in the next few years at most.

In the case of inflation Austrians started predicting that inflation would rise substantially about 2 years ago and many people made big noises about it 1 year ago saying that it would happen imminently. Result? The inflation is now _lower_ than it was back then.
This is sufficiently long time to test the validity of the predictions. Again, only the liquidity trap theory predicted _exactly_ that. The current budget impasse will likely muddle the waters and the effects are hard to predict in detail, but the situation will probably not change significantly soon in terms of inflation.

Ah. You study at the University of Chicago or something, I guess? Such opinion is extreme ideology and is hard to take seriously. For example:

> Government spending does not "stimulate" the economy.
I see. I suppose that is why the major banks downgrade their GDP estimates as a result of the prospects of decreased government spending? And why the UK economy nose dived as a result of the austerity package?

> The value of a theory is measured by its ability to predict... yet Keynesians have never predicted any major economic events... even though Monetarist and Austrian economists have.

Heh.
Monetarism quite correctly predicted the stagflation. It is failing miserably in the current situation of liquidity trap, however. It predicted that Japan will recover 10 years ago after increasing the money supply, for example. Nothing happened. The effects of QE2 were far smaller than predicted, etc.
Austrians: Even Milton Friedman did not think that theory had much to do with reality.

Also, all of those theories were predicting massive increase in inflation and/or long term interest rates due to the economic policy in the past two years. What happened? The interest rates are at historic lows instead. That's a massive failure of the predictions.
Only the liquidity trap theory predicted what happens accurately. And it is a consequence of the Keynesian theory.

The way to fix the economy in a normal, non-liquidity trap situation (i.e. almost always) is through monetary policy exclusively, no question about that.
Monetary policy does _not_ work well at this very moment, however. A fiscal stimulation is needed to get the economy out of the swamp and get the interest rates above 0%. After that we can revert to monetary policy again.

I thought the judicial system has the monopoly on that!

During 2001 the IPCC made a number of predictions as to what would happen as a result of the climate change. At the time their results were widely mocked and ignored by the "climate change deniers" circles.

It now turns out that the actual effects measured today are _worse_ than what was predicted. For example, the rise of the ocean level is 80% greater.

I think people should concentrate on the larger picture -- the predicted effects are happening. The whole CRU emails issue is peanuts and only diverts the attention from the real issue, even if we assume that everything that is being claimed there is true.

The vCloud API (http://www.vmware.com/solutions/cloud-computing/vcloud-api.html) is a protocol for placing VMs into a cloud, managing them, and downloading them back when needed. It was published at the last VMworld using a very liberal license and has been submitted to DMTF for standardization. While it has been initiated by VMware, the vCloud API has been designed to be implementation-agnostic and could be implemented over Xen or other hypervisors.

The whole idea is for people to be able to download the data/VMs that they have in one cloud and move it easily to another if they desire to do so. In short, this is the opposite of Hotel California.

> As the idea about competition in architecture... well, the ones who can't code :P should also have their competition, so why not? :P

Heh. I have awards from both IOI and ACM (gold medals, etc) and have significant industry experience, so I feel that I know what I am talking about :)

Most programmers apparently believe that the best way to make a scalable application is to (re)write it in assembler. The responses to my post seem to prove that point. Perhaps architecture/design competition is more necessary than I thought.

Software Development consists of several relatively independent skills:
    - programming (knowing how to use the tools)
    - algorithms
    - architecture and design
    - knowledge of processes (development methodologies, etc)
    - enabling teamwork (allowing many developers to work together)
etc.

The IOI competition is for high-school students and tests mainly the 'algorithms' aspect.
The ACM competition is for college students and tests mainly the programming aspect. (strange, one would think that the aims of those two would be reversed)

There does not seem to be a big competition for testing the architecture and design abilities, although arguably they are even more important (unless you count the Real World competition). Part of the difficulty perhaps is that it is tricky to come up with an objective measurement. An approach that I have been using is the following:
- give a task and provide plenty of time
- at 50% of the time change the requirements of the task slightly
- at 90% of the time change the requirements significantly
If proper design has been used, then making appropriate modifications would be easy and the task would be accomplished in time. This closely mirrors the situations in reality.