« Comment of the Year - 2020 | Main | Endings and Reincarnations »


Stephen Murray

Kartik, great article as always. Overjoyed to hear you will be kicking it up to the next step in your work. Will be keeping a close eye on it!


Falcon 9 delivers 22,800 kilograms to LEO for a launch price of $62 million. That leads to a cost of $2,719 per kilogram. Starship should be able to reach orbit for $300/kg. Which, given the likely path of the starship, is right about in line with your curve, predicted to be operational by 2025.

I'm not positive it can go a lot cheaper than that - certainly more launches will reduce the cost per launch. It could go to around $2 million per launch just for fuel, meaning your minimum cost might be $20/kg, but that is quite a reach. A more likely scenario would be $150/kg, which is still very cheap.

"Technology always increasing the supply of, or alternatives to, any commodity."

So true. And precious metals like platinum, gold, rhodium, iridium have amazing chemical and physical properties that make them valuable as industrial metals. If their price dropped significantly, we would no doubt have an incredible industrial expansion.

Rhodium, for example, is very hard, very stable and has a high melting point. It also has very low electrical resistance as well as a low and stable contact resistance. We don't use it for much because it is very rare and has a sky high price ($70,000/kg). Even at that insane price there is industrial demand for the metal. Imagine what it would be at $700/kg.

Regarding planetary probes you have is exactly right:

1) Launch costs are high, so we can only afford to send one probe to Jupiter per decade.

2) Only one probe? Then it has to be larded up with every imaginable instrument.

3) Because it has every instrument, the probe is very heavy. Meaning launch costs are even higher. Back to number 1.

4) Also every probe must be custom built. Making them even more expensive.

I disagree with one thing - New Horizons was pretty cheap overall. It represented a terrific advance in space exploration. Total cost was only around $700 million. The Cassini probe to Saturn (for comparison) cost about $3 billion. So Pluto was cheaper than Saturn, despite being 8 times further away.

But what Nasa should do is mass manufacture probes. Create a standard multitasking planetary probe, and build 50 or 100 of them all at once. Once that is done, then launch as funds become available. We lack so much basic information about so many celestial bodies. Just having long term observation of changes in various planets would be extremely valuable. Call it a standard mark one orbiter - every planet has 1-2 in orbit all the time, measuring and mapping. Every major satellite as well.

Look at the moon (Luna) - half the moon is rarely observed, because satellite coverage is uneven. Even today, in 2020, there are only four satellites in orbit around the moon, and one(?) active lander. Build 20 surface landers, and land them all over to provide continuous 360 observations of the surface. It wouldn't be hard - make them small enough that one Falcon heavy rocket can carry all 20.

Kartik Gada


A more likely scenario would be $150/kg, which is still very cheap.

That is true, and is still fine, when the electronics of Voyager 1 and 2 (1977, one ton each) are inferior to what is now found in a 200 gram Raspberry Pi setup.

I bet if the full schematic and reference design of Voyager 1 & 2 were posted online, someone could easily make a small equivalent all by themselves. It is just a camera, battery, solar cells, antenna, and transmitter, with some radiation shielding.

So Pluto was cheaper than Saturn, despite being 8 times further away.

Four times farther, to be exact. Pluto is 8 times farther than Jupiter.

But what Nasa should do is mass manufacture probes. Create a standard multitasking planetary probe, and build 50 or 100 of them all at once.

Fully agree, but it should be subcontracted to some electronics manufacturer. Given that all images are posted on the Internet (unlike back in the day, when I had to check out library books to see images that Voyager 1 and 2 took of the outer planets), billions of person-views of the now huge number of images would happen almost immediately.

It is sort of pathetic that there are only four satellites in orbit around the Moon, even in 2021. A simple Moore's Law application of electronics combined with low launch costs should apply.


Just looked up what is planned for the moon this year. Looks to be a very active year, with as near as I can tell, 2 orbiters, and 7 landers/ rovers. Oh, and 13 cubesats. Or is that 25?

2022 has 3 more orbiters and 3 landers.

Ah the singularity - you complain about something then turnaround and the problem is already solved. Seems like single launches are delivering multiple rovers and satellites.

My dream? One of those Boston Robotics dog like robots, nuclear powered, running across the moon at 15 mph. That would be fun to watch.


About having too few moon orbiters. There's the problem of diminishing returns. The first pictures are extremely valuable but having taking photographs of the same thing over and over again is not that appealing, especially for static places like bodies devoid of atmosphere. It is ironic that we had so many huge government pilot projects which fail to trickle down to cost effective private commercialization.

Supersonic airplanes. Cool but too expensive to be visible. Nuclear... Too expensive. Space... Without soaceX it was too expensive that US, the richest country in the world, abandoned human space flights.

My point is that without a commercial ROI those projects are too expensive for dinner virtue signaling.

The moon race shouldn't have taken place. It was fine for wrong reasons and than the program was scrapped.


Well, yes and no. Orbiting the moon - the moon is a very big place. Recent orbiters have discovered many additional features, and landers have been few and few between.

For example. We used to think the last time the moon was volcanically active was a billion years ago. Then we realized, no, it was 100 million years ago. latest estimate was 10 million years ago. Could it be 1 million years ago? 500 years ago? https://www.sciencemag.org/news/2014/10/recent-volcanic-eruptions-moon

Certainly there has been observation of mysterious phenomena on the moon. Flashes of light. Glows. Say, what is inside lunar lava tubes? How many are there? can we get in? Maybe colonize them?

You are right up to a point - if all you send is an occasional orbiter, that costs a few billion, that last a couple of years, you might not learn much new. But more data brings new discoveries. And the cost to collect and process that data has been falling rapidly with launch costs.

How about mapping the entire moon at the level of google street view? You saying we won't learn a lot of interesting things?

Boom thinks supersonic aircraft will work. Nuclear - SMRs are being deployed in Utah right now that could radically alter prices.

"Without Space X..." well you can play the game that without this we would never have that...but that is not how it works. Without Space X...we might very well have a different company very much like Space X.

SpaceX has shown that we were going about space exploration in a tired, old fashioned way. There would be no SpaceX at all without NASA COTS - NASA decided to change the way of paying for and creating space access, and ditched the cost plus mechanism that brought us Apollo, and moved to a more innovative funding mechanism, paying for goals achieved. Within 10 years it was a spectacular success, reducing costs to access space by an order of magnitude. because of that price drop, suddenly something like starlink becomes viable. Suddenly we can afford a dozen lunar orbiters. COTS has paid for itself in reduced space access costs 1,000 over. next came crewed access to space. That goal was just achieved. next is lunar landings. By 2025 no less. Then Mars.


I can see a drop of water with my eye, why pay for a microscope? A drop of water is boring. I've seen a million of them.

Now that I have a microscope, I can see more. I can see bacteria. Why get a better microscope? What more could I possible learn?

Now I have an electron microscope, do I really need something better? It's just a drop of water. Sure now I can see even smaller structures, but how much more could I learn?

Science often advances by gathering more data on finer and finer details. Things tend to get more interesting as our resolution of the object increases. To paraphrase Richard Feynman, there is a lot of room for interesting things at the bottom.

Planets and moons are very big places. Our resolution is quite poor at the moment. Small increases in our resolution will likely yield many important discoveries - there is a looong way to go to the bottom.

Kartik Gada

All true.

Surely the cameras in the newest smartphones are far better than what Voyager 1 and 2 have (given that they were launched in 1977). Since both launch and electronics costs have fallen so much, we really ought to have multiple cameras around every significant body in the solar system. Every object down to 500 miles in diameter.

Cost plus was absurd, and unfortunately still exists in most parts of government. Robert Zubrin wrote about this as far back as 2002, and anticipated that removal of this is all it would take for private space companies to exist.


Universal basic income doesn’t impact worker productivity.


Providing workers with a universal basic income did not reduce the amount of effort they put into their work, according to an experiment conducted by Spanish economists, a sign that the policy initiative could help mitigate inequalities and debunking a common criticism of the proposal.

Examining a universal basic income worth about one-fifth of workers' median wages, the researchers also found that the threat of being replaced by robots did not impact workers’ productivity, nor did a tax on firms when they replace a worker with a robot or automated process, though the latter successfully created a disincentive for managers.

Kartik Gada


Good find. Yes, UBI ought not impact worker productivity (as long as it is UNIVERSAL, which is the key word). It does not have the perverse incentives that means-tested welfare programs have.

A tax on firms that replace workers with robots will never work, as such a corporation is very easy to move elsewhere. Rather, I contend (as we all know) that there should be no income tax, so that all the energy spent on game playing to lower tax, is not wasted at all. QE will be enough to cover the expenditures that tax currently funds.


My 2c.
I would say that such tax on robots can be circumvented on technicalities. If it is a tax that replaces workers there will be a new entity that replaces a whole department. If it is on the number of installed robots you will see megarobots with many appendages or devices that don't qualify as robots on some technical definition.

Kartik Gada


I would say that such tax on robots can be circumvented on technicalities.

Of course. That is why I have said for many years that the long-term trend is towards no income tax on humans, or corporations owned by humans. It just will be harder and harder to enforce, and arguably already costs more than the dollars collected.

A consumption tax is more durable and fairer, but even that is not going to be necessary forever. Monetary creation is on the verge of being enough to render at least income tax as (technologically and economically) unnecessary. Political constraints are another matter.


Much like a gas tax for supporting roads. As EVs increase in sales, and ICE cars get better mileage, the concept of taxing gasoline to pay for road maintenance is becoming an increasingly untenable concept.

What is a robot? When you start to think about it, the line is getting very blurred. I'm old enough to remember having typists type up reports that I wrote. That position has been eliminated. Is my laptop a "robot" that saves "labor" and therefore should be taxed?

This is one of the many problems with the singularity - old forms of stable government management of things becomes increasingly unstable and unpredictable. Whatever scheme you dream up to outwit the singularity will only work for so long before the singularity generates a work around. Your only hope of collecting any taxes is making taxes so low they are simply a nuisance, these avoiding the hungry gaze of the ravenous ATOM.

Low, simple, and universal taxes are not a preference, they are a necessity.

Kartik Gada


What is a robot? When you start to think about it, the line is getting very blurred. I'm old enough to remember having typists type up reports that I wrote. That position has been eliminated. Is my laptop a "robot" that saves "labor" and therefore should be taxed?

Yes, which is exactly why so few people understand technology (including many people working in the industry).

A robot, with many mechanical appendages, is actually far too information-inefficient (and materialized) to have emerged quickly. Here is a clip from Tom and Jerry from way back in 1952, where a robotic cat that can catch mice was portrayed as something that anyone could order via mail order even at that time. But something like this does not exist even in 2021 :


'Robots' have been 10 years away for 70 years.

By contrast, software, particularly in the cloud, cannot be partitioned into 'individuals'. If a certain type of job is eliminated by cloud-based software, and 1 million workers are displaced, there is nothing close to 1:1 matching of the individual to one one-millionth of the indivisible software in the cloud. And where is the cloud? That is a vastly higher ATOM material efficiency, and also completely borderless.


Well said.

It is not a one for one replacement, it is one for a million. And it is disbursed. You might as well try and tax raindrops.

An no one considers the following scenario - I hire a guy in Bangladesh and using modern communications I have him teleoperate a machine in Illinois. I can't tax the machine - it's not a robot. I can't tax the Bangladeshi guy, he not in my jurisdiction and could argue he is doing the work in Bangladesh. Do I tax the communication link? But that is the same as million of other links being used - how do I call it out?

Taxing robots is a foolish quest.

Kartik Gada

Coming back to space :

When you take a 100X decline in launch costs, and a 1 million X decline in the weight of electronics for a given functionality, only the government would keep the basic paradigm of space probes unchanged for 40+ years, rather than adjust to the 100 million X cost improvement. It is truly shocking, when you think about it.

I think SpaceX will probably just do probes on its own at some point. As mentioned above, Voyager I and II as a small IC can probably be produced in no time at all, and SpaceX can just launch them and start collecting images by the billions.


I was just thinking about scale and how it has two benefits - financial, in the obvious of more money for doing more, and the productivity increase that often comes with doing more and learning. But of course there are areas that don't see that productivity increase with scale - including our favorites of government, education and medicine, probably because there the feedback mechanisms are reversed in cost-plus systems.

On the space side, I'm also surprised that no wealthy donor - such as Paul Allen's estate - isn't simply sending up lots of cube satellites to gather images from earth orbit and eventually beyond. Even if using cell phone cameras, I'm sure that using multiple viewpoints to combine the views would also increase our knowledge greatly.

Kartik Gada


All true. This is one of the reasons that the EU does not work. The entire project was a reaction to being envious of the scale of the US, but sectors that worsen with scale become powerful by the time the EU had begun, so the entire vision was stillborn.

As far as images from space, yet another dimension is that the productivity of image access is easily about 1000X greater than in the 1980s, due to the Internet, enabling consumption of thousands of times more images across a larger group of people. As I said above, back in the day, one had to go to the library to check out books that might have images you wanted to see (and only a few countries even had/have decent libraries to begin with).


Imagine google street view. For the moon. Just being able to wander around virtually. At some point, forget science and start thinking in terms of entertainment.

In 2020 the total known asteroid count passed 1 million. We know most (>=90%) near-Earth asteroids above 1 km - there are about 1,000 of those. Those are planet killers.

The next goal is to find all the near-Earth asteroids over 140 meters. A 140-meter asteroid will impact Earth on average every 20,000 years and will unleash 300 megatons of energy, causing regional scale devastation. Around 15,000 such objects have so far been discovered.

The flip side is also interesting. A 140 meter asteroid is small enough we could tow it around and even get it into lunar orbit. Just 10 cubic meters of an S-type asteroid contains about 650,000 kg (1,433,000 lb) of metals. That is just about the weight of the international space station. M-type asteroids are rare but contain up to 10 times more metal than S-types.

So we have something that is both potentially very dangerous, and potentially very valuable. If such an object was heading toward earth you could rendezvous, but instead of pushing it away from earth, push it into a stable orbit of the moon, or one of the Lagrange points. Then send up a team to mine it, carving up chunks and returning them, ten cubic meters at a time, for processing. No need for astronauts - just use teleoperated robots.

Want to build the big orbital space wheel? That is how you do it. It wouldn't take too many 140 meter rocks to build such an object. And again, there thousands of such rocks.

And once you build one, why not a dozen?

And imagine each Space X starship returning to Earth with a couple of tons of gold and platinum. Left overs from habitat construction. Every 10 meters of a type S asteroid contains around 110 pounds of gold and platinum.

The point is there is a way for launch costs to "go negative". That is to say, each launch brings back more money in precious metals than the cost to launch. We would no longer much concern ourselves with reducing launch costs.

Kartik Gada


The next goal is to find all the near-Earth asteroids over 140 meters.

Almost all of those are known at this point as well, with no known impact expected before 2036 (and even that won't actually hit). Recall this :


I think that the majority of thought has already moved away from impact deflection worry and towards eventual mining. Mining is an ATOM progression, just so that all those rare metals become common and commoditized, enabling industrial applications that one could never have thought of using Rhodium, Palladium, Platinum, or Gold for.

A 1 km asteroid has 150 billion tons of metal, and asteroids with a high ppm of gold and platinum have already been identified.

But if 60m asteroids just burn up in the atmosphere, perhaps we don't have to let those go to 'waste' anymore. Perhaps a large parachute can be affixed to it, so that it neither burns up nor impacts at high velocity. It lands softly and can be mined.

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.


Post a comment

Your Information

(Name and email address are required. Email address will not be displayed with the comment.)