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

Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

This site may earn commission on affiliate links.
You must remember that Tesla's criteria is based off data. What you describe is how many experienced drivers tend to drive, but it doesn't mean, statistically, it's the safest behavior. In Tesla's quest for 10x safer FSD, it seems likely that some of it's behaviors might not be what we would (want it to) do...
well at least you made me think...
"Experienced drivers tend to drive"? Aren't all us old farts "experienced drivers"?
When I first started imagining have a level 5 (non)driving experience. You know, where the car drives and I just do other things like sleep. I immediately thought of how frustrated I would be. I figured it would be not some feeble-minded granny, but something akin to a Mommy that had a kid in the backseat and obeyed ALL the laws, and knew how to drive "safest." NO modest acceleration out of turns and such. But also no real knowledge and understanding of how to drive safe where the rules do not define what is safest. For example positioning yourself so you can see the cars in front of the car you are behind. And frankly it takes some mental energy to remain silent when I drive with such drivers.
But what is worst is to have to maneuver/drive a car around the majority of drivers who have no real understanding of how to drive. The legally medicated, the drunks, the phone users, the afraid, the uncaring, and the CLUELESS etc. etc. THAT is what Level 5 will give me. Reducing my frustration level. I blame my unreasonable frustration on one of my first jobs. I was a parts delivery guy back in the 1970's. Ol man Spurlock, Owner of Spurlock Diesel Parts and Service when he interviewed me told me I was hired because I had a clean driving record. He told my job would be to get the part to the mechanic as fast as I could, and he would pay any speeding tickets I got. (Some of the stories I have.) So I grew very impatient of people that did not drive correctly which caused me to arrive later than I would have.
But now level 5? I will leave earlier, and just let the machine drive.
Eventually all my crazy views on how to drive were validated for me when I got to take some Government-sponsored (aggressive) driving schools, but that made my frustration of clueless drivers worse.
So, bring me peace of mind, level 5. But that is just one road to peace of mind...all the drunks and phone abusers, and clueless will have peace of mind as well.
 
  • Like
Reactions: MC3OZ and Compton
The question of KoGuan’s flakiness can be put to rest by the filename of his KQID “Time Engine” paper:

View attachment 714524

Honestly, it sounds more like a religious doctrine based in Taoism than any physics work. Plenty of folks are able to simultaneously hold what many would consider to be absurd religious beliefs while being extremely intelligent.

So I would just let the man have his personal beliefs, and recognize that they don't necessarily reflect on his financial abilities.
 
One of the first 10.1 YouTube videos I’ve watched was not very impressive. Probably a good idea to watch good and bad for objectivity. Kim’s style is critical and she does NOT give FSD a long leash just to see what it might do, rather she intervenes quickly and often. Either way, watch the first 3 to 4 minutes to see fairly poor performance in busy city driving. BTW, not all bad, there was excellent interaction with an illegal cyclist around 5:20…
 
One of the first 10.1 YouTube videos I’ve watched was not very impressive. Probably a good idea to watch good and bad for objectivity. Kim’s style is critical and she does NOT give FSD a long leash just to see what it might do, rather she intervenes quickly and often. Either way, watch the first 3 to 4 minutes to see fairly poor performance in busy city driving. BTW, not all bad, there was excellent interaction with an illegal cyclist around 5:20…
Her v10 errors are being trained right now and we wouldn't see her v10 improvements until next release. Current improvements are from internal v10 users (employees)which still has zero error correcting from the public Beta testing fleet just bear in mind.
 
Most people expect a German left-wing coalition including significant number of Greens to win the national election today. And perhaps policies more favorable to electric cars and renewable energy.

Exit polls suggest that may not be the case. 5% is the threshold to enter the Bundestag.

1632672663403.png
 
It would benefit Tesla at some point to make an announcement that FSD beta is now available on car dashboards for those who buy FSD via fully paid or a sub.

That point is likely to be in Q12022 (or whenever Dojo becomes operational). Telsa is primarily doing this Beta button to increase the amount of training data it receives, but they have limited abilities to process that data right now. So I wouldn't expect Tesla to "open the floodgates" until they're ready to "drink from the firehose". :p

Just as important is the Insurance data they will collect from the fleet (I expect a large fraction of FSD owners/subscribers will accept the data release conditions associated with pressing "the Button"). But equally, I expect that button to result in a limited increase in FSD beta recipients until Dojo is deployed to deal with that deluge of data.

Cheers!
 
I wonder if parking and logistics will move to the new area being prepared, and then building construction continues where parking and logistics is currently.

Lol, you think a parking lot needs 5m of vertical fill? No, this is for heavy foundations. They'll need another logistics yard for the 2M Models 2 per year that Phase 3 will be cranking out by 2024-ish.

snapshot.jpg


I hope by 2024 Tesla FSD can automate much of the factory-lot logistics. Would be especially nice to see new cars drive themselves to the Port, or even onto tractor-trailers.

Cheers!
 
Last edited:
I don’t think people understand what exactly Tesla is aiming to do with their new safety score. Yes, it potentially provides a way to gate entry into the FSD Beta program, but that’s more of the carrot rather than the actual purpose. I think we can all agree that Tesla would have better ways of screening for FSD babysitters if that was the main purpose such as monitoring driver attentiveness/phone usage/etc via the cabin camera and interventional/corrective abilities.

The real magic behind this is that it offers a glimpse into how Tesla intends to deal with regulators on autonomous driving, and that is to essentially create a new set of parameters that define safe driving. Until now, the only metric that Tesla or anyone else has to measure safety has been accidents per million miles. As we all know, this can be a very flawed statistic, with the least of the problems being that it requires a very large data set and lots of time to collect that data (Tesla has the advantage here but will be severely hamstrung if this statistic remains the gold standard). Why not directly measure driver safety with parameters that most correlate to accidents, and then compare auto pilot/FSD on those criteria? Since the score comes from Tesla drivers, it eliminates all the confounders critics of Autopilot stats enjoy citing, such as vehicle age, driver socioeconomic status, etc. This allows a much more rapid assessment of FSD progress [from a safety standpoint] instead of waiting for millions and millions of miles on each iteration of the software. In that sense, folks using the new safety score truly are helping advance FSD, even if it’s not via the dataset you were thinking you’d be a part of creating.

And no, I don’t think an opt-in with the only carrot being a potential insurance discount would have gotten a good enough take rate to create the dataset Tesla wants.
 
That point is likely to be in Q12022 (or whenever Dojo becomes operational). Telsa is primarily doing this Beta button to increase the amount of training data it receives, but they have limited abilities to process that data right now. So I wouldn't expect Tesla to "open the floodgates" until they're ready to "drink from the firehose". :p

Just as important is the Insurance data they will collect from the fleet (I expect a large fraction of FSD owners/subscribers will accept the data release conditions associated with pressing "the Button"). But equally, I expect that button to result in a limited increase in FSD beta recipients until Dojo is deployed to deal with that deluge of data.

Cheers!

Absolutely agree. Should have written out "at some point" with more emphasis. :)

Even now, I am sure there is appreciable upsell for those wanting in on the beta. A hypothetical 25,000 activations * 200 just in September is a cool 5 million straight to marginal Q3 profit.

Now you mention insurance, that is tremendous upsell opportunity.

"Based on your safety score, five clicks should save you five percent!"
 
Until such a time as when FSD is L5 and widely deployed….
Tesla‘s Safety score really needs to be on the main screen/GUI for real-time feedback. Then Tesla can “train” insurees closed-loop style with what they statistically deem riskier. Right now we see data after the fact and try to guess which portions of our drive might be associated with poor scores. While this closed loop training method might not be everyones cup of tea, it’s fantastic for the liable insurer…
 
Until such a time as when FSD is L5 and widely deployed….
Tesla‘s Safety score really needs to be on the main screen/GUI for real-time feedback. Then Tesla can “train” insurees closed-loop style with what they statistically deem riskier. Right now we see data after the fact and try to guess which portions of our drive might be associated with poor scores. While this closed loop training method might not be everyones cup of tea, it’s fantastic for the liable insurer…

Insurers have been trying to do this for some time by having folks agree to place a tracking device in their car or a phone app. I don’t know about the success of these programs but I’m going to guess they aren’t getting a lot of opt-in, and rather imperfect data out the other end. Tesla has the ability to measure and record parameters that these devices cannot (eg following distance). As such, the safety score is the first iteration of this methodology that could actually work accurately in terms of precisely risk-stratifying drivers. This has obvious implications for compelling insurance product but that’s a side effect IMHO.

It’s more about redefining safety statistics for the entire auto industry and using that to dictate (as much as possible) the discussion with regulators. If I am a self-driving competitor, or auto OEM, I would be wise to ensure, starting immediately, that my vehicles can measure the parameters Tesla is incorporating into this driver assessment.
 
This video is so unreal. Elon does not talk this way. He must be reading a prepared statement. It was not prepared by him. He is reading it like a robot.

What is going on?
Elon wisely steps carefully in this tightly scripted piece. Notice how polite and effusive his wording is. He‘s even thankful to regulators for restricting Chinese driving data to China and offers a specific example of what little goes abroad. "If you bow at all, bow low" as the Chinese proverb goes.

Two major forces are cause for the care:

The first is the clampdown on entrepreneurs in China.* (Can the Chinese Communist Party throttle business leaders without throttling commerce? Enquiring minds want to know. Hint: Evergrande‘s travails are a symptom not the disease.).

The second is the sharpening great power competition between China and the US. Politicians everywhere are stepping lively. I don’t think Elon needs to worry overmuch on this score. China wants EV’s because of its relative strength in coal and weak access to oil. Also, though currently EV’s are a minute part of the relationship, China is presumably planning to dump their EV’s in the US market. Still it doesn’t hurt to be polite.

Since you’re there, I’ll observe that Germany needs to both step lively and with care. It likely needs access to the Chinese EV market if any of its legacy OEM’s are to survive. However as with the Russian gas pipeline, this may pull Germany eastward towards the authoritarian regimes and put further strains on the fissures within the EU as well as on the Atlantic alliance.

* There’s a new book out that dishes the dirt on doing business at the highest levels within China. It’s fascinating, chilling, and a quick read.

 
Last edited:
Insurers have been trying to do this for some time by having folks agree to place a tracking device in their car or a phone app. I don’t know about the success of these programs but I’m going to guess they aren’t getting a lot of opt-in, and rather imperfect data out the other end. Tesla has the ability to measure and record parameters that these devices cannot (eg following distance). As such, the safety score is the first iteration of this methodology that could actually work accurately in terms of precisely risk-stratifying drivers. This has obvious implications for compelling insurance product but that’s a side effect IMHO.

It’s more about redefining safety statistics for the entire auto industry and using that to dictate (as much as possible) the discussion with regulators. If I am a self-driving competitor, or auto OEM, I would be wise to ensure, starting immediately, that my vehicles can measure the parameters Tesla is incorporating into this driver assessment.
No one trusts insurance companies. (Exception: They trust them to deny claims.)