Hacker News Reader

> It's impossible, it'll never work, no one will want to use it.

Of the issues I've heard tied to Starlink, none are these three.

The issues usually are: Expensive, slow, weather related connection issues, blocking sky, space junk and/or falling debri.

I'd include ethics of giving money to Musk.

The real test would be if you gave end users in the most remote areas the money directly and let them make their own decisions on how they got internet would they wait 5 years and pay for some microwave tower govt thing or go buy starlink today.

> would they wait 5 years and pay for some microwave tower govt thing or go buy starlink today

They'd probably pay to keep their electricity on for another month. Or 4 months because it's a starlink-level of cash.

They didn't get the FCC money because they couldn't get a consistent 100Mbps down 20Mbps up, not because anyone claimed it didn't provide access.

Those numbers are the current FCC definition of "broadband" and I think it's fair to use the broadband standard as the main threshold for that money. And the latency requirement is a generous 100ms, no shenanigans there. Maybe partial funding would be good for connections that can at least meet 25/3, but we shouldn't consider 25/3 to be good enough.

"Winning bidders have committed to deploy broadband to more than 5.2 million homes and small businesses in census blocks that previously lacked broadband service with minimum speeds of 25 megabits per second downstream and 3 megabits per second upstream (25/3 Mbps) as determined by FCC Form 477 data."

The change in definition to 100/20 was in March 2024. Well after the RDOF auctions.

Even ignoring that - the shenanigans are pretty clear :) The FCC claimed that starlink couldn't meet the standard because it currently does not deliver that speed. The irony though is all the competitors (who got lots MORE money) basically haven't deployed anything AT ALL, and are at 0/0 AND are very very very unlikely to ever deploy to the hardest to reach rural locations. It's all going to be games playing and cherry picking in near rural areas or crazy high build out costs.

The proof is actually in the pudding. People are chosing to pay to use Starlink and not the subsidized RDOF deployments in part because the cost per deployment is so high to very rural areas they simply are not out in all of alaska etc yet (and starlink is going to be not only in all of alaska but on most bigger boats and many planes and more).

Note the awards are pretty comical

"The CCA paper argues that funding was unfairly awarded to areas like Fisherman’s Wharf in San Francisco; parts of downtown Chicago; some of the largest and busiest airports in the world; and tech-heavy locations such as Apple Headquarters in Cupertino, Calif."

“This whole thing is a sham,” Settles said. “Having the 477 data is a joke, because you’re basically self-dealing to the incumbents because they influence what the information is.”

Settles added there should be more accountability on how FCC broadband money is used after it’s awarded. Billions of FCC dollars have been given out for many years. “And what do we have to show for it? I would contend that we have little to show for it,” Settles said.

A 2018 study by the Institute for Local Self-Reliance indicates that Form 477 reporting enables “de facto monopoly providers” to “overstate their coverage and territory to hide the unreliable and slow nature of their service in many communities.”

Having used Starlink a couple of times recently while travelling (think rural accommodations and the like such as rural areas of New Zealand and the Samoa islands) I've not been impressed with it. In every single case my 4G phone had better service from the local cellco (even where it was 1, bar of service) than Starlink. Don't get me wrong, Starlink is great when it is the only option. But I keep seeing it deployed in places there are other better options in including fibre, xDSL, 4G/5G, etc. It seems the hype for Starlink is very real, and lay people are sucked into thinking it's the best thing since sliced bread when it's not the only option (and actually not the best option). A great example is my current holiday accommodation is literally within 5G coverage and probably can easily get xDSL (if not fiber optic) but no, they go with Starlink... So damn shifty I gave up after a few days and now use a local LTE 4G provider instead at my own cost.

I've traveled internationally extensively. These days my phone just works. I have not traveled to Africa or South America though, so maybe we're just doing different traveling. I mention this because you say it should be obvious if you've ever traveled.

My provider is Google Fi.

I remember when they announced Starlink (and 3 or 4 competitors started angling at the space too) that I made a rough prediction that its never going to become a global internet, its a niche product for rural users that will immediately hit a ground station in the same country.

Despite consistent, vocal, misunderstanding from rusted on musk stans, this continues to be the truth.

In fact Starlink continues to obey every Carrier and Satellite related law in every country that it does business. Musk spends significant time shitting on the australian government, but he absolutely pays for his spectrum and he absolutely submits all requested information to the australian government.

Starlink's ground stations basically operate like a traditional ISP, just worse. To the point where to really get any traction its more or less a Vocus operation in Australia these days.

>Yes, Biden and the FCC successfully fought Starlink, they got the awards for rural connectivity revoked, and the appeal of the revocation also kept the award revoked because starlink supposedly couldn't provide rural connectivity. So at least on paper starlink has "failed" to meet the needs of rural folks.

This doesnt surprise me in the least. Where my customers use starlink in an area with any kind of population its a race to the exit. Where it is doing amazing is on remote cattle stations with near to no density. I have seen what passes for Rural in the USA and its positively urban by Australian standards. Where WISPs are professionalising they arent really becoming threatened by Starlink (where they suck they are getting creamed by them however)

>T-Mobile is betting pretty clearly that starlink WILL be able to deliver messaging via a space based backhaul.

T-Mobile probably has to meet Starlink at every base station in the country they are working on.

>And rumor has it that starlink is bringing in some real money even if the govt has determined its not technically feasible so there is some market validation of their ideas.

This doesnt surprise me. I have evaluated several alternative satellite internet providers, and starlink is extremely cheap in comparison. Lack of layer 2 services is a dealbreaker for some companies however, which is why I was still seeing 15k pa 20M sat services as of 2024. Theres some interesting things happening in MEO. Ironically, if Starlink ever hits multigig, we will probably see MEO services hit widespread availability of roughly 200M and the corresponding cost saving might throw starlink under the bus.

>space based backhaul.

This remains the untested claim. Much like Tesla cars becoming commonplace, Spaceship going to Mars, hyperloops etc, the actual stated goal of the company is what lies unachieved.

My rough calculations were that space based backhaul is prohibitively expensive, and doesnt benefit well from scale out of small satellites. Its all about routing density. He has a line of larger sats going up, but even then I just dont see it. Every increase in routing capacity is going to come with increases in heat and failure. The starlink constellation is good at dropping the link at the nearest ground station, its very unlikely to ever become good at sending your data to a ground station on another continent.

I was chuckling a bit at your 'just worse' comment. I do not know whether you have had any actual experience with it, but just to add a few anecdotes:

- I ordered starlink for testing even though I have Telekom DSL in a small town. Did not even install it the first day (as I had to put it up on a pole), then Telekom failed for me and lots of my neighbors. No phone, nothing. The next day I installed starlink and had Internet again at 200Mbs (faster than my DSL, albeit more expensive at first, they have since reduced price to parity).

It took 7 days for Deutsche Telekom to fix the issue. And that repeated every once in a while. I have since setup a failover with PRIMARY usage via Starlink.

- A customer of ours has Telekom and Vodafone Cable. The entire VOIP runs over Telekom , but is so slow as to be unusable (Rather big city in Germany) for that many people. They got cable as a secondary web surfing link. Number of failures over the year 2022-2023 was not funny, we had a LOT of calls to make to get this back up.

- Another customer, still big city but on the outskirts in the industrial area does not even get any good DSL at any sensible cost. He installed Starlink with us and its his primary connection for years now. Including daily and hour-long video calls.

- A school I know has had a huge number of issues with their fiber internet connection often resulting in very intermittent internet which disrupts classes etc.

So to say its always better is just not true and having some alternatives definitely dont hurt.

Tested a starlink in a depopulated outer suburb of a city and it performed worse than our government copper vdsl. Lifted that very antenna inland 400km and suddenly it was performing perfectly.

This is part of the issue. You really do need to "Suck it and see" when it comes to starlink. Is the ground station over subscribed? The Satellite? Its not like their support will tell you anything.

There's nothing wrong with having alternatives, I put my father on to starlink because his area is perfect for it. But there are cheaper alternatives in metro areas. I was slinging 2-300M 60GHz links for cheaper than starlink 6 years ago. 60GHz is getting cheaper and faster at an impressive pace. I worked on a POC in canada for 5GHz at 80MHz that was doing 150M for residential. Half the price of Starlink.