Chat > Tech Chat
Why Starlink is doomed to fail
celso:
I don't see how Musk's bold claims affect your Tesla. Isn't it fast? Doesn't it have the range? The charging network isn't better than the alternatives? How does a car being late to production affects your ability to drive your car?
The only way I could see someone being "scammed" is if they bought the self driving hardware early on. If they do it now, I find it a bit harder to be sad for them because even I know their self driving isn't ready... and I don't even drive. Be it a car, computer, phone, etc, buy it for what it is, not what it may become in the future.
What's left? So Musk made a bad business decision for Tesla by buying SolarCity to save his cousins(?) rear and now doesn't sell enough solar panels? I don't own Tesla shares though. If I did, I'd probably look into this stuff before investing. This may say a lot about Elon Musk the person, but again, it doesn't affect the performance of your car or home battery.
Regarding SpaceX, you must understand that they said similar things about early versions of Falcon. There's no way a small private company can create a rocket. There's no way they can make their own engine. There's no way you can land parts of it! Transporting cargo and people to and from the ISS? No way! Yet, they're doing it. They seem to deliver their contracts on time or at least faster (and cheaper) than Boeing. And refurbishment times keep dropping (slowly). For example, recently one of the boosters was used 21 days after it landed (which means that they do it in less than 21 days). Again, keep in mind that people had doubts that this could be done. That's why I give them the benefit of doubt.
Starship... look, maybe it's a scam, but those videos pick the wrong things to "prove" it's going to fail. Things change? Of course they do. Why are people surprised that it looks different from the computer rendering? Not only it's still in early development, but anyone that pays attention to these new rocket companies knows that SpaceX doesn't work like the old space companies. They iterate fast, test often, and blow up a lot of stuff. You don't have to like it, but that's how Falcon, Merlin, and Dragon were developed and... it works. Is it late? Yes, but late doesn't mean ponzi scheme and prioritising launch and landing on land doesn't mean that they'll never land on some old oil rig (and plans can't change?).
While skipping parts of that video, I landed where he says something about engine melting (not a secret, Musk said it himself). Obviously a problem, but they developed engines before, it's not like they don't know how to do it. And they're still developing this one... v1 and v2 are not the final design. Look at
this video from last year. The first minutes don't make it clear that it's not the final version? If this is a scam, then they're being very public about it.
Regarding Starlink, the calculations are guesses. We can guess about satellite costs or how much a Falcon 9 launch cost, but we don't know for sure. How can you make a proper prediction, when you don't know how many users the version 1 and v2 of their satellites can handle? How much more capable are the new v2 (larger, heavier, apparently big improvements in capacity)? And economically, how many millions will some stock trading company in the City of London pay for something that can make information travel the Atlantic a few milliseconds faster than the alternatives? If the reports coming from Ukraine are true, how much money will they make from military contracts? We need to take this into consideration when guessing. He doesn't and even if he did... he would still be guessing. We don't know their real costs, we don't really know how much certain users are willing to pay, and there are things we can't predict. I wouldn't trust his (or anyone else outside SpaceX) calculations that much.
The number of satellites is a problem, but singling one company out when others are or will do the same doesn't look good. Why is Starlink bad, but OneWeb or Kuiper okay? And why are you both assuming most of those satellites can't be de-orbited at end-of-life? Do they have a flaw I'm not aware of? Looking at some of the satellites on sites like https://satellitemap.space/, they seem to be able to change altitudes fairly quickly. What am I missing?
You said above that the channel debunks things using facts, but those facts are sometimes twisted to fit a narrative. The point I made about the speeds, for example. The fact he presented about the speeds being bad is probably true, but he's comparing it to fibre in cities when Musk - which makes bold claims - says it's not competing with that (as quoted on my previous comment). Not only that, but those aren't the speeds most users are seeing (see the Starlink reddit sub, Adrian's blog post, etc). Knowing this, it's hard for me to believe the narrative because clearly he's cherry picking facts to fit his narrative.
All this makes me question the intentions of the person making those videos. I mean, any neutral person can see that Tesla had failures (maybe solar roof doesn't work!) and that they are strong in some areas (charging network, casting of parts for their cars, developing their own battery cells, etc). Starship may be late, but Falcon 9 works, they can land anywhere, Dragon works, refurbishment times are dropping, etc. Does Starlink success depends on if Musk really cares about Mars? And why is this guy focusing only on the negatives and ignoring anything that works? Has he shorted Tesla stocks or something like that?
Some of the points made on those videos are valid, but some are wrong and some are misleading, and that makes me raise my eyebrow. Then it seems that there are two groups when it comes to Musk: one thinks that everything he does is bad and that he's a con artist while the other thinks Lord Elon is a genius that can do no wrong and will defend him no matter what. This guy is in the later group and I personally think both groups are wrong. The truth is somewhere between the two.
The headlines I see on my news feed makes me believe that Musk sometimes is like a train without brakes... it can end badly. I'm somewhat aware of SpaceX risks. I also know that Tesla is about to experience competition from legacy car manufacturers, so there's a lot of risk for the company. And I'm happy to change my view if there's some new information. Until then, Musk's companies track record is to deliver some things, fail to deliver other things, they're often behind schedule, etc. Nothing has changed recently for me to change my mind about them.
(edit: I linked to the wrong YouTube video. It's fixed now. The forum is embedding it for some reason though...)
Alex Atkin UK:
--- Quote from: celso on June 11, 2022, 08:33:14 PM ---I don't see how Musk's bold claims affect your Tesla. Isn't it fast? Doesn't it have the range? The charging network isn't better than the alternatives? How does a car being late to production affects your ability to drive your car?
--- End quote ---
Again, its not about if it delivers, its about can it deliver long-term? They forbid you from making any modifications to the car and repairs are not economically viable.
--- Quote from: celso on June 11, 2022, 08:33:14 PM ---The only way I could see someone being "scammed" is if they bought the self driving hardware early on. If they do it now, I find it a bit harder to be sad for them because even I know their self driving isn't ready... and I don't even drive. Be it a car, computer, phone, etc, buy it for what it is, not what it may become in the future.
--- End quote ---
Its literally called Self Driving, but they admitted its Level 1, which is advanced cruise control at best.
Any system that self-drives "some of the time" but will randomly fail, has no business being used on public roads. It increases rather than decreases the chance of accidents as it flies in the face of hazard perception, how can you react to a hazard developing if you don't know its a hazard until the self driving system fails? I haven't driven much, but enough to understand how that's just a terrible situation. If you're relying on the car to drive itself the majority of the time, you're not going to react as quickly when something goes wrong which can be the difference between life and death.
What about Tesla trucks that never were manufactured? What about faulty battery management controllers but they just wrote software to hide the error messages?
Also Tesla aren't good at making battery cells, they buy them from Panasonic and forbid you from replacing the cells or repairing a bad battery management controller, making a new battery pack rather uneconomical.
As for Starlink vs other services:
--- Quote ---The OneWeb satellites fly at higher altitudes than the Starlink spacecraft. The difference in architecture means OneWeb can reach global internet coverage with 648 satellites, a significantly smaller constellation than Starlink.
--- End quote ---
Kuiper definitely sounds just as bad as Starlink however.
I still think that is bad mind you, when competitors can achieve global coverage with THREE satellites. When you're dealing with something as risky as space junk, minimalism is a must. Its not a case of if we'll no longer be able to launch satellites, but when. The more go up, the sooner that day will come.
Overall when it comes to "what about the stuff Musk has done that works", again, its not about if it works or not - you can make a lot of things work with brute force, throwing pots of cash at the problem. That's no good if you can't cover the running costs without constantly begging for government hand-outs and conning investors, which is how he is supporting everything right now and the numbers just don't add up for that changing, ever.
Alex Atkin UK:
Just stumbled onto this one from someone who used to believe the hype that basically goes over everything.
https://www.youtube.com/watch?v=iC3pnJmYaxA
celso:
--- Quote from: Alex Atkin UK on June 11, 2022, 10:22:35 PM ---Again, its not about if it delivers, its about can it deliver long-term? They forbid you from making any modifications to the car and repairs are not economically viable.
--- End quote ---
From a customer point of view, Tesla is selling cars, battery packs for homes, charging stations, etc, and things seem to work. So they're delivering in that regard. Will they keep delivering when Toyota, Volkswagen, etc, all go electric and start releasing affordable EVs? I don't know, but that seems to be a problem for investors and Elon's pockets, not the consumer.
Starlink, I really have no idea how many users the 2nd gen of their satellites can support and if the satellite-to-satellite connection will work. And that's why I think we should be cautious when looking at overly negative or positive predictions. If we guess that one satellite can support 50 connections and it actually can support 200, then our calculations will be wrong.
Regarding repairs, both remind me of some Apple products. They're expensive, not the easiest to repair, and they're certainly not for everyone.
But you don't need Tesla to change your tires, fix your suspension, change bulbs, or to fix a small bump on your car. Check Bjørn Nyland's (from Norway) channel on YT, some of his maintenance is made on a regular shop that also works on electric cars. This change from ICE cars to electric means that those providing car services will have to learn and change... right now you have less options available, but it's not because it's impossible to fix the cars.
Now, some repairs are too expensive (maybe it's smarter to buy a cheaper car?) and need to be fixed by Tesla. But it's not like you can't salvage an electric motor, install a "refurbished" battery, replace a motherboard repaired by a 3rd party, etc. They won't take responsibility for it, but no brand does when we use after market parts. There are lots of videos of repair shops and "car guys" about this on YT too. Eg:
https://www.youtube.com/watch?v=DvlvNS9kQeU
Starlink terminals, I believe you only really need the antenna and the power-over-ethernet connector. Maybe their router is bad, but I using my own router too because I don't like the one PlusNet sent me.
I didn't know, but the cable I talked about on the previous post seems to be replaceable on the new terminal (Dishy v2). Proprietary connector, but at least you can change it.
--- Quote from: Alex Atkin UK on June 11, 2022, 10:22:35 PM ---Its literally called Self Driving, but they admitted its Level 1, which is advanced cruise control at best.
Any system that self-drives "some of the time" but will randomly fail, has no business being used on public roads. It increases rather than decreases the chance of accidents as it flies in the face of hazard perception, how can you react to a hazard developing if you don't know its a hazard until the self driving system fails? I haven't driven much, but enough to understand how that's just a terrible situation. If you're relying on the car to drive itself the majority of the time, you're not going to react as quickly when something goes wrong which can be the difference between life and death.
--- End quote ---
I believe you're a bit wrong here, if I understand what you're saying correctly. They have the "free" Autopilot, which is essentially a fancy line keeping + adaptable cruise control. From here:
> Autopilot advanced safety and convenience features are designed to assist you with the most burdensome parts of driving. Autopilot introduces new features and improves existing functionality to make your Tesla safer and more capable over time.
> Autopilot enables your car to steer, accelerate and brake automatically within its lane.
> Current Autopilot features require active driver supervision and do not make the vehicle autonomous.
And then they have paid upgrades that unlock the features you mentioned. This is from the UK website, while trying to order a Model 3:
So, I think there are a few criticisms we can make here.
The first is the name "Autopilot", which may be misleading, but I think it's widely known that the commercial planes we fly on has autopilot and 2 human pilots in the cockpit. If one is too dumb to know that, the car tells the driver to be ready to take over and if they ignore it, it will beep at them. If you try to trick the system and take a nap... is it really the car's fault? At some point we have to take responsibility for what we're doing with our car. It has some problems, but functionality wise, it's not that different from what many Toyotas (and other brands) have on cars released in the past 5/10 years.
Then there's the "advanced" stuff. This is what I was referring to when I wrote this in my previous comment:
> The only way I could see someone being "scammed" is if they bought the self driving hardware early on. If they do it now, I find it a bit harder to be sad for them because even I know their self driving isn't ready... and I don't even drive. Be it a car, computer, phone, etc, buy it for what it is, not what it may become in the future.
You're paying for something that isn't ready. If I had paid for this a few years ago, I'd be fuming. I mean, changing lines in the motorway should work fine, but there's no full self driving yet. People have been referring to autopilot as "self driving", but Self Driving was only available in 2020 with a restricted beta and expanded in September 2021.
Should you pay in advance before this is ready? I wouldn't. Should it be allowed to be tested in public roads by careless people trying to showoff? Yeah... probably not. Is the system as advanced as it needs to be? I don't think it is.
I think your criticism (edit: and points) are valid here.
What I won't do is side with morons that are not paying attention when using autopilot or similar systems. Be it a Tesla or a fancy Mercedes EV with their own level 2/3 system, they know what the system does, they learn its flaws as they use it, and yet they decide drunk drive, play angry birds, take a nap, check facebook while driving.
--- Quote from: Alex Atkin UK on June 11, 2022, 10:22:35 PM ---What about Tesla trucks that never were manufactured? What about faulty battery management controllers but they just wrote software to hide the error messages?
--- End quote ---
The Tesla Semi announced in 2017? It seems to be delayed, I think. The Cybertruck announced in 2019 is also delayed. They're saying production is supposed to start in 2023 in their Texas factory, which - I believe - they only finished this year.
Is there a reason why you're doubting this? I remember the Model 3 and Y being delayed, but they're selling them now. Again, late... but late and scams are different things.
Regarding faulty BMS, what happens in practice? Are cars shutting down randomly, catching fire? Is the "fix" creating any problems? Honest question because I don't know.
--- Quote from: Alex Atkin UK on June 11, 2022, 10:22:35 PM ---Also Tesla aren't good at making battery cells, they buy them from Panasonic and forbid you from replacing the cells or repairing a bad battery management controller, making a new battery pack rather uneconomical.
--- End quote ---
I follow that Bjørn Nyland channel I mentioned above and the guy tests the different battery packs Tesla and other brands use. I'm working from memory here, but I believe Tesla uses Panasonic for their "performance" cars and LG for their "standard" cars. I don't know if it's all models, but the cheaper Model 3 also uses LFP batteries from CATL.
The batteries I was thinking of are the ones they talked about in 2020. They're using a new size (4680) and different process (from a company they acquired). This is made by Tesla, not Panasonic. (Search for Tesla Battery Day 2020 on YouTube to see the presentation.) Apparently Panasonic will also make their own bigger batteries for Tesla, but that seems to be the same deal as the current 18650 batteries they buy from Panasonic/LG.
Regarding repairs, I don't know how it will be with the new cars where the battery pack is part of the structure of the car, but at least until now you could remove the battery, replace modules, etc. If you'll find a company in the UK doing this, that's a different question, but they're not going to call the police on you. They just don't want to have anything to do with it if the car catches fire.
https://www.youtube.com/watch?v=F-B_8oMZNeI
To be clear, Tesla repairs are not easy. Your point is valid.
What I don't know is if they're better or worse than EVs from VW, Mercedes, Audi, Porche, Renault, Volvo, Polestar, etc. I'd be surprised if it was as easy as ICE cars.
--- Quote from: Alex Atkin UK on June 11, 2022, 10:22:35 PM ---As for Starlink vs other services:
Kuiper definitely sounds just as bad as Starlink however.
I still think that is bad mind you, when competitors can achieve global coverage with THREE satellites. When you're dealing with something as risky as space junk, minimalism is a must. Its not a case of if we'll no longer be able to launch satellites, but when. The more go up, the sooner that day will come.
--- End quote ---
I agree that more satellites is worse.
There two things that we should keep in mind, though:
- Starlink, Kuiper, etc, are in Low Earth Orbit. Pings of 20/30ms if coverage is good.
- ViaSat, Hughes, are in Geostationary orbit. You'll find pings online between 400-600ms.
Now, that image from ULA's Tory Bruno is useful if we're worried about the Kessler syndrome.
We've been sending these big, heavy, and expensive satellites to GEO for years, and they'll be there "forever" (in the "graveyard" orbit) and I assume it's harder to reach them if we start cleaning up space because it's very far away. How good is this?
And then there's LEO where the drag eventually brings stuff down after just a few years. It seems we're about to witness an international commercial space race over the next few years... I could be wrong, but if we're going to have a Kessler syndrome, this seems the best place for a "wake up" call. The chances of me being wrong are high though :P
Then we have the quality of service. The alternatives with 3 or 4 satellites all have high latencies. Viasat seems to be the best option here, so I looked for speed tests on YouTube... and it's bad. 1-20Mbps and pings of 600ms. The best I could find was this result, which reached 74Mbps down under their 100GB "premium data". But the ping... almost 600ms:
https://www.youtube.com/watch?v=ADgdRJmow34
Searched online for ViaSat vs Starlink and the first result on DuckDuckGo is this blog post from someone in Arizona. I'd recommend reading it, but essentially the ping is around 700ms with ViaSat which causes problems with calls, SSH doesn't work well, etc. Starlink? Over 300Mbps down, 20Mbps up, ping 33ms, jitter 7ms. During the beta.
I also checked /r/ViaSat. This is what I see:
The first post is calling the service a scam (ha!). Paying 150 dollars a month (Starlink was 100, now 110) and it seems to be slow. There's a reply, which I find to be a bit... funny:
> When I had Viasat I had no problems streaming Youtube in at least 480p or occasionally streaming sports. I just had to be careful that i didn't go over priority data.
And here am I complaining about my FTTC connection which plays 1440p YouTube just fine (the default picked by them for this high dpi display). :D
There's also this post from someone asking if is actually that bad. I'll let you read the replies.
I think it's fair to say that ViaSat is way worse than Starlink. Some can get good speeds with ViaSat until a certain point, but there's no way around the latency. I've never used either, but LEO and GEO internet seem to be very different things.
The worse results I see from residential Starlink are like slow/average 4G (40Mbps/10Mbps) while the best ones are up there with good 4G (100-300Mbps/40Mbps). In some cases, even a "slow 4G-like" Starlink is still many times superior to ViaSat.
I don't mean to say with this that we should ignore the problem of so many satellites in space, but you can't do Starlink with 5 or 10 GEO satellites.
--- Quote from: Alex Atkin UK on June 11, 2022, 10:22:35 PM ---Overall when it comes to "what about the stuff Musk has done that works", again, its not about if it works or not - you can make a lot of things work with brute force, throwing pots of cash at the problem. That's no good if you can't cover the running costs without constantly begging for government hand-outs and conning investors, which is how he is supporting everything right now and the numbers just don't add up for that changing, ever.
--- End quote ---
Do you mean government contracts? SpaceX seems to operate in the same way (but cheaper) that older space companies do in the US. I don't know enough about US politics to make a comment about that. I don't think SpaceX is making a profit and they're investing a lot in Starship and Starlink. It could be a problem, but I don't know if they're conning investors or running out of money.
If it's EV discounts, governments in the "west" support EVs in these early years because they are more expensive than ICE cars. But this is for everyone, not only Tesla.
If it's the money that Tesla received after the 2008 financial crisis, it was paid back already. I don't know if they received more since then.
Tesla made a profit of "$3.3 billion" from "$18.7 billion in revenue" in Q1 2022... while everyone is struggling with a shortage of chips and other components. Considering that they finished 2 factories, are investing in their own battery tech, keep expanding their charger stations world wide, the work in AI/self driving/custom processors (doesn't sound "cheap"), and working on new cars, is it that bad? Since I don't follow companies and stocks, I don't know if 3.3B is good or bad in this business.
I think the main difference between us is that you're looking at this from an investor point of view and I'm looking at it as a possible customer. If I get what I pay for, I'm happy.
--- Quote from: Alex Atkin UK on June 12, 2022, 01:50:44 AM ---Just stumbled onto this one from someone who used to believe the hype that basically goes over everything.
watch?v=iC3pnJmYaxA
--- End quote ---
It's a long video, I didn't watch most of it, but why does it matter for you as a customer if the guy was poor or rich, if he was a school dropout or not, if he was the creator of Tesla or bought it early on, etc?
Windows doesn't crash more often just because Bill Gates dropped out from... Harvard University... was from a wealthy family, got lucky with IBM because his mom worked there, tried to control the internet, etc.
I can't afford a Tesla and Starlink isn't for me, but if I was going to buy anything from them, I'd look at the product. Musk's good PR doesn't lower Starlink's latency or increases a Tesla's range.
My suggestion: don't invest in his companies if you don't trust him, don't pay for self driving until it's ready, and look at their products like... products.
I don't know if I'll have the time to write another reply like this, so if I stop posing, it's not because I ran way :P
Alex Atkin UK:
--- Quote from: celso on June 12, 2022, 03:27:05 AM ---The first is the name "Autopilot", which may be misleading, but I think it's widely known that the commercial planes we fly on has autopilot and 2 human pilots in the cockpit. If one is too dumb to know that, the car tells the driver to be ready to take over and if they ignore it, it will beep at them.
--- End quote ---
Most people wont have a clue how Autopilot on a plane works whereas pilots will have had extensive training on EXACTLY how to use it, what it does and does not do. Someone buying a car will not have the same knowledge, nor do you have the same amount of time to react to a problem when you're on a road compared to a plane.
Even if it tells you exactly what it does and does not do in the manual, it has to be to a higher standard as you cannot expect someone who is just making a trip to the shops, potentially with many distractions, to be aware of everything it does and does not do at all times. I stand by my claim, its dangerous for ANY car to "partly" self drive, as it causes complacency by the driver if most of the time they do not have to steer. At most it should be able to emergency brake, park, stay in lane, maintain speed on a motorway. The driver should ALWAYS be the one changing lanes, turning corners, etc. We have decades of proof that with the best intentions, people become complacent once a majority of the time they don't have to do something, making them slow to react when they are required to. By its nature, by the time you realise its gone wrong, you may be too late to react.
Its risky to even expect a car to emergency brake:
https://www.youtube.com/watch?v=p7lp5f0aqzU
Had this been a real-world scenario with traffic coming in the other direction, things could have gone badly.
My point being, you shouldn't be promoting a system as able to do something unless it can do it 100% of the time. The driver shouldn't just have their hands on the wheel, they should be steering, braking, etc, you should NEVER be relying on the car to react period. But this is not how Tesla are promoting their tech however.
--- Quote ---The Tesla Semi announced in 2017? It seems to be delayed, I think. The Cybertruck announced in 2019 is also delayed. They're saying production is supposed to start in 2023 in their Texas factory, which - I believe - they only finished this year.
Is there a reason why you're doubting this? I remember the Model 3 and Y being delayed, but they're selling them now. Again, late... but late and scams are different things.
--- End quote ---
I can't find the video right now, but this again comes down to economics. The weight of the batteries required for a truck means the trucks can haul a tiny minority of what a diesel truck can. Musk made claims about cost per distance, but didn't factor this in, making it appear cheaper and more efficient when its not.
In fact, a lot of the problems come down to energy storage. We've pretty much peaked with battery storage, we can't make it much more efficient, the laws of physics dictate so. The only way to make electric trucks work is this:
https://www.youtube.com/watch?v=_3P_S7pL7Yg
--- Quote from: celso on June 12, 2022, 03:27:05 AM ---Regarding faulty BMS, what happens in practice? Are cars shutting down randomly, catching fire? Is the "fix" creating any problems? Honest question because I don't know.
--- End quote ---
Inconsistent range due to improper charging, that's not a small issue especially for long distance travel in rural areas where you could get stuck in the middle of nowhere.
https://www.youtube.com/watch?v=eSOHzmqLWjU
As for Viasats negatives, the problem is you're looking at it from a "this is what we want" scenario, rather than "this is what we can do".
We all want perfect broadband that works for everything, but if its not economical or damaging the environment, then we absolutely should NOT be doing that. The end does not justify the means.
Its the same kind of logic as "we need land for x, lets just cut down all the trees and kill off all the wildlife". Sure, it will wipe out all life on earth eventually, but things will be fine for a short while. Are you honestly advocating that we should just ignore all long-term problems in order for a quick fix today?
This is the crux of why I'm critical of Elon Musk, even though I wont be using any of his technology. Because it WILL impact everyone in the long term.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version