Ludicrous launches confirmed 100% safe!
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable
- Thread starter AudubonB
- Start date
Ludicrous launches confirmed 100% safe!
ruralectric
Member
Another data point: I've got the latest ios app download and still no safety score, but the FSD button presented in the car and clicking it gave me the "in the queue" response, so pretty sure that the vehicle and/or mothership knows the score and the info displayed on the app is just informational.
Have you driven it since opting in? My score showed up 15 minutes after I got back from a drive.
I couldn't find where speeding was a problem but have trouble thinking that's not the criteria at least through intersections. I bet this is just a subset of reports available to Tesla internally, we're just getting some tips is all to keep the insurance tech advantage they already have.Interesting how differently those criteria are weighted. Forced Disengagement is 20X worse than Aggressive Turning and 200X worse than Unsafe Following. Looks like if you can avoid Forced Disengagement and Hard Braking that should be enough. Also notice no penalty for 0-60 launches (unless you brake hard afterwards)
This could also be an experiement in driver awareness. The act of observing, including immediate +/- feedback, will make us safer drivers. That data could be useful to Telsa's defense should it be necessary, as well as others who actually care about safety (vs just using it as a tool for generating fear).
Tesla stated that they will not upload any data telemetry until we allow it (the Button). Why wouldn't they already have our data on past driving archived in the vehicle, and also uploaded on this same permission today? (I'd have to read it again to confirm.) This would create a pre/post human performance dataset predicting safety improvement just from being observed (I assume).
This is Efficient and Effetive Training for sure. Causing someone to change behavior is the very definition of a Level 3 Learning Assessment (ya we have levels too). A Level 4 Learning Assessment is defined as the business level impact in the form of $$$ (fewer accidents, or maybe even lawsuits won). I hope they did/do this Pre/post collection. Solid L4 data is extremely difficult to measure and they could have it already. Training pros globally would consider this a serious score on a massive scale as time correlates better and better with the assessments. (Sorry, maybe geeking out on training a bit. There is no Level 4 assessment if there is no pre-testing, period.)
This is Efficient and Effetive Training for sure. Causing someone to change behavior is the very definition of a Level 3 Learning Assessment (ya we have levels too). A Level 4 Learning Assessment is defined as the business level impact in the form of $$$ (fewer accidents, or maybe even lawsuits won). I hope they did/do this Pre/post collection. Solid L4 data is extremely difficult to measure and they could have it already. Training pros globally would consider this a serious score on a massive scale as time correlates better and better with the assessments. (Sorry, maybe geeking out on training a bit. There is no Level 4 assessment if there is no pre-testing, period.)
DragonWatch
Small FootPrint
Forward Observer
We are relaxing forty-five miles away from Xena II (MX) ~ upgraded the software this morning over the air (2021.32.22). Tesla app on the iPhone has been updated along the way.
Sitting here and reading this blog raises two questions to beat the evaluation during the seven days. First, can I Summon the car? Clearly Xena can drive better than me over those forty-five miles.
Second, and by far the best is to ensure I give Sandy the credit she deserves by ensuring Xena believes Sandy is driving. Win, win for me. I do not get blamed for the poor score.
We are relaxing forty-five miles away from Xena II (MX) ~ upgraded the software this morning over the air (2021.32.22). Tesla app on the iPhone has been updated along the way.
Sitting here and reading this blog raises two questions to beat the evaluation during the seven days. First, can I Summon the car? Clearly Xena can drive better than me over those forty-five miles.
Second, and by far the best is to ensure I give Sandy the credit she deserves by ensuring Xena believes Sandy is driving. Win, win for me. I do not get blamed for the poor score.
What creeps me out more but I think is the only reasonable hypothesis given how these journalists behave almost across the board is that it is the job itself that turns them into these narrow visioned pessimists. I think they pick up on the latent incentive structure and internalize it to remove cognitive dissonance. CNBC is going to be rewarding engagement and so if she gets more engagement by being negative and especially if she can shift to a moralizing tone about safety and greed then that works even better. Eventually instead of that being just a conscious stance to take for the job it becomes her actual thought process because it is easier. It seems to me almost every writer that principally covers just Tesla ends up this way. I don't think it helps that Tesla fans can be obnoxious because then she has a personal reason. So what is more creepy about this is that it suggests these monstruosos can be manufactured rather easily.
I finally made it through Walter Lippmann's public choice and he rips apart newspapers back in 1922 in quite familiar terms. They are fundamentally businesses not bastions of truth and objectivity. He also puts a lot of ink into the basic fact that advertisers are the real clients though he doesn't stress direct competitors trashing each other so much as needing to perform in a way that generates sales for the advertisers. So yeah trashing Tesla might accomplish more purchases of GM. It presumably makes GM happier to see negative Tesla articles. I'd love to see how exactly this all plays out behind the scenes in editor/publisher emails etc. Might be as subtle as simply being told 'no that's not newsworthy' every time she suggested maybe writing about how FSD may save lives. I assume you figure out the game pretty quick and then what are you gonna do? Quit?
Nope, not sure why. Only tried the 3, will try the X later.
jhm
Well-Known Member
Cool! I could tell you were sincere in your effort to understand these things. So I didn't mind writing detailed response.
At the pace you suggest, one thing remains that could slow things up. It may be important to test for a whole year to get exposure to all the seasons. Inclement weather is pretty important to get right.
The 115 appears just in the final safety score formula (it is not mentioned as a max in the description), and the safety score itself can only be 100 at max. For the best possible PCF, all of the exponents would be zero yielding a PCF of 0.682854. Multiplying this with 22.526504 in the final formula, you get 15.38 making the Safety score 100 (plus some rounding error).
EDIT: @evermore was quicker!
Last edited:
jhm
Well-Known Member
You are correct. I would expect that regulators would want to analyze interventions. Tesla, too, can evaluate how good its system is by the frequency of interventions. Sometimes intervene when AP is taking a curve too wide or too fast for my comfort. This is not necessarily an indication that AP was unsafe, just not smooth enough for my comfort. Even so, as the FSD becomes so sophisticated that occupants are never uncomfortable with the driving, never feel like they need to take over, this is a really good thing.
Still from a safety testing perspective, interventions should probably be analyzed to determine if any immanent safety hazard was present. If so, FSD needs to be fine tuned to anticipate that more quickly than the human driver does. Human drivers are still quite capable of anticipating risks before FSD kicks in. So it is an opportunity for the AI to learn more thoroughly from human drivers.
Thanks for posting.
I updated all 3 Teslas; Model S, Model X and Model 3 this morning. Each vehicle is given an individual safety score so we'll see how my wife's car compares to mine. This is not only exciting but fun. So off to driving the S next.
I don't have Tesla insurance but am considering it for all 3 although our current insurance is pretty cheap right now as we are good drivers. I wonder how our safety score will rate?
Buckminster
Well-Known Member
V10.1 in 10 hours - Christmas has come early!
iOS here and the new app version is available, but without a button to press, all looks the same as before
jhm
Well-Known Member
Ah, this model looks like the work of a statistician, not AI! I have to wonder what score FSD is able to maintain, which would depend heavily on avoidance of forced disengagements.
Tesla psychology on the FSD testing seems good. They are not starting with a hammer, but rather lots of positive feedback ( I scored 100 ).
Important to not look past their “game psychology” advantage when compared to their competitors of note.
When trying to figure out who their competitors of note are, I struggle.
Apple’s money making App Store machine was devised by Jobs. He is not there any more. I don’t think they are near as fast.
Amazon wins with scale, and they are pretty smart as well. Not as good at gaming, and software discipline on Alexa is … mumble mumble.
Google has some fundamental “managed as a country club” issues. This happens when a single application pays the freight and management is blind on what matters when promoting. I think they literally can’t see when they are making a mistake. Like the information is coming in right where the optic nerve lands and there are no pixels available.
Speaking of pixels, maybe if Nintendo bought Honda in 1985 when old Man Honda was handing over the reins there would be a worthy competitor…. That time has passed.
In summary, the FSD testing psychology exposes a huge gap between Tesla and the competition. The truck with Betty Boop eyes might find a niche…
I find this disconcerting as gaps this large encourage concentration. And have to be tested/treated as confirmation bias.
This is the kind of gap not seen in nature. I struggle to pattern match on this and do not know what is going to happen.
There is my review of the FSD test experience. For all you people who don’t like being evaluated.
Applying gaming psychology to improve traffic safety [that is a euphemism for saving lives and limbs] is a great service to society, and a pretty smart idea.
Important to not look past their “game psychology” advantage when compared to their competitors of note.
When trying to figure out who their competitors of note are, I struggle.
Apple’s money making App Store machine was devised by Jobs. He is not there any more. I don’t think they are near as fast.
Amazon wins with scale, and they are pretty smart as well. Not as good at gaming, and software discipline on Alexa is … mumble mumble.
Google has some fundamental “managed as a country club” issues. This happens when a single application pays the freight and management is blind on what matters when promoting. I think they literally can’t see when they are making a mistake. Like the information is coming in right where the optic nerve lands and there are no pixels available.
Speaking of pixels, maybe if Nintendo bought Honda in 1985 when old Man Honda was handing over the reins there would be a worthy competitor…. That time has passed.
In summary, the FSD testing psychology exposes a huge gap between Tesla and the competition. The truck with Betty Boop eyes might find a niche…
I find this disconcerting as gaps this large encourage concentration. And have to be tested/treated as confirmation bias.
This is the kind of gap not seen in nature. I struggle to pattern match on this and do not know what is going to happen.
There is my review of the FSD test experience. For all you people who don’t like being evaluated.
Applying gaming psychology to improve traffic safety [that is a euphemism for saving lives and limbs] is a great service to society, and a pretty smart idea.
Last edited by a moderator:
Thumper
Active Member
I hope Tesla will not hold it against me, but I "run" this stop sign regularly.
Just to close on this, I can report that it does not record without the correct 4.1 Android App is installed. We can draw some theories on this.
Edit: Others did not report this same outcome. So ignore.
My first drive data (after the Button) is not shown in my daily, only the 2nd drive (as "1 of 1"). This implies it's not tracking unless you have the correct App installed. This means it's possible your phone's acceleration telemetry is combined with vehicle red flags to create the report. It's only speculation, but maybe don't throw your phone around the car or let it slide off the dash on a turn?)
After the 4.1 App update and a drive, it shows my second drive only. 100% so far
After the 4.1 App update and a drive, it shows my second drive only. 100% so far
Last edited:
Similar threads
- Locked
- Locked
- Locked
