Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

TSLA Market Action: 2018 Investor Roundtable

This site may earn commission on affiliate links.
Status
Not open for further replies.
If it breaks out of the $340 resistance we will see a double candle wick back flip crescendo. Of course we could also see a hammerhead star shoot followed by a cigar top.

I was about to say "you made those terms up", but seeing the gobbledegook that the chartists utter, I'm not so convinced...
 
You've provided the best evidence proving me wrong. Thank you!

Now, I need to search through my data to figure out why my info differs from yours.

Let's put real numbers behind it. Here's the key points from my traceroutes to various sites. I'll eliminate hops that don't meaningfully affect the delay, and boldface the first hop that leaves Iceland.

[meme@fwiffo ~]$ traceroute teslamotorsclub.com
1 dsldevice.lan (192.168.1.254) 1.411 ms 3.677 ms 3.719 ms
2 157-157-184-1.dsl.dynamic.simnet.is (157.157.184.1) 17.189 ms 19.658 ms 20.471 ms
3 rix-tg-gw.cloudflare.com (195.130.211.50) 22.262 ms 24.639 ms 25.235 ms
4 rix-tg-gw.cloudflare.com (195.130.211.50) 25.693 ms 26.587 ms 27.604 ms
5 104.25.55.22 (104.25.55.22) 30.972 ms 31.383 ms 31.393 ms

[meme@fwiffo ~]$ traceroute model3ownersclub.com
1 dsldevice.lan (192.168.1.254) 1.532 ms 1.959 ms 1.960 ms
2 157-157-184-1.dsl.dynamic.simnet.is (157.157.184.1) 18.172 ms 19.123 ms 20.095 ms
3 siminn-linx-gw-1.isholf.is (195.66.225.26) 59.098 ms 62.070 ms 62.497 ms
5 be100-1298.nwk-5-a9.nj.us (192.99.146.133) 134.363 ms 135.791 ms 136.550 ms
10 158.69.47.68 (158.69.47.68) 135.841 ms 137.222 ms 137.781 ms
11 158.69.61.11 (158.69.61.11) 138.672 ms 140.169 ms 129.617 ms

[meme@fwiffo ~]$ traceroute teslarati.com
1 dsldevice.lan (192.168.1.254) 1.275 ms 2.043 ms 2.044 ms
2 157-157-184-1.dsl.dynamic.simnet.is (157.157.184.1) 18.695 ms 20.897 ms 21.589 ms
3 teledesign-plc.demarc.cogentco.com (149.6.148.46) 61.044 ms 61.472 ms 62.687 ms
4 te0-4-0-11.rcr21.b023101-0.lon01.atlas.cogentco.com (149.6.148.45) 65.227 ms 67.659 ms 67.694 ms
5 be2949.ccr21.lon01.atlas.cogentco.com (130.117.2.13) 68.186 ms be2950.ccr22.lon01.atlas.cogentco.com (130.117.2.109) 67.689 ms 68.594 ms
6 be2871.ccr42.lon13.atlas.cogentco.com (154.54.58.185) 70.008 ms be2870.ccr41.lon13.atlas.cogentco.com (154.54.58.173) 66.011 ms 68.633 ms
7 be2982.ccr31.bos01.atlas.cogentco.com (154.54.1.117) 125.359 ms 113.439 ms be2983.ccr32.bos01.atlas.cogentco.com (154.54.1.178) 113.813 ms
8 be3600.ccr22.alb02.atlas.cogentco.com (154.54.0.221) 114.942 ms 113.031 ms be3599.ccr21.alb02.atlas.cogentco.com (66.28.4.237) 156.504 ms
9 be2879.ccr22.cle04.atlas.cogentco.com (154.54.29.173) 156.578 ms be2878.ccr21.cle04.atlas.cogentco.com (154.54.26.129) 156.568 ms 156.557 ms
12 be3035.ccr21.den01.atlas.cogentco.com (154.54.5.89) 165.408 ms 155.180 ms be3036.ccr22.den01.atlas.cogentco.com (154.54.31.89) 155.640 ms
14 be3110.ccr22.sfo01.atlas.cogentco.com (154.54.44.141) 180.286 ms be3109.ccr21.sfo01.atlas.cogentco.com (154.54.44.137) 180.970 ms be3110.ccr22.sfo01.atlas.cogentco.com (154.54.44.141) 181.329 ms
15 be3670.ccr41.sjc03.atlas.cogentco.com (154.54.43.14) 183.972 ms be3669.ccr41.sjc03.atlas.cogentco.com (154.54.43.10) 185.081 ms 186.064 ms
26 52.93.12.133 (52.93.12.133) 203.272 ms 52.93.12.67 (52.93.12.67) 203.225 ms 52.93.12.169 (52.93.12.169) 203.251 ms
27 54.239.48.191 (54.239.48.191) 240.720 ms * 52.93.240.91 (52.93.240.91) 203.240 ms

[meme@fwiffo ~]$ traceroute insideevs.com
1 dsldevice.lan (192.168.1.254) 1.650 ms 2.022 ms 2.014 ms
2 157-157-184-1.dsl.dynamic.simnet.is (157.157.184.1) 19.167 ms 20.893 ms 23.228 ms
3 teledesign-plc.demarc.cogentco.com (149.6.148.46) 59.818 ms 64.648 ms 65.143 ms
4 te0-4-0-11.rcr21.b023101-0.lon01.atlas.cogentco.com (149.6.148.45) 65.137 ms 65.129 ms 66.604 ms
6 be2870.ccr41.lon13.atlas.cogentco.com (154.54.58.173) 69.830 ms be2871.ccr42.lon13.atlas.cogentco.com (154.54.58.185) 66.144 ms 67.899 ms
9 38.140.107.42 (38.140.107.42) 119.162 ms 38.140.106.162 (38.140.106.162) 116.016 ms 118.720 ms
20 52.95.1.4 (52.95.1.4) 137.934 ms 140.198 ms 52.95.3.126 (52.95.3.126) 138.353 ms

[meme@fwiffo ~]$ traceroute cleantechnica.com
1 dsldevice.lan (192.168.1.254) 1.251 ms 2.158 ms 2.159 ms
2 157-157-184-1.dsl.dynamic.simnet.is (157.157.184.1) 19.293 ms 20.519 ms 21.010 ms
3 be201.am5.ams.nl.ip.siminn.is (157.157.55.249) 84.693 ms 84.684 ms 84.667 ms
4 157.157.55.226 (157.157.55.226) 80.430 ms 84.592 ms 84.605 ms
5 108.170.241.236 (108.170.241.236) 84.613 ms 108.170.241.173 (108.170.241.173) 84.608 ms 108.170.241.141 (108.170.241.141) 84.589 ms
8 172.253.51.160 (172.253.51.160) 133.971 ms 172.253.51.158 (172.253.51.158) 132.047 ms 172.253.51.156 (172.253.51.156) 134.019 ms
9 216.239.58.254 (216.239.58.254) 152.604 ms 153.504 ms 154.778 ms
12 * * 216.239.57.25 (216.239.57.25) 161.498 ms

So, even if Starlink could *only* get me to the place where my communications leave Iceland, it would cut out 17 to 85 milliseconds. (Ed: while GeoIP says that be201.am5.ams.nl.ip.siminn.is is in Iceland, even that's wrong, it doesn't change anything - see below)

Now, if we assume an average Starlink trip - after accounting for angles, and both higher and lower altitude satellites, etc - of 2000km - that's 2.000.000 / 300.000.000 = under 7 milliseconds. Even if you add a couple milliseconds for in-satellite switching (realistically it should only be a small fraction of a millisecond), it's clearly the fastest route to get to my destination.

The key aspect is that it's cutting out a lot of hops (including numerous direct passthroughs that don't show up in a traceroute***). These hops are A) in fibre which isn't as fast as light traveling in a vacuum, B) don't all go in the right direction, C) requires receipt, processing, and retransmission each time. Also, home users generally don't have many possible routes (like people connecting starlink would have). If some router or line leading up to where you get on a backbone is overloaded.... well, tough luck.


*** Traceroute only shows you level 3 hops. There may be hundreds of switches between them.
 
Re: FSD Discussion

I think the discrediting of possible end-to-end NN for driving is a bit premature. Based on Karpathy's blogs I would think he would try to give it a go. Keep in mind the output is really only a two variable state (wheel turn, acceleration level). There is no reason to not have this output then be monitored by an independent control algorithm to make sure it doesn't do anything crazy.

In regards to cars being autonomous on the road as long as they are safe: Well that's true in theory, but can you imagine the uproar if there are thousands of Tesla's slowly roaming the streets of a city, being "safe" while creating road rage? Not the best situation. Some minimum competency will be needed...
 
Um, no. satellite latency is pretty high (longer distance traveled (at least 120 miles more for uplink/downlink and then the additional arc length of path traveled). landline fiber is faster (even with all the routing and switching).
Sorry, incorrect. The speed of light in fiber is about 1/3rd speed of light in a vacuum. You can get from anywhere on earth to the exact opposite point with only two intermediate (and two endpoint) satellites. Theoretically, at least, Starlink will be much lower latency for long distances.
 
Good vibes:

Just got approval to manage project to install Tesla chargers at work.

From project proposal to budgeted in less than 72 hours.

And about 30 people waiting to test drive my Model 3.

Attached initial request, I was busy it’s sloppy writing have mercy.
 

Attachments

  • E6518DC0-D0DC-46C7-AAB8-DF5775D0815F.jpeg
    E6518DC0-D0DC-46C7-AAB8-DF5775D0815F.jpeg
    351.8 KB · Views: 90
Last edited:
Re: FSD Discussion

I think the discrediting of possible end-to-end NN for driving is a bit premature. Based on Karpathy's blogs I would think he would try to give it a go. Keep in mind the output is really only a two variable state (wheel turn, acceleration level). There is no reason to not have this output then be monitored by an independent control algorithm to make sure it doesn't do anything crazy.

In regards to cars being autonomous on the road as long as they are safe: Well that's true in theory, but can you imagine the uproar if there are thousands of Tesla's slowly roaming the streets of a city, being "safe" while creating road rage? Not the best situation. Some minimum competency will be needed...
Are you talking about... The Teslaminator?
 
California has two different speed limits on our freeways, 55 mph vehicles with trailers, and 65-70 mph for others. Whether those limits are observed is another matter, but people seem to manage the speed differential with trucks just fine. If a given jurisdiction can get away politically with a lower speed limit for fossil vehicles, and actually enforce it, that’ll generate a major surplus of demand for Teslas!

Really looking forward to seeing what happens in Austria. Regulatory developments like this could be major for TSLA, not to mention the global climate.

I supppot a fine for any Tesla caught being slower off the line than a Camry. :D
 
This is what many people used to californian style roads do not understand./ I live in rural UK. In my trip to the nearby town I will drive down 50% single-lane roads with zero road markings, and thick hedges on both sides. I will likely encounter a few mini-roundabouts (they are EVERYWHERE in the UK), and probably at least one dead badger in the road. Its pretty likely there will be pedestrians in the road at some point, and I may well have to slowly drive around some people riding horses.

At its ABSOLUTE widest, there is no room for me to pass a cyclist safely if there is a car coming in the opposite direction.

Needless to say cell signal is intermittent at best along this entire route.

FSD on the streets of LA is one thing. FSD in the snow where I live is likely several decades away, and I say that as a professional computer programmer with huge faith in Tesla.
FSD determines the safe path of the car. Your examples are just special things that exclude some paths from the set of possible safe paths, and I don’t see why FSD would need many more years to cover those cases.
 
Status
Not open for further replies.