Hacker News Reader

Comment by wongarsu | Hacker News Reader

On the ISS you can produce higher quality optical fiber than on earth [1]. A demonstration fabrication unit is already on the ISS, and NASA has recently published prices for using the ISS for commercial tasks such as this. I suspect many other manufacturing methods can also benefit from zero gravity, and as launch costs drop more of these will be explored.

1: https://upward.issnationallab.org/the-race-to-manufacture-zb...

View on HN

It's interesting to think about that from the other side, too. Just delivering the raw materials to the manufacturing facility would cost ~$3000/kg (assuming rock bottom SpaceX prices). Based on some loose googling, I think that 1kg of raw material would yield only something on the order of 100m of cable. It looks like a typical 100m fiber optic patch cable retails for about $60.

So, there's clearly a lot of value to somebody for fiber optic cable with even lower signal loss.

That really drives home to me that, even though information technology is just my boring day job and it's still space travel that gets me as excited as it did when I was a kid, it's still true that the space age is over and we're definitely living in the information age.

If you didn't have to source your material from Earth, it would be a lot cheaper.

Orbital manufacturing makes very little sense if you source material on Earth. You need to get your raw material from the Moon or a captured asteroid.

Which is why this fiber optic thing is a big deal. They're demonstrating that you can build useful stuff in space that are a) already have a demand, and b) are infeasible or impossible to manufacture downwell. More attention = more funding, more attempts... until, I'm hoping, we'll reach critical mass and someone will start mining near-Earth asteroids, or perhaps capture one and move it a bit closer to us, at which point the prices of space material will drop. = more funding, more attempts, until someone figures out how to make stuff useful for space missions in space, and this will be the true beginning of a self-sustaining in-orbit economy.

So, no idea the cost of O-O amplification or O-E-O these days, but it's expensive. You have to run power (usually there's power with the optics), and the bigger problem is real estate (At least in urban areas). 100x lower signal loss would do away with most of these.

For example, the MAREA (https://en.wikipedia.org/wiki/MAREA) undersea cable is 6,600 km long. Current optical fiber requires repeaters every ~100km. If it's to be believed, ZBLAN could get rid of these repeaters all together.

The cable's design is _much_ more complicated because of the requirement to have electricity and repeaters.

Presumably lower loss also means more bandwidth when background thermal noise becomes the limiting factor.

Lasers. This stuff may be necessary for a new class of weapons. That is the reason for all the speculation.

> I think that 1kg of raw material would yield only something on the order of 100m of cable.

That sounds like you’re including the weight of the whole final cable, including yarn strengthening members, plastic shroud and sometimes a braided metal shield layer. Only the glass fibre itself actually needs to be manufactured in space and is usually on the order of a kg or so per km.

For the time being, they are targeting specialty applications where it is not uncommon to charge in excess of $1000 per meter even for normal earthly fibers.

For scaling up, if one uses the smallest commercial fiber standard (80um diameter cladding--they would put the buffer and whatnot on after it comes back to earth) the mass of the fiber is only 53grams/km so using your $3000/kg the cargo cost is only $0.16/meter which seems pretty palatable to me.

What’s the application of super high-quality earth cable?

That's kind a obvious: high-frequency trading. Thousands of banks and companies will happily to pay premium to get some edge over competitors. As soon as single entity on market have it everyone else would want it as well so there will be plenty of demand.

is asteroid mining viable compare to shipping raw material up the earth gravity well?

Definitelly - just has a high startup cost and possibly takes some time to get the material back.

High frequency traders will want the best connectivity humanely possible. It’s a zero sum game. If they can beat a competitor firm’s signals by the smallest fraction of a second, even a nanosecond, they reaps in the profits. So $100m for some better fibre? Go nuts!

Eh, that's a really sad thing to optimise for, don't you think? Revealing of a broken incentive system. I'd prefer if HFT practices were banned.

Can you be more specific about what is so broken about the need for efficient and fast trade?

Volatility is the main concern [1]. Making trading decisions based on microseconds is speculation. How can anyone or anything reach an understanding of something as complex as markets in this short time?

Other concerns include unnecessary waste and unfairness [2]. The point is that this is a zero-sum game that makes us invest in infrastructure that will soon get obsolete once someone else builds a faster infrastructure.

Fast trade isn't "efficient". Long-term stability is.

[1] https://en.wikipedia.org/wiki/2010_flash_crash [2] https://en.wikipedia.org/wiki/Flash_Boys

It is a speculation, yes. That does not mean it is not useful, or is harmful to the society.

How is it useful?

People that need to sell something will find a buyer faster.

> How can anyone or anything reach an understanding of something as complex as markets in this short time?

Computers mostly.

No you wouldn’t. The development of HFT has created massive liquidity for markets and reduced bid ask spreads in a way that benefits retail investors more than past alternatives.

Does it benefit society though? I'd argue that increased liquidity in markets is a net negative. That means it's easier for investors to sell their assets quickly, which is exactly what you don't want if you want to encourage long-term investment.

No, if people need access to their money they need to be able to sell off their investments damn near instantly. This isn’t real estate.

People buy and flip houses, despite the fact that it takes months. What timeline would you think would discourage short-term investment?

Ultimately I'd rather place the responsibility for long/short term investing on the investor. The fact that most retail investors should be investing passively and long term is well known at this point, the issue is one of education.

Don't dumb down the system when it's easy to make people stronger.

> People buy and flip houses, despite the fact that it takes months. What timeline would you think would discourage short-term investment?

Months would be pretty good. But I suspect that wouldn't fly in practice. Flipping houses is long-term investment, no? Even if the asset isn't held for a long time, long-term value is added. This is very different to speculative trading where no value is added.

> The fact that most retail investors should be investing passively and long term is well known at this point

Nope! The primary purpose of the stock markets isn't to generate income for the investors. The purpose is to effectively distribute capital to those business that can use if effectively. Passive investment does not help to achieve this. It reduces the "wisdom of the crowd" effect, because anyone investing passively is not contributing their wisdom.

> Don't dumb down the system when it's easy to make people stronger.

I have completely the opposite view on this. If there's an obviously correct way of doing something, then don't leave it to chance. Design the system so that people can't fail.

Flipping houses is also speculative, as those who have lost money doing so can attest to. If you spend a lot of time/money flipping a house only to have the market tank or you just chose a bad area, then no one buys and all that "value" is lost in wasted resources.

Buying stock by definition is purchasing the right to participate in a company's future profits. That's the reward for taking a risk on the company, and the primary metric by which "who can use it effectively" is determined. It isn't and has never been some board of financial justices trying to figure out how reallocate capital out of altruism, otherwise hedge funds and active traders would all be non-profits.

There isn't an obviously correct way to do this, it depends on one's objectives. Financially illiterate employees contributing to their 401ks have different needs than hedge funds who have different needs than people trying to build wealth for non-retirement time horizons in taxable accounts who have different needs than people saving in HSAs and 529 accounts who have different needs than Pension funds, among many other users of the market. It's up to the investor to determine their needs, not be coddled by some nameless, faceless system that only permits the lowest common denominator of participation.

On a more philosophical note, systems where people can't fail are systems I've experienced professionally (working for companies with political/economic moats so vast they can fail repeatably and not be punished). All it produces in my experience is complacent dead weight and rules to maintain said dead weight that unite to stifle the progress of those who wish to perform or make progress. The systems become benign tumors at best.

Now granted there should be some safety nets in place, not everyone can be or wants to be a high performer, but that doesn't mean we should make every room with padded walls. People who throw money into the market without the slightest bit of research should be punished for their lack of planning, same way people who jump off a roof without thinking are punished by injury.

Personally I waffle on this myself. My instinct tells me it’s bad but then I try to decide what trading cadence I would limit it to. Assume there is a CEO who wants to do something nefarious to manipulate the stock price. If they make a decision how fast can a company actually act on it? Would or could they make the company act on 1 hour boundaries? Even if we limited trades to once per week is that the long term we want a CEO to consider? At what time frame does a stock stop being a stock and convert to something else? Then I decide even by minute trading is not really something that management can even consider, even if they want to, so there is really no point to artificial limits imposed by law and take a more laissez faire attitude.

Don't limit cadence. Just tax transactions, and cadence will limit itself.

I waffle on this one too. Institutions that do HFT have a huge capital cost and pay large fees before they even make their first trade. They also employ programmers, traders, etc. Basically, these taxes are going to need to sum to something really big to dissuade them. Too small, they ignore it. Medium, the market will just consolidate. Large, you'll stop HFT, but at what consequence? The probability of unintended consequences is high.

This is when I decide I don't really understand the market as well as I think, and I should stop solving the world's problems, and go back to designing circuits.

A tiny tax on each order (bid, ask) entry would cut the total volume of orders by a huge factor, cost them only a small fraction of their takings, and probably fund free college for everyone in the US.

High frequency traders do NOT want to be limited by the speed of light in a fiber - due to its refractive index (1.5 for ZBLAN according to Wikipedia) this is only 66% of the speed of a radio signal. Over 10km this amounts to an 11 microseconds difference - already significant for HFT.

This is why they have been leveling hills to carve a (mostly, with minimal amount of repeater towers) line of sight path between NYC and Chicago, gaining 3ms (1), or set up a line of sight laser link between NYSE and NASDAQ to eke out a handful of microseconds (2), or set up multiple links across the English channel to reduce latency between London, Amsterdam and Frankfurt (3).

1. https://www.nytimes.com/2014/04/14/opinion/krugman-three-exp...

2. https://www.extremetech.com/extreme/176551-new-laser-network...

3. https://sniperinmahwah.wordpress.com/2016/01/26/hft-in-the-b...

I've seen speculation that low latency may be a big part of the business plan for Starlink.

It should be faster for long distances than cables for sure. Also less routing.

One could use hollow core fiber, which has much higher limits. Up to 99% of light in a vacuum IIRC.

A wave guide?

Sort of? It came up ocassionally here:

https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...

How much of that material needs to come from Earth, though? If we can get that from the Moon, for example, then you wouldn't need nearly as much fuel to get the raw materials into orbit for manufacturing. Then all you need to do is get the finished product to wherever it needs to go (e.g. in an atmospheric entry vehicle, or to some other orbital factory, or to some far out colony, or what have you).

Long term - definitely! But to get space manufacturing started in the short term, lowering the launch costs is the best bet.

Agreed 100%. The more we can lower costs getting from Earth to LEO, the cheaper and faster it'll be to bootstrap that manufacturing infrastructure.

> That really drives home to me that, even though information technology is just my boring day job and it's still space travel that gets me as excited as it did when I was a kid, it's still true that the space age is over and we're definitely living in the information age.

Just the fact that the historical events the past couple of days didn't involve space, but something filmed on a mobile phone shared through information technology ...

> the raw materials to the manufacturing facility would cost ~$3000/kg

This is also assuming the price per kg to orbit stays stable but Musk has stated he intends to drive the price down much further. He's even given figures for what he thinks they might be able to achieve with Starship[0] and SpaceX's rapid progress seems to indicate that they will, eventually, get there. The best comparison from history is probably the construction of railroads which had long development timelines and capital costs but eventually began paying back massive amounts even while driving the cost of freight down and providing a backbone for commercial and industrial development.

[0] https://www.inverse.com/article/60712-spacex-starship-elon-m...

People used railroads to profitably carry freight (originally with carts pulled by horses) right from the get-go. They became a lot cheaper to make once steel could be economically mass produced.

Steam engines were originally used for pumping water, but there was an obvious huge demand for locomotives to move freight by land once they could be made cheaper or better than draft animals.

I’m sure there exist some commercial uses for space flight, but it really doesn’t seem all that similar to railroads in terms of obvious market demand or effect on the economy.

Yeah; I would compare railroads/rockets to microcomputers/VR. The latter is definitely impressive, and unique, and evocative for futurists, and has some niche use-cases, but doesn't have a clear "killer app". While the former had people - real businesses - chomping at the bit even before it became economical.

The points you bring up don't seem to support your argument. It took a decade from when people first attempted to haul freight with a steam locomotive until the first commercially successful railway was built. If you start counting from when the steam engine was first invented then it would be more than a century. You're correct that a lot of that success depended on improvements in manufacturing capability but I don't think it's so obvious that there was a huge market for steam locomotives right from the start, early steam engines were notoriously fickle, required constant attention and a trained engineer, needed specialized tooling and materials to repair when they inevitably broke, and they were incredibly inefficient and fuel hungry at first. That's why they started out so linked to coal mines, the mine produced enough cheap fuel for the engine and already had a lot of the metalworking tools and capable engineers needed to maintain it.

I'm not sure why you don't think space is commercially viable right now, companies and governments are paying a lot of money to have things delivered to orbit. SpaceX is making money with every launch otherwise they would have gone bust years ago. If you mean there's no market for human spaceflight, well there wasn't a market for passenger railroad travel for a long time either. You need to build the railways before people can make practical use of them. Back when most railways only extended for 1 or 2 km it was pretty hard to imagine getting paying passengers. You have to build something that connects to places people want to go before they'll pay to ride it.

With spaceflight we're currently in the shorthaul stage, all the profit is in delivering cargo to LEO. The cost is coming down and people are finding more uses for cheap satellites. The rockets already have their use case, they move people and cargo the same as a railroad, it's what we do with that capability that we need to get creative with and as we continue to build out capability and lower cost there will be more and more uses for space. The main reason I compare it to railroads is that it is a type of infrastructure that makes new things possible and there's always a profit to be made in a frontier as long as you're daring enough to take it.

Wait until some HFT investors will step up to buy such fiber cable to be microseconds faster than competition. Those $3000/kg could turn out to be a small change.

You know...

Launching the raw materials to orbit from Earth might cost $3000/kg.

While Earth is great for being able to sustain life, the gravity well rather sucks.

Launching the same materials from most other interesting places in the solar system would cost a lot less, (at least, were there people there to provide some sort of settlement)

For instance, presuming most of what you need is silicon oxide anyway; launching the raw material from the moon and transporting it to earth orbit easily costs MUCH less fuel than launching it from earth.

* Earth->LEO: 9400 m/s * Moon->LEO: 1730+680+3260=5760 m/s on rockets alone * Moon->LEO 1730+680=2400 m/s with aerobraking (smarter) * Mars->LEO 6300 m/s with aerobraking * Asteroid belt->LEO between 1060-3360 m/s

Moral of the story: Don't launch your stuff from Earth if you can at all help it.

Now, once you have a somewhat established moon settlement that can take on projects it gets even more interesting. Unlike on earth, Lunar gravity is so low that things like a space elevator or a linear accelerator are practical with even todays technology and materials science. Combined with aerobraking on the other end, you could ultimately ship materials or goods to earth for cheaper and faster than china post ship from china. ;-)

The only downside: first you do need that settlement on the moon, and that's one heck of an initial investment. On the other hand: once established and marginally autonomous, you can intuit that -despite harsh conditions- such a settlement could be quite a competitive location to establish mining and industry.

Reference (assuming this one is well made): https://www.reddit.com/r/space/comments/29cxi6/i_made_a_delt...

The company that sent that demonstrator does not intend manufacture at the station. The plan is, once the tech is proven, to launch a dedicated robotic craft. It would do the work and return on its own. That moots all sorts of docking/navigation problems, dramatically reducing the complexity of the craft.

How pressing of a problem is the quality of optical fiber? Something tells me tells me that even if you imagine that all the capacity already existed, the mere shipping of raw goods up and finished goods down would probably eat any marginal profits over earth-manufactured fiber.

For regular data transmission, I can't imagine it to be worth it - far bigger gains could be made by just making repeaters smaller and cheaper by iterative design improvement.

Perhaps there are scientific uses where loss in fiberoptics makes a project infesible? Or perhaps quantum entangled photons, which one can't amplify?

I suppose the issue is no matter how small and cheap the repeaters get, they will still need power, backups, replacements and all the complexity that implies.

I also have a feeling I've read the power needed for the repeaters generates magnetic fields that attract sharks. Shark bites are a not entirely trivial source of fiber cuts, which require very expensive fleets of cable repair ships to be maintained on standby and simply driving down the number of fiber cuts could be a fairly significant source of cost savings.

Reply on news.ycombinator.com