Hacker News new | past | comments | ask | show | jobs | submit login
Amazon has acquired Facebook's satellite internet team (engadget.com)
208 points by alexrustic 65 days ago | hide | past | favorite | 182 comments

Imagine Walmart at the height of their retail power, but they also control the infrastructure running a huge portion of the internet, several media properties, and are working on creating a new ISP - that's Amazon. We're heading into a world where the likes of Standard Oil will look like a quaint "lifestyle business" compared to the FAANG giants.

Rockefeller's wealth was ~420 billion (inflation adjusted) at its peak. That's significantly more than Bezos and Gates combined. So no, Standard Oil won't look like a lifestyle business compared to FAANG giants. You can make a point without using hyperbole.

> also control the infrastructure running a huge portion of the internet

AWS has less market share than Microsoft. The Cloud is a pretty competitive space.

> are working on creating a new ISP

This is great, look what Google Fiber did to fiber availability in the US. More competition is great in this space.

I find it ironic that you're complaining about FAANGs in the context of ISPs, who control literal monopolies in several regions.

Well actually, Standard Oil peak market cap was about $1 trillion. Rockefeller owned about 25% of this. The rest of Rockefeller's wealth came from the rise in share values post-split. Standard Oil at it's peak actually had lower market cap than any of Apple(2.5T), Google(1.7T), Amazon(1.9T), Microsoft(2.1T) have today, and is about on par with Facebook (1T).

FAANG is huge and market caps already greatly exceed that of the old-school monopolies.

It is fair to say this is quite distorted by the impact of very low interest rates on expectations of future profits though, which is a key component in the share price.

The parent comment was using Rockefeller's peak net worth as compared to FAANG CEO's net worths to show that FAANG is smaller than Standard Oil was. I mainly wanted to point out that if you instead compare company valuations, each of FAANG is bigger.

Also, at it's peak Standard Oil was worth about 6% of the US stock market. Apple is currently worth 5.3%. This comparison should be robust to differences in interest rates. In this light, Apple is slightly smaller than Standard Oil was. But, the market is broader now and there are more lines of business in existence. So a company that dominates an entire line of business would be expected to have a lower proportion of total stock market cap now than in 1910.

At the very least 2021 Apple and 1910 Standard Oil are very comparable in size. There might not be a clear way to tell which is bigger.

>AWS has less market share than Microsoft. The Cloud is a pretty competitive space.

They didn't say a "majority" though, they just said a huge portion.

In the context of the internet and how many devices are out there, I think he would absolutely be right in saying "a huge portion."

> AWS has less market share than Microsoft. The Cloud is a pretty competitive space.

According to gartner[0] AWS has over double Azure’s market share by revenue.

What metric are you using for your stat?

[0] https://www.gartner.com/en/newsroom/press-releases/2021-06-2...

Microsoft likes to throw around cloud figures that include Office 365 to make their market share sound bigger in comparison with AWS.

"Rockefeller's wealth was ~420 billion (inflation adjusted) at its peak."

Dumb question. Why do people use inflation? Like Rockefeller would take the cash and put in a %2.5 account for 100 years? He probably would have put it in an investment at minimum gaining %7-10 account for 100 years - giving him multiple trillions.

Because the point is to give an apples-to-apples comparison of the value represented. It’s not Rockefeller’s fortune and it’s growth that’s interesting but rather how many Big Macs his fortune could buy in 1921 vs how much that many Big Macs would cost in 2021.

But they were comparing Amazon to STD oil, not Bezos (diluted in divorce) to Rockefeller. Rockefeller owned a full third, Bezos 11%, Gates 1% (but similar foundation setup to Rockefeller maybe, and at one point was 49% or something, but of a much smaller MS than today).

Even if you factor in investment returns you're not getting the full picture IMO unless you take into account price increases over time as well.

A common refrain is that Carnegie Hall only cost $1M ($29,581,868.13 when adjusted for inflation) to build, so why do we still credit its founder with their name? Shouldn't we rename it in honor of someone who's contributed more to the Hall?

What this doesn't take into account is what it would cost to _build_ a new Carnegie Hall today. Labor is far more expensive (highest $/hr ever in 2019 if I'm not mistaken [1]) today and so are building materials [2].

So it's true he'd see compounding returns from investing, but to do what they did back then would cost significantly more today. IE, their dollars took them further back then.

Also, worth noting Rockefeller donated 6% of his salary to charity every pay check every single year of his life, not just when he could "afford" it [3]. So if you take into account the _missed_ compound returns of those charitable contributions, you can start to get a sense for just otherworldly their charitable efforts were.

Not trying to say these guys were angels. And yet, as rich as they were, I think too often that overshadows the gargantuan contributions to charity they made.

[1](https://fredblog.stlouisfed.org/2018/02/are-wages-increasing...) [2](https://tradingeconomics.com/united-states/consumer-price-in...) [3](https://www.philanthropyroundtable.org/almanac/people/hall-o...)

> And yet, as rich as they were, I think too often that overshadows the gargantuan contributions to charity they made.

Is it really such a great thing to give away money you don't need. IMO tha should be the absolute minimum baseline expectation for someone who controls so much wealth.

I think it is a great thing. They could just as easily keep it all to themselves and pass it down to generations. It’s fully within their legal rights to do so.

> Why do people use inflation?

Because it is very hard to think about the currency values in terms of purchasing power.

Inflation is the lowest metric in that count, but reflects a somewhat uniform drop in purchasing power (per dollar) across the whole market.

My friend's mom told me her mortgage for a Bay Area house was 75$ a month, the two cars parked in front were worth more than a 2 bed house when she got the cars.

So it was much cheaper to buy a house, but much more expensive to get a car and I can't even have a ballpark figure for what a gigabyte of computer memory would have cost in 1971.

So, inflation adjustment is at best a rough proxy for purchasing power at current costs.

It's because the question isn't "how much money would Rockefeller have now if he'd time traveled while his investments sat there," it's "how much did he have relative to today's wealthy" and the easiest comparison point we have there is to inflation-adjust and compare that to our existing mental feeling for "how rich is Bezos."

I don't know why people are downvoting this. I want to say thank you for taking the time and asking a question, and for actively making an effort to understand something better. I hope the downvotes don't dissuade you (or anyone else) from asking questions in the future.

Because it helps the average person get a sense of scale - instead of saying '100 million, but that's 5 times ore than....' you just do the inflated number and people get a better sense of scale.

Presumably to compare purchasing power. Your investment idea actually distorts the comparison since it involves additional labor and wealth creation.

I prefer fraction of total global wealth as the units. it’s hard to get good estimates though.

Walmart sells a lot more stuff than Amazon. People over estimate the size of Amazon and total retail sales.

Amazon has higher retail sales than Walmart?

Yes, Amazon does more retail sales than Walmart. Their "revenue" is only 15% on all third party sales, which are now more than half of total sales on the platform. They don't report on total sales on the platform on purpose to avoid drawing attention to the fact that they are bigger than Walmart.

I think amazon may have higher total sales on its platform, but like 60% of their sales are actually sold by other marketplace sellers, not Amazon.

Walmart meanwhile does also have a marketplace, but I think still sells the overwhelming majority of their stuff direct in their physical stores.

Remember Walmart has revenue of like 600 billion a year or something crazy.

Shopify for comparison is almost half the size of Amazon Marketplace but on trajectory to match it in the next two years.

But that 15% that amazon takes (not counting the extra cut they take for Fulfilled by Amazon items) is larger than Walmart's average markup. So that isn't really an argument against their size.

That question mark is confusing

If you have to ask what it symbolizes - it doesn't

I'm not confused by it, though I'm a native English speaker and live in an English speaking country and spend most of my internet time on English speaking internet, so maybe that's why

They are creating an ISP, but how successful is it going to be? The clear leader in satellite Internet constellations is SpaceX Starlink, and Kuiper compares poorly to what Starlink offers. SpaceX is going to have a bigger constellation. SpaceX already has a bigger constellation – over 1600 satellites in orbit, compared to zero launched yet for Kuiper. SpaceX has a huge advantage in having their own launch vehicles (Falcon 9, with Starship under development). Amazon has to rely on third party launch providers, and it isn't willing to choose SpaceX (SpaceX COO Gwynne Shotwell has said they are happy to launch competitors' constellations, but obviously Amazon doesn't want to fund its main competitor). For now it has chosen ULA, but ULA costs significantly more than SpaceX, and with Starship the gap is going to become even bigger. Amazon's other major option is Blue Origin – which, although owned by Bezos, is not an Amazon company – but while Blue Origin's launch system New Glenn is competitive on paper with SpaceX's (being in between Falcon 9 and Starship in capability), its maiden launch probably isn't until 2023 at the earliest. I think it is likely that Starship will be launching Starlink satellites before New Glenn is launching Kuiper.

starlink won't scale and everyone knows it.

At most, starlink can replace one tmobile.

amz seem to be buying into everything under the sun and trumpling usage agreements (i.e. "reselling" consumer broadband with people joining via dark-patterns despite it being against their ISP terms of service, via amz sidewalk product)... and that can scale up to O.G. Bell levels of monopoly.

If Starlink won't scale, Kuiper won't scale even moreso. FCC has approved 12,000 satellites for SpaceX, and another 30,000 are going through the approval process. Project Kuiper has only applied for a 3,236 satellite constellation. And SpaceX is way ahead on deployment, and is likely to stay ahead.

Other than Kuiper, Amazon has shown no signs of becoming a serious player in the consumer ISP space. Sidewalk is not a consumer ISP service in its own right, it needs to piggyback on a third-party consumer ISP service. Amazon is nowhere near becoming anything resembling Bell.

The number of satellites is an extremely poor metric for capacity. A single GEO satellite has more USA capacity than about 900 LEOs

Don't you mean to say coverage, not capacity?

Capacity would be aggregate bandwidth, which is determined by the satellite's communications technology (number of transponders, spectrum allocated, modulation used, etc) rather than directly by which orbit it is in.

Comparing constellation sizes does make sense in determining total capacity (aggregate bandwidth of the constellation) assuming that Starlink and Kuiper use similar capacity satellites. I think that is a reasonable assumption, but we have limited info on what Kuiper are planning. Current Starlink satellites have on average 20 GBps capacity each. I can't find any public info on the planned capacity of Kuiper's satellites. Even if Kuiper has higher bandwidth satellites, they'd need a fourfold advantage in bandwidth to overcome Starlink's fourfold advantage in satellite count. And SpaceX is likely to increase the capacity of Starlink in future iterations, so even if Kuiper did have such a bandwidth advantage, it might not last.

Both. I specifically said the USA because that's where they're going to make the most money, and at a given time only about 10% of their satellites are over it.

The satellite capability is one thing, but spectrum is another. SpaceX doesn't have priority on a lot of the spectrum they have, so it might be completely unusable by the time they go live.

Is Kuiper's spectrum situation any better than Starlink's? From what I've read, Starlink and OneWeb have priority over Kuiper for the same spectrum (since they were approved first).

I feel like we are arguing two different things. I was saying Kuiper can't beat Starlink. You are arguing Starlink is not going to succeed. I'm more optimistic about its prospects than you are, but even if you are right, it wouldn't change my original point – if both fail, then Kuiper still doesn't beat Starlink.

How can you say this? Starlink is in beta and nowhere near the full capacity they need to sustain the business.

My main point was they are way ahead of Kuiper, which is obviously true. SpaceX has launched over 1600 satellites, Kuiper has launched zero so far. SpaceX has launched over 700 satellites so far this year. SpaceX has a public beta in multiple countries. Kuiper can't have even a private beta since they haven't launched anything yet.

I think you're overestimating what it takes to catch up here. There are no contracts. If one is better than the other, people will cancel their service and switch at the drop of a hat.

For Kuiper to catch up they need to solve their launch systems gap with Starlink. I think you're underestimating how hard it is going to be for them to close that gap.

Blue origin has been at it for a while. While kuiper is new, blue origin is not.

Kuiper and Blue Origin are independent companies. Amazon owns Kuiper. Amazon does not own Blue Origin, Jeff Bezos owns that independently of Amazon. Despite having some ownership in common, they are legally required to operate at arms length.

Blue Origin is two years older than SpaceX, and yet has achieved far less than SpaceX has in the same timeframe despite a two year head start. Over that timeframe, SpaceX has successfully delivered 124 Falcon 9 payloads to orbit. Blue Origin hasn't sent anything to orbit yet, New Glenn is targeted to "late 2022" but most observers think it won't launch until 2023 (or even later).

Blue Origin is the one that needs to prove itself here. SpaceX has a proven capacity to deliver payload to orbit. With Starship, that capacity is going to greatly increase. But even if Starship runs into trouble and gets delayed, they still have a proven capacity with Falcon 9 to rely on while any issues with Starship gets delayed. Blue Origin is working through the issues on New Glenn without any fallback.

They operate independently, but both are funded by bezos. They don't need to be owned by the same company. They need to both be good at their respective goals and cheap launches will follow.

But to be clear, none of this matters. The business case to sell a $1000 terminal for under $300 isn't closing. They're subsidized right now, and public perception will be very different once the real sticker shock comes in.

> They need to both be good at their respective goals and cheap launches will follow.

Is Blue Origin good at its goals? On New Glenn, which is what matters here (not New Shepard), the jury is still out. By contrast, for Falcon 9, the jury returned a verdict long ago.

> The business case to sell a $1000 terminal for under $300 isn't closing. They're subsidized right now, and public perception will be very different once the real sticker shock comes in.

They say they've been cutting the manufacturing cost from $3000 to $1500 to $1000; you are assuming they aren't going to succeed in cutting it further. Even if they can't get it all the way down to $500 (the current sticker price), if they can get it down to $600 or $700, a $100-$200 price rise for new customers caused by withdrawing the subsidy is unlikely to be a huge issue. (The $300 figure was just an aspirational goal Musk set, nobody is paying that little right now, who knows if anyone ever will.)

None other then Elon musk says they need $20-30B in investments to stay afloat recently. They are nowhere near where it needs to be for the terminal cost. The launches can be free and the business will still sink without more government subsidies.

Doesn't Biden plan to go big on anti trust laws?

Some stuff I read sounded rather bleak for the likes of FAANG.

Politicians' plans and politicians' accomplishments tend to be vastly different things.

We're starting to get some political alignment on anti-trust against Amazon at least. Though one party seems motivated by worker/consumer right, the other by retaliation for Bezos' ownership of WaPo.

Considering Bezos is hardly involved with Amazon anymore it’ll be hard to tie WaPo into its antitrust case.

Do they need to tie WaPo into whatever antitrust may hit Amazon? As long as it's hurting Jeff Bezos, I'm sure some will be happy about it.

I fully expect Amazon's lobbying arms to be able to exploit that sort of ideological split.

It would be very easy for Bezos to divest WaPo or tilt its coverage more conservative to buy off an anti-trust issue.


So; how did antitrust work out for Bell? Not too terribly bad, in the long run. (Though there were huge benefits for us consumers. . . )

The various Standard Oil split results had some nice successes, too.

Honestly I'm not sure of the main corporate opposition to anti-trust other than ego and/or laziness. It'll force the resulting business to get even better. If you got your position by being the best, you can still be the best in the new world.

This is basic infrastructure/utility and should be run by the government (with companies as contractors).

I was recently looking at getting starlink. During my research I got a good chuckle at the realization that it’s more cost effective for SpaceX to launch thousands of satellites into space for starlink then it is for someone to dig a goddamn hole to provide adequate service.

Even slightly rural places it’s not cost effective for traditional isps to dig a cable so what’s the markets solution? Launch thousands of satellites! I don’t know why you’re being downvoted. The market has failed, requiring external forces to intervene(the government). Whether it be Title 2ing them or allowing municipal competition something needs to happen

Yes. It should absolutely become less agile, over-regulated, more expensive, and less convenient.

Don't forget that government institutions invented the internet, then companies messed it up.


'messed it up'- care to elaborate?

Can you imagine working at FB, which despite its considerable ethical failings, is a decent employer from what I've seen and suddenly being told you're going to Amazon and need to start memorizing The Principles?

Landing a job at either Facebook or Amazon is an unachievable dream for most people so I have a hard time to feel sorry for anyone here. In their position they can get any job they want.

What makes it unachievable? The process is very straightforward. Every friend of mine that wanted to work at FAANG has been able to and you really just need to spend a couple weeks/months practicing.

The process is straightforward to become an Olympian too, you still don't see people doing it every day. You can run every damn day if you want, but I'll bet dollars it would still be difficult to run a sub 3:00 marathon for most people.

Working at FAANG is easily regarded as one of the higher end jobs in the SWE world (from comp and prestige) and does generally require you to be in the top % of people who know how to code. That much is fairly indisputable so unachievable is a good way to put it for most folks (even if they are into programming already)

I think this is half the problem when people interview at FAANGs -- they think it's like preparing for olympics and you truly have to be in the top 1%.

It's really not. There's thousands and thousands of engineers, from mediocre to truly great. With the right preparation, a lot of it just comes down to getting lucky with who interviews, how they're feeling that day, etc -- luck.

You might find this interesting: https://www.youtube.com/watch?v=r8RxkpUvxK0

I mean, you can say the SAME thing about getting into an Ivy League school. Your application and skillset is likely similar to THOUSANDs and yet still, you have to prepare like it is the Olympics AND have luck. No one lucks their way into FAANG, you still have to answer the questions well and contrary to much of the HN crowd, these are not easy questions for most folks EVEN after studying...

You can't get in without both luck and preparation. That part I won't deny, but most exclusive things in life involve that bit of luck since everyone starts to hit the "threshold" of qualifications pretty fast

>I mean, you can say the SAME thing about getting into an Ivy League school.

No, you cannot say the same thing about Ivy League admissions, at least not at the current point in time. Ivy League has a massive supply of qualified candidates to fill the demand (aka the fixed number of seats they have) many many times over. Tech companies don't have that luxury. They have a lot of candidate supply, but the "qualified" candidate supply is much more difficult to come by.

Ivy League admission officers admitted multiple times in public interviews and public statements that there are a lot of people that would qualify and succeed at those schools just fine, but the schools have very limited numbers of seats, so they have to have the bar be set higher and higher in order to accommodate truly the "best" (based on whatever metrics they use to determine the best).

With FAANG? They are desperate to hire good engineers. As someone who interviewed candidates at one of them, for every single interview I went with the mindset that I want to hire that person, as long as they demonstrate they are competent enough to do the job. There is no such thing as "limited" number of spots (as it is with Ivy League admissions). If every single candidate we interview is qualified, we will hire every single one of them, we won't be artificially raising the bar just to accommodate a specific X number of people because we cannot fit more. We can. Of course, my specific team cannot hire 50 people, but if all 50 candidates are competent, we will try to set them up with one of our sister teams or any other team in the company that desperately needs engineers, and there are tons of those teams.

The real problem is that a lot of those people we interview tend to not even be able to do fizzbuzz, and I am not exaggerating. I used to think that people said that statement as a joke back then, but after interviewing enough candidates, I realized that they weren't kidding.

TL;DR: they aren't similar, because Ivy League has a fixed number of seats available, and even if every candidate is qualified, they can only accommodate a predetermined number of them, so they have to artificially raise the cutoff metrics. For every candidate my team interviews, we can easily hire every single one of them if they are competent, without artificially raising the bar just to have a fixed number of hires.

> they think it's like preparing for olympics and you truly have to be in the top 1%.

You have to be capable and interested in passing a FAANG interview. Presumably those that failed the FAANG interview process are either incapable or uninterested. My money says it's mostly the former.

You have to pass-as what FAANG deems as the top 1%. Most people don't.

"In my bubble it has been achievable so it must be achievable for everyone"

There is a reason that most people do not work at FAANG.

One of the more “hn” comments I’ve seen lately. Assuming every person in the world is an ivy league graduate, with a high IQ, and A+ coding talent is a little bit out of touch with reality.

I just don't think those are the requirements to get a job at FAANG. I went to a mid sized public school in the midwest that isn't even in the top 100 US schools, my IQ is definitely average, and I seriously didn't understand sorting algorithms when I first saw them in my intro to programming class. With time and practice anything is possible.

And I want to echo a point filoleg made. You don't need to be the top 1%. It's not like Ivy League school acceptance where they are artificially throwing people out. FAANG is dying to hire. My team has been trying to hire multiple people for months and we have gotten exactly 0. Every hiring manager goes into interviews wanting to say yes and the requirements are extremely clear and manageable for anyone willing to put in the time.

My understanding is that for any given age the average IQ is 100. Are you seriously suggesting that you would score 100 on an IQ test? I highly doubt that considering that your manager is so motivated to hire but simply can’t.

Put another way, the modal IQ on HN is likely higher than 100. So I have a hard time taking someone seriously when they claim they have an avg IQ and think getting into FAANG is easy. That just means they have no clue how much smarter they are than most.

visa issue.

this is spoken so casually and without any consideration for the underlying privileges one requires in terms of education and lifestyle to enable such study that I must now laugh

I am an architect at a "prestigious" midwest consulting firm without a formal IT background. It would take me 6 months of prep to pass a FAANG interview, and that's from a starting point of knowing virtually everything I don't know.

AND I am only talking about the algorithm and behavioral parts of the interview and working on the assumption that I will pass any general questions on system design, ML, etc.

Other than the CS degree, you are also make a 130+ IQ assumption here.

Even if it took 6 months, is that such a big deal for what is probably a large pay raise and potentially the opportunity to work on cooler and more impactful things with better coworkers? And for the record the idea that 130+ IQ is necessary is ridiculous. I can assure you I'm well below that.

Point by point:

1. IQ is going to be in the 120-130 range for MIT and other FAANG target schools (https://www.iqcomparisonsite.com/occupations.aspx and factor in the GPA necessary and those correlations). If you are NOT from those schools, you are less competitive and have more to catch up on (like me). Whatever you say, either you had a lot of study free time or your IQ is higher than you think, as IQ translates directly to learning speed, especially on computer-related tasks.

2. Big deal. For a young student out of school? No, it's the rational choice, though a lot of people out of school/college NEED to work to keep the bills paid. For older people - dedicating a good 3-4 hours a day is going to lead them to underperform at work (at least it would for me), so it's kind of a big risk.

The pay increase is also "debatable", arguments like COL, vesting, decrease in quality of life due to commute, relatively-high (by non-FAANG) standards salary, taxes, etc come up. And that's without mentioning potential kids.

I've worked at Amazon and other FANG companies and I honestly miss amazon's leadership principles. They're really just assigning a phrase to a Good Idea (TM).

The leadership principles are an important part of Amazon's solution to the problem of getting hundreds of thousands of employees pulling in the same general direction. They are easy to get people to sign on to because they are, for the most part, obviously good ideas succinctly stated. People at Amazon actually use them to make decisions daily.

Other companies don't propagate their corporate culture so explicitly, so it seems a little weird and cultish. But if you're a big company doing things the same way other big companies do them, you're going to get the same results that other big companies get. Amazon wants to get better results than other big companies.

When I started at Amazon, I felt like 14 was simply too many. I tried to do a Carlin-style winnowing, but darned if I could get the number below twelve without starting to lose things that the company obviously thinks are important. (They recently added two more, by the way.)

As a bonus, after a stint at Amazon, your familiarity with their leadership principles can make you very desirable to other companies. I know a number of people who were individual contributors at Amazon and were snagged by Microsoft for leadership positions, either team leads or product managers.

I would have trouble with the blatant hypocrisy. To take an example from just yesterday, Amazon has internal documents that instruct managers to hide from employees when they're on a PIP. How can you reconcile that with "Strive to be Earth's Best Employer"?



There are probably better examples you could've used. They've literally added "Strive to be Earth's Best Employer" to the list of principles like yesterday. At the very least they seem to have recognized the problem, shouldn't we give them an opportunity to actually try to fix it?

That LP didn't exist until after the article was published. There is some hope that it could be leveraged to remove that kind of guidance from mid-level leaders.


It seems like this is a sorta "pre-PIP"?

> If Amazon employees don’t improve while unknowingly in the Focus program, they are then placed into the “Pivot” program, according to previous reporting from Business Insider. Employees told Business Insider that if they were placed in Pivot, they were either offered a severance package or given a chance to be put on a performance improvement plan.

So "Focus" is more a manager-focused thing and then if the manager isn't able to turn things around, out comes the real PIP, is my impression.

Honestly, I've had to put someone on a PIP before, and more support and training on what to do before it got that bad would've been very appreciated.

I think the manager needs to be transparent about "you need to do better" but not about internal management practices necessarily at that point.

“Strive to be earth’s best employer” isn’t a leadership principle. - oh shit, it is now! My bad!

And “instruct managers to hide from employees when they're on a PIP”

This is not as clear cut. The document cited by Seattle times clearly states to not go into details of the Focus application, and instead go into how they can improve.

A quote: “Do not discuss Focus with employees. Instead, tell the employee that their performance is not meeting expectations, the specific areas where they need to improve, and offer feedback and support to help them improve.”

“Do not discuss Focus with employees." I don't know how much more direct this could possibly be. The only time the policy allows managers to tell employees they're on a PIP is if the employee asks directly. That's literally hiding the truth as much as they possibly can without actually lying.

If your manager is telling you that your performance is lacking and that you need to improve in specific areas, how is that really any different from being told you are on a PIP, and the rest of that? That is, what would you, as an employee, do differently if you were told specifically that you had been placed on a formal PIP, rather than just being told how you needed to improve?

> what would you, as an employee, do differently

Start looking for another job before I get fired. Which I assume is about 10x easier than doing so after the fact.

The Principles are merely words used as tools - often for selfish or even evil purposes - not some sort of adamantine moral backbone. The idea that they exist somehow gives a higher sanction to the acts performed ostensibly in their name.

amazon subsidiaries still operate with their previous culture to some degree. Twitch is much different than the rest of amazon. However, since they will be merging into an existing org not sure if their culture will remain

I'm sure Amazon Corp HQ is taking a boiled frog approach to absorbing things like this.

I heard they are still very much different from amazon after the merge which is good to hear. However, it was told by a hiring manager :) So it's not from my personal experience.

It's going to be at least a year before the incoming FB team will be taking their vows.

still very different but amazon has its influences in many parts of the company now.

This is America and if you are upper or middle casted, then Amazon will give you better accommodations and compensation. It’s comfy, don’t worry about those people.

Is this a reference to Amazon being disproportionately Indian?

America has a caste system too but we call it "class" typically to make it seem more meritocratic. In conversation we typically focus on the exceptional cases of economic mobility rather than the rule. I did use caste on purpose to draw the parallel to India, a place where Britain famously used the caste system to rule.

That's a little bit intellectually dishonest. Class mobility does exist, albeit more rarely than it should, and unlike places with enforced social systems, you can't tell someone's social class by their name or clothes.

Class is also a generalization: I'm a gainfully employed SWE, and I (by choice) live in a neighborhood where I hear gunshots nightly. I don't own any new clothing, my nicest pair of shoes are my work boots, and my friends are almost entirely blue-collar people. I could easily afford to live like an upper-middle-class yuppie, but I border on lower class at first glance.

Is class entirely economic, or is there a social aspect?

perhaps they can dry their tears with hundred-dollar bills /s

> Amazon aims to have a 3,236-satellite constellation in orbit by 2029

Assuming Starlink is similar, are there any risks of "imprisonment" on Earth having ~8k low-orbit satellites flying around? In that they gravely affect efforts to fly Humans to the Moon/Mars?

On most days there are up to 20,000 airplanes (which are 100-1000x larger than satellites) in the sky. And the satellites, being at 100x higher altitude, are 100x more spread out than the airplanes.

So just from a "unusable space" point of view, it's on the order of 10000x less of a problem than airplanes. The caveats here are the satellites are moving much, much faster than airplanes, and they stay in the sky much, much longer.

But it's not really a huge problem unless stuff goes wrong and you get Kessler Syndrome. This is more of a risk with the higher constellations like Amazon's and OneWeb's than it is with SpaceX Starlink (which is in a low enough orbit to de-orbit all debris within a few years, rather than centuries).

> And the satellites, being at 100x higher altitude, are 100x more spread out than the airplanes.

This doesn't feel right. Shouldn't the right frame of reference be the distance to the center of the earth, not the sea level? Then it is not a 100x difference, but more like a 10% one.

Here's the actual correct math:


Area of sphere: 4πr^2

Airplane cruising height: ~10km

Satellite orbit height: ~550km

Radius of earth: 6371km

thus, relative increase: (6371+550)/(6371+10) = 1.084 = 8.4% increase

Squared (because of first formula) that corresponds with a 1.084^2 = 1.175 = 17.5% increase in area.


Still, a few caveats:

1. Earth is huge. 510.1 million km² is a lot of space (+20% at 550km altitude). We could have a million sats with each having more than 1km^2 to themselves.

2. Satellites orbit at different heights. Amazon's and SpaceX's satellites will not be on the same orbit.

3. Starlink satellites are at a sufficiently low altitude that Kessler Syndrom is not a problem; even if they all simultaneously turned into millions of pieces of dead debris at the same time, the atmospheric drag would make them lower their orbits and burn up in just a few months.

Agreed rough math there doesn't seem right. Also, I wonder if there's an added benefit from the inverse square law here: https://en.wikipedia.org/wiki/Inverse-square_law

If we were also launching from the center of the earth I would agree, but as we are launching from sea level, I think the general idea still holds.

Thought experiment:

You are standing in a field.

There is a 10x10 meter plate hovering 10 meters above you. There is a 1 square meter target on it. You fire a gun upwards at a random location on the plate. There is a 1 in 100 probability you hit the target.

Now imagine there is a 20x20 meter plate hovering 20 meters above you. It is perfectly occluded by the original 10x10 meter plate. It also has a 1 square meter target on it. When you fire at a random location on the original plate, the bullet passes through it and continues on to the higher plate. What is the probability you hit the target on the higher plate? I believe it is 1 in 400.

From this thought experiment it seems that altitude from launch point is what counts.

If your original claim were true, then you'd also be saying that there is twice as much area at 2m above sea level as there is 1m above sea level, and infinitely more than at sea level. If you truly believe this, then I will be willing to make you a trade where I give you one tenth floor condo in exchange for 10 first floor condos.

Now expand your thought experiment to 4 people, standing in a 10 meter square from each other. They each have their own 10x10 and 20x20 plates, and their own targets.

Clearly the 10x10 plates are lining up (modulo curvature). But the 20x20 plates are not, they're overlapping. So when I shoot through a random location of my own 10x10 plate, there's a chance that I'll hit a target from somebody else's 20x20 plate. Sum up those additional chances, and they'll cancel out the 4x difference you found.

This feels like it would make a nice puzzle, your phrasing makes for a great misdirect / sleight of hand.

I totally agree with you and the other commenters that the total area of the spherical shell is only increasing by a few %, but I still think for the purposes of "escaping earth" as the top level comment was talking about, all that matters is how spread out things are directly above you.

When talking about the area of the spherical shells, it conflates what is above me as equally relevant to things that are on the other side of the planet from me.

That is, satellite X and Y may be over me at 500km and 1000km distance, respectively. Later, they may be directly through the earth from me at distance 13200km and 13700km distance.

In the first case, if I shine a laser straight up, my probability of hitting X is 4 times higher than my probability of hitting Y.

In the second case, (if I could somehow shine a laser straight through the earth), the probabilities are nearly equal.

But my intuition is that for the purpose of escaping earth, this second case does not matter, because we are just dealing with what is above us, not the entire spherical shell.

Except you can't generalize a square plate to a spherical shell...

Can you explain why?

If I am launching a rocket, and there is a 1x1 meter satellite orbiting 1000km above me. What is the probability that my rocket hits that satellite compared to an identical satellite at 2000km above me? The area of the angular sector of the sky that the 2000km altitude satellite is 1/4 of the 1000km altitude satellite.

The orbital launch will follow a curved path through succesive spherical shells. The percentage of the spherical shell occupied by the 1000km satellite is not 4x larger than the percentage of the spherical shell occupied by the 2000km satellite.

This is the wrong mental model entirely unless your rocket launch is directly vertical and instantaneous.

(10x10):(20x20) sounds like 1:4 to me, instead of 1:400, no?

It's not 1:400 -- it's 100:400 which is equal to the 1:4 you mention.

That is, you 2x the height of the target which results in the probability of hitting being 4x less.

They're at 100x higher altitude relative to the sea level but nearly the same altitude relative to the Earth's center which is what matters here.

But we know where all the planes are, and actively monitor and move them around to avoid collisions. If there are a sufficient large number of satellites would there not be some risk of accidental collision with spacecraft taking off?

We know where all the satellites are and they are actively tracked by both governments and private citizens.

Every satellite is in a very deterministic orbit which requires energy to change (enormous amounts of energy for a significant change) so they don't change their orbit significantly nor often.

Ref: https://sos.noaa.gov/catalog/datasets/space-trash-and-satell...

> But we know where all the planes are

We really don't. We mostly know for commercial flights, but not so much for general aviation. What aviation does is have zones around airports with restricted airspace that is well-controlled.

This is a fun site for tracking stuff in space:


NASA & the ESA actively monitors all satellites and actually maintains a database of every little piece of space debris they find (payload fairings, failed satellites, etc). They determine the orbit and track it until it falls back to earth.

> But we know where all the planes are, and actively monitor and move them around to avoid collisions.

This is also true for satellites.

> But we know where all the planes are, and actively monitor and move them around to avoid collisions.

Do you think they don't know where their satellites are?

A similar monitoring applies to satellites as well. We're not launching spacecrafts and just hope that everything goes well.

But we can ground airplanes in a matter of hours. It doesn't really feel comparable.

* Altitude is essentially the same 6300km vs 6700km, so satellites aren't signficantly more spread out.

* Those 20k planes don't all fly at the same time.

* Retired / Broken planes don't fly, satellites may still be on orbit for decades.

* There are no debris flying in the sky at 11000 km/h

* Planes can adjust their path at will instantly for avoidance.

* Planes can be grounded instantly if we need to.

Orbital space is a limited resource that gets depleted very fast and recovers very slowly. We are talking about launching in the next ten years 5x the total number of satellites that were ever launched so far. I am sure humans in 50 years would still be able to launch a thing or two in space as well.

This is an actual problem, and unlike for planes, once the problem is apparent, you can't just take some of them out of the sky while you figure out a solution.

> Those 20k planes don't all fly at the same time.

They absolutely do. Check out https://www.flightradar24.com, there are currently 15,673 planes in the air worldwide at the moment I am typing this comment.

Fair point, but does that really change the conclusion?

Broken planes fall down, while satellites create space debris.

Dead sats in LEO fall.

I haven't looked at the Amazon plans, but the risks of something like Kessler syndrome "trapping" us on earth from Starlink are low.

The main reason is that the Starlink satellites fly in a very low altitude, such that even if they lose all control, they will deorbit in a few years. Which means, if something went horrifically wrong, the Starlink system debris would clear itself within a short period of time.

It looks like the Amazon Project Kuiper satellites will be slightly higher up, but still have a natural orbital decay time of between 5-7 years. https://spacenews.com/amazon-lays-out-constellation-service-...

So, the long term risks from these kinds of low-orbit mega constellations is fairly low. If anything goes catastrophically wrong we wait a decade and it's gone.

I think personally, the longer term risks we should be wary of are medium altitude and geostationary orbits that won't naturally clear themselves for decades or centuries if something goes wrong.

Probably not. Even 500,000 sattelites would still leave more than enough space between them for whatever you want. The risk would be if many of them blew up or something, creating billions of tiny pieces of debris. In that scenario, we might just wait 5 years and then most of the debris will have burned up in the atmosphere.

SpaceX satellites are around 550km attitude. If someone puts satellites higher than that, collision debris will last longer, the satellites' fuel will last longer so the satellites won't have to be replaced as often, but network latencies will be higher. Seems like 500-600km is the optimal zone for the primary constellations of internet satellites.

Video showing decay of debris vs. its altitude. The "X" shape is because each debris is plotted twice, once at it's perigee and again at its apogee (describing the ellipse of the orbit as they generally are not perfectly circular)


Higher altitude satellites would be a bigger concern but these aren't a big deal.

What you're asking about is basically https://en.wikipedia.org/wiki/Kessler_syndrome

In general the question of collision risk and debris is something evaluated for every launch/constellation. Starlink, for example, mostly avoids it being an issue by flying so low that debris quickly falls to earth and burns up in the atmosphere (they also design their satellites to fully burn up in the atmosphere). On the flip sides Starlink is planning on an order of magnitude more satellites than this.

Even the worst case though doesn't really impact humans ability to fly to the moon/mars. You can make low earth orbit relatively unusable because there is a high collision risk if you hang out there for a year, but you should basically always be able to fly through low earth orbit to a higher orbit with negligible collision risk.

Suppose there were only 8000 cars in the world, evenly spread all over it (I'm being generous here, since 2/3 of Earth is covered by water). How likely would you be to ever see one?

The starlink satellites are traveling a little bit faster than your average car though. They complete an orbit about every 100 minutes. So your 8000 satalites are covering much more ground than the same 8000 cars.

At any given moment the distribution is similar, and we track and regulate the orbits, leaving swathes of space open. Everything about the satellites is engineered, including their orbits. Other countries have less stringent standards, resulting in more or less random de-orbits that can drop space junk anywhere in their path, but the US has a lot of forethought put into orbital standards.

Space junk is an issue, but it's not anywhere near crisis level yet.

The planned Starlink constellation is actually quite a bit bigger, but even if there's 20k satellites whizzing around it's not a huge problem.

I think it's sometimes hard to reason about the vastness of space, but imagine if the planet Earth had exactly 20,000 cars on it's surface. Even if you crossed the street without looking, your odds of getting hit by a car would be incredibly low. And ofc in reality low earth orbit is bigger than the surface of the earth AND we know where every obstacle is located. If humanity ever gets to the point where we decide it's too crowded, most of these constellations are low enough that they'd naturally deorbit in less than a decade.

Space is big. You just won't believe how vastly, hugely, mind-bogglingly big it is. I mean, you may think it's a long way down the road to the drug store, but that's just peanuts to space. - Douglas Adams

You quoted that Amazon planned to have 3,236 satellites.

Yet, SpaceX already launched 1730 satellites, 1630 of which are active, with a planned constellation of 12,000 satellites.[1]

Amazon's Kuiper Systems hadn't even launched yet, and they're going with ULA for their first launch, which AFAIK, is much more expensive than SpaceX, with only 9 satellites as opposed to 60 satellites at a time.[2]

[1] See https://en.wikipedia.org/wiki/Starlink

[2] See https://en.wikipedia.org/wiki/Kuiper_Systems

And if/when SpaceX finishes Starship plus Superheavy, SpaceX will be launching up to 400 satellites per launch.

Also, they've submitted authorization for up to 42,000 total starlink satellites: https://spacenews.com/spacex-submits-paperwork-for-30000-mor...

And they applied for permission to launch additional 30k satellites, bringing the total to 42k. Amazon Kuiper is a toy by comparison..

No. This wouldn't happen as space is much bigger and there are far more airplanes today in the sky than satellites with a higher more distant apart orbit.

These low earth orbit constellations will naturally experience orbital decay and at the end of their useful life will simply burn up on re-entry. They're explicitly designed to prevent Kessler Syndrome:

See also: https://en.wikipedia.org/wiki/Kessler_syndrome

There are two probable problems:

a) debris damaging other satellites or space stations. There currently is no proper liability currently and different monitoring systems are still in development

b) astronomy from earth will see problems.

As long as they are on their orbits there is enough space (haha) and if they don't cause conflict with radio frequencies they also don't cause issues

There are already approximately 250,000 objects in low earth orbit 2cm and up. The risk from these constellations isn't from the satellites themselves directly, but from Kessler syndrome if they start getting smashed by untracked debris.

Atmospheric drag brings down anything below 800km:


Starlink operates at 550km altitude:


At 800km it can take a long time (decades)

Starlink is at low enough orbits that even if the entire constellation spontaneously exploded all the debris would deorbit within ~5 years

pardon the dumb question - how do rockets from spacex, etc avoid crashing into a satellite on its way up? Is there a way to keep track of all the satellites and debris in space and time the launch so there is a clear path?

Yes, there is a database and radar tracking stations, and they'll avoid launching if there's an increased chance of a collision. The density of satellites and debris in orbit is so low in absolute terms, though, that the probability of any given rocket colliding is negligible even if launching totally blindly.

Most launches are timed for minimizing fuel to achieve the desired orbit, and there's only a few seconds of wiggle room for a launch window. So, a particular launch window may be preferred over another depending on the relative probability of a collision. Nobody is explicitly timing their launches or ascent profile with regard to other satellites other than for space stations and other explicit destinations. Most launches the rocket just gets the satellites up into roughly the right orbit, and then the satellites use their own propulsion systems to maneuver into precise orbits over several months.

Short answer: yes.

They do this thing called "COLA", Collision On Launch Assessment, an analysis of launch trajectory to ensure they it won't hit known objects.

we are in the new age of consolidation.

Just like the combustion engine development created the transportation and assembly revolution some ~60 years after, profits from the newly created peripheral markets led to massive profits in specific tech sectors (oil, steel)... Flush with cash, they started horizontal, vertical acquisitions, leading to the massive corps of the time (1950s).

We now have the internet. We are roughly almost 60 years into the internet creation revolution cycle.

We have seen how this movie plays out.

It's amazing how big these tech companies are getting:


Fascinating blog post. Satellite internet seems just as valuable to FB as Amzn I wonder why they’d spin it off when they have so many other projects that are less useful

Are these global internet initiatives targeting faster internet (10+ Gbps) or just more accessible internet for poor regions?

Faster than what, fiber? DOCSIS 3.0? No.

10x faster than anything that has ever been available in many regions of the globe that currently have some form of limited access? Yes.

And then of course this Internet being available absolutely everywhere on Earth.

It varies from constellation to constellation.

O3B's MPower constellation promises 10 Gbps terminals, but those terminals are very likely to go to big enterprise customers, as they will cost several $10k's.

Telesat's Lightspeed constellation is aimed at rural areas of Canada and 5G backhaul. To do the latter, they'll have to deliver at least 1 Gbps. Telesat's been in the game for a long time, so I wouldn't doubt that they'll deliver.

There's two Chinese constellations going up. They'll probably deliver anything and everything they can, but it remains to be seen what the satellite and terminal capacities will be. The west still has an edge here.

Starlink started with residential service but they'll try to expand into everything they can. One of applications I expect to see is to use Starlink to get data out of Teslas and back to the ML team so they can improve their self-driving code.

Kuiper? I mean, who knows. It's Amazon, so, they'll probably also try to expand into everything they can. I wouldn't be surprised if their delivery trucks will use it as backhaul.

You might notice that none of the above, so far, are actually aimed at connecting the unconnected. That's because the terminals, so far, are too expensive and too power-hungry. The only two initiative I'm aware of that are actually trying to connect the unconnected are:

OneWeb, which has truly global coverage (including the poles) and has a quite smart design, including working crosslinks and a relatively affordable terminal. Also, in the arctic, they can provide connectivity to militaries, so there's some good cashflow there.

Curvanet, an initiative by Tom Choi. They promise a sub $100 terminal that can run on 5 W (!) and does not require an ESA but can still deliver up to 50 Mbps. If they can actually pull this off I will buy a unit and stick on my roof just for fun. Also, if they succeed, they would be absolutely best-positioned in the market, as no one can (so far) match their cost or power consumption.

I believe it's both, but mostly accessibility. Access to internet is seen as limitation on growth for the large companies. Bandwith is certainly factor though when considering AMZN's media empire.

It's very unlikely to be faster internet (compared to what is generally available in urban/suburban America).

Ultimately, we're using radio similar to any 4G or 5G connection. There are some differences, but they probably work against satellite internet more than they work for it. Terrestrial wireless networks like Verizon or T-Mobile can easily split cells to get more capacity (install another tower and split traffic between them). It's harder to do that with satellites. Plus the satellites will be over ocean so much of the time when their capacity can't be used (and other satellites in the network will take that traffic). You do get better line-of-sight which is helpful, but the amount of capacity is somewhat limited. Elon Musk has even said that they'll "most likely" be able to serve 500,000 customers. "More of a challenge when we get into the several million user range."

These are unlikely to be services that offer a faster internet completion for people who are already well-served. Of course, there are a lot of people who aren't well-served.

Over the coming years, 5G home internet is going to become more common. T-Mobile is looking to sign up 7-8M customers which would make them the 4th largest home broadband provider. Verizon has announced that they want to cover 50M households for home internet by the end of 2024 (T-Mobile already offers home internet coverage to 30M households; also, remember that while there are 330M people in the US, there's only around 130M households). While 5G won't reach everywhere that satellite will, it will fill in some of the broadband gaps that we currently have. 5G will also offer an alternative to wired home internet in many areas.


A lot of the interest in satellite broadband is driven by government money. The US government is offering $9.2B to expand access to 5.2M home/businesses in this one instance and there's a lot more money where that came from. The Universal Service Fund in the US shells out billions every year to companies providing rural connectivity. It might not even be that there's much interest in these programs beyond government subsidies. Of course, once the satellites are traveling over other areas, might as well make the service available and get some extra money.

Musk has said that Starlink might require a $30B investment to be viable, but might become cash-flow positive after $5-10B. $5-10B is probably well within the realm of government subsidy in the US. I mean, the government will definitely be spending much more than that subsidizing rural broadband over the coming years, but whether Starlink/SpaceX will get that money remains to be seen.

I wouldn't say this is just about poor regions. There are plenty of far-flung places that I wouldn't classify as "poor" that don't have great internet.

I think an underrated part of this is applications for merchant ships and edge compute in LEO (though I don’t think Amazon has discussed this explicitly). The enterprise use cases could be interesting.

I think they're targeting accessibility and reduced latency on long journeys since light travels faster in air than wire. I don't believe they're targeting faster bandwidth though.

Comparing the launch costs:

> "The average launch cost for Atlas was about $225 million per lift," Bruno told FLORIDA TODAY this week. "Where Atlas V is at today is almost half of that just by virtue of all the changes we have done in this business."

from [1], so around 100Mio per launch.

The same article cites 50-60Mio per Falcon 9 launch.

[2] says the payload mass to LEO is between 9.5 and 18.5 metric tonnes for the Atlas V (depending on the configuration, mostly how many boosters you attach), and 16.8T for a reusable Falcon 9.

So I'd say Amazon pays at least twice as much as Starlink, measured by mass in LEO.

[1]: https://eu.floridatoday.com/story/tech/science/space/2020/11... [2]: https://en.wikipedia.org/wiki/Comparison_of_orbital_launch_s...

Was it ever a serious effort; or was it an excuse to get the FCC to hinder similar projects?

Disclosure: I worked at FB, but not on connectivity projects, although I was adjacent to internet.org.

As far as I could tell, these were serious efforts, but it's important to understand the organizational goals. Facebook benefits from increased connectivity regardless of who operates the network (as long as it's not actively hostile to FB, anyway). So there was a focus on research and publication rather than build-out. If FB can find a viable improvement in networks, and convince networks to use it, that increases connectivity which is good for FB. Additionaly, if network providers use FB developed technology in their networks, it might improve relations with those networks, which are sometimes strained because of competition in the messaging space.

In my mind, this is the same as Google Fiber. Google Fiber was a terrible business for Google, but as a result of announcing their plans to build out in specific cities, the incumbent networks built out high speed fiber in most (or all) of those cities and maybe a few other places, which increases penetration of high speed connectivity, which is good for Google as a whole, so it's still a win.

"A rising tide lifts all boats" sounds like a pretty good reason to me. Thanks very much for illuminating more of the details.

From Facebook's POV? Very serious.

Edit: For added context, I worked on this along with several other FB Connectivity projects.

That's interesting, can you share anything about why we never really seemed to see much of this team's work?

as somebody who worked in this org: they are mostly focused on research, enablement and getting new tech used by operators of different sizes. Low level network tech is not sexy and doesn't really makes news

That makes sense, thanks!

I guess they must have something if Amazon are buying it instead of it just being abandoned by Facebook.

This makes it sound like a sports team trarde.

Amazon has significantly ramped up hiring efforts for Project Kuiper at its Redmond headquarters, with 500 employees currently aboard and 200 open positions

This makes it sound like a trade.

Amazon War: fight with dollaroni$$$

Is there any reason we need several private companies doing their own thing instead of making this a world wide project where every company can buy in?

I mean sure nobody would have done anything like that without Starlink. But it's like every mobile provider would setup their own antennas in a way.

In a broader sense, I think that redundant competing efforts to accomplish the same task is a much more reliable path to success.

This worked for GPS [0] because it was conceived as a purely military project (and still is mostly). Once established, it was opened up to civilian and commercial uses. But even so, other countries and supra-national organizations decided they needed their own navigation satellite systems, and now we have Galileo [1] and Beidou [2].

Maybe if the military had done this first, a similar path would follow (not likely though). As it is, the military is planning their own massive LEO satellite system [3].

[0]. https://en.wikipedia.org/wiki/Global_Positioning_System [1]. https://en.wikipedia.org/wiki/Galileo_(satellite_navigation) [2]. https://en.wikipedia.org/wiki/BeiDou [3]. https://www.sda.mil/us-military-places-a-bet-on-leo-for-spac...

Also GLONASS (Sovient/Russian) and a couple of regional systems from India and Japan.


You are getting downvoted but this is 100% true. Imagine if every maps provider or weather app had to launch their own satellites to offer basic functionality. A large collective investment into core infrastructure benefits everyone. Not every aspect of life needs to be a privatized free for all.

Yeah. I too totally realize that this is simply how competition works and prolly even leads to a more competitive and better market. Our mobile market is a expensive mess because of the regulation.

I understand all of that. I guess it's just wishful thinking that this could be one thing where the world gets their shit together.

It might be a more efficient network if it were a single system.

However, having the option to switch networks increases the efficiency of administration, customer service, etc.

Competition is the best way to produce desired outcomes.

While this is true in a way it is to simple. Ex. Health care, heavily regulated in Europe and hell we are glad it is if the free market would be something like in the U.S. Or credit card fees, there is to little competition in the free market we (Europeans) are glad that the fees are somewhat regulated and therefore much lower. And honestly our local internet and mobile market is not bad either, sure a more open market could lower the prices but private companies would have no interest bringing fiber or proper 4g or 5g in regions with only a handful of people, but there is. Germany for example has bad internet coverage and they given away most of the responsibility to private companies

TL:Dr it's not black and white

Yea, competition doesn't work when there are market failures. There is no competition is the us healthcare market because pricing is opaque and consumers have very limited information compared to doctors. Natural monopolies like wired internet can't have competition either.

I don't see how satellite internet falls victim to such egregious market failures.

Various reasons - capitalism, the degraded ability of bureaucracies to manage things compared to agile private enterprises as technology improves, conflicts of interest between nation states, and so forth. Government is not a great mechanism to drive innovation by itself, especially cooperative multi-state efforts. Things are better managed by governments setting good rules in place and private sector entities doing whatever they can to profit within the bounds of those rules.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact