Blog Tesla Looking for Good Drivers to Test Full Self-Driving System

If you want to be a beta tester of Tesla’s Full Self-Driving system, you’re going to need to be a good driver.

Tesla CEO Elon Musk tweeted that owners will soon be able to request access via a button in the car. This will give Tesla permission to evaluate owners’ driving for seven days before they’re included in the beta testing group.

https://twitter.com/x/status/1438751064765906945

Musk also tweeted that that Version 10.1 of the FSD Beta is estimated to arrive on September 24th. The beta request button will be included in the update.

https://twitter.com/x/status/1438747500010168322

The V10 update was well-received, with some reviewers showing their cars navigating through areas that it was previously unable to complete without driver intervention. Musk has said the next version will be another noticeable step in performance for the system. 

 

[brianman]

This system is kind of flawed. I drove the past two days, now mind you I commute 100 miles round trip. 90% is on the freeway. once on the freeway I tap into FSD. Bay area driving with Bart off to the left of the vehicle which has red/green lights for the BART. If the Tesla detects the red light in its sensors it will abruptly break. going from 75MPH hard break down to 50 in seconds, all on FSD. traffic will be going smoothly, all of a sudden traffic will slow down and the Tesla will not detect the break lights sooooo what happens.....HARD BREAK all on FSD. First day score was at 68. Second day I was able to get my score up to 89. I would take over if BART was on my left or if I saw traffic slow down I would take over. So if this is supposed to score off FSD by far needs more work to hand over FSD BETA.

If you're in the FSD beta already, I'm surprised you're posting here.

 

[prashanthsha]

So far from my experience, full regen does not trigger hard braking at all. On autopilot or off.

Just stay completely off the brake pedal and you’ll be fine

I've driven on full regen & one-pedal driving and am being heavily dinged. I don't even touch the brake pedal. I feel that is a flaw in the algorithm .

I feel Full regen should be encouraged, not penalized.

 

[redmunds]

keep a finger on one of the scroll wheels. Scrolling will do the same as torque. I always get nagged despite having hand on wheel and applying some (I guess not enough) torque. I found scrolling dismisses it. I generally lower volume and then back up.

These are the Tips I come here for!

I haven't been able to use FSD for very long because I haven't gotten the "touch" right. Either not enough and FSD cancels me, or too much and I cancel FSD.

 

[levi.peters193]

You'll have to restrict your driving hours to off-Bart hours. If this doesn't work for you you'll be required to move to a NON-Bart state

unfortunately the BART lights are on 24 hours. If they aren't running the lights are RED. This is something that should be fixed by Tesla to not detect a red light in a completely different lane.

If you're in the FSD beta already, I'm surprised you're posting here.

Not on FSD Beta. I have FSD and since this week with the new Safety score on the app. these are the issues im having while on FSD. I requested FSD beta the day it came out.

 

[TravelFree]

"Unsafe following is only measured when your vehicle is traveling at least 50 mph and is incorporated into the Safety Score formula as a percentage."

I just did a test run this morning that was two and a half hours at less than 50mph, mostly in the country roads. No Interstates and many traffic lights. Roads I used had maximum 45mph. So after driving 75.4 miles I check my strikes for the drive and I had unsafe following 3.2%. Something isn't right. Also, there was hardly any other vehicles on the road and a few I did encounter ahead of me took off at much higher speed than me. The closest car was a 15 seconds at my speed away.

My only other strike has been hard braking using regen only. So when approaching a light I begin to slow much sooner and never allow regen to go full. That worked and resulted in 0 hard braking.

So my score as of today is 98 with 279 miles of travel. I did 101 of that at 75mph on AP.

 

[TravelFree]

 

[brianman]

Not on FSD Beta.

Then you shouldn't have describe your driving trips as "on FSD". Only beta users can say that.

 

[T3M3T]

I've driven on full regen & one-pedal driving and am being heavily dinged. I don't even touch the brake pedal. I feel that is a flaw in the algorithm .

I feel Full regen should be encouraged, not penalized.

I do the same...

 

[dgatwood]

It kind of sounds like the code trying to decide if you're a safe driver is at best the same pile of neural nets that you're supposed to be monitoring, and at worst, even worse at judging correctness. So doesn't that completely defeat the purpose of doing this analysis in the first place?

It seems like what they should really be doing is awarding FSD access to the people with the largest number of legitimate disengagements. And if you can't determine if the disengagements are legitimate (e.g. by running some reasonable portion of the FSD stack in parallel), then it probably isn't good enough to release yet.

For example, consider hard braking:

• To avoid a collision, stopping within a short distance of another vehicle or obstruction: driver error.
• To maximize regen before coasting slowly up to another vehicle or obstruction: desirable driver behavior.
• Because you didn't slow down for the yellow light that just turned red: driver error.
• Because Autopilot thought a flashing yellow light on the freeway was a traffic light: car error; bonus points for driver quickly stomping the gas pedal to avoid getting rear-ended.
• Because Autopilot in highway mode thought another vehicle cut into your lane: car error; bonus points for driver quickly stomping the gas pedal to avoid getting rear-ended.

And unfortunately, the legitimate causes of hard braking greatly exceed the driver error causes, both in frequency and in importance, so if they can't do better at weeding out the noise, they should drop that one entirely.

Then again, we MCU1 users are still swearing under our breath wondering if they'll ever even give us a button to push, so... whatevs.

 

[USCGDAV]

It kind of sounds like the code trying to decide if you're a safe driver is at best the same pile of neural nets that you're supposed to be monitoring, and at worst, even worse at judging correctness. So doesn't that completely defeat the purpose of doing this analysis in the first place?

It seems like what they should really be doing is awarding FSD access to the people with the largest number of legitimate disengagements. And if you can't determine if the disengagements are legitimate (e.g. by running some reasonable portion of the FSD stack in parallel), then it probably isn't good enough to release yet.

For example, consider hard braking:

• To avoid a collision, stopping within a short distance of another vehicle or obstruction: driver error.
• To maximize regen before coasting slowly up to another vehicle or obstruction: desirable driver behavior.
• Because you didn't slow down for the yellow light that just turned red: driver error.
• Because Autopilot thought a flashing yellow light on the freeway was a traffic light: car error; bonus points for driver quickly stomping the gas pedal to avoid getting rear-ended.
• Because Autopilot in highway mode thought another vehicle cut into your lane: car error; bonus points for driver quickly stomping the gas pedal to avoid getting rear-ended.

And unfortunately, the legitimate causes of hard braking greatly exceed the driver error causes, both in frequency and in importance, so if they can't do better at weeding out the noise, they should drop that one entirely.

Then again, we MCU1 users are still swearing under our breath wondering if they'll ever even give us a button to push, so... whatevs.

There should be no contest, game or test to award something that people have paid for in the first place. This may be the only Ponzi Scheme ever created that sells a buyer on spending 10K on a product (years ago) without ever delivering a finished working product that would be rolled out to those who OWN it in their vehicles. You spend 10K and that money goes into creating/building plants, hiring employees, manufacturing more vehicles (which only have a 3.0 computer without FSD) and the cycle goes on and on and on. So, your 10K buys nothing for years but the car company keeps on growing and its stock is how high? If you could make one of these out stamps you could call it Ponzi Motors. Now, let's talk "testers". 2,000 people (we don't know who they are) have been driving around with FSD for how long? Just how much testing (how many years) is required to "test" a car's feature for full release? Does, Ford, GM, etc roll out software updates to your car on a monthly basis? The only thing they can do is hook your car and run diagnostic and turn off the faults. This should tell us all something. There are so many lawsuits against Tesla for FSD and the promises made but not kept that will keep it rolling in and out of courtrooms for years to come. I've come to the conclusion to not download any updates as who knows what's in those updates that just might shut off parts of what you have and activate others you may not want. For those who paid 10K they have a legitimate beef and child play games should come to an end. All that said, I love my car the way it is.

 

[T3M3T]

There should be no contest, game or test to award something that people have paid for in the first place. This may be the only Ponzi Scheme ever created that sells a buyer on spending 10K on a product (years ago) without ever delivering a finished working product that would be rolled out to those who OWN it in their vehicles. You spend 10K and that money goes into creating/building plants, hiring employees, manufacturing more vehicles (which only have a 3.0 computer without FSD) and the cycle goes on and on and on. So, your 10K buys nothing for years but the car company keeps on growing and its stock is how high? If you could make one of these out stamps you could call it Ponzi Motors. Now, let's talk "testers". 2,000 people (we don't know who they are) have been driving around with FSD for how long? Just how much testing (how many years) is required to "test" a car's feature for full release? Does, Ford, GM, etc roll out software updates to your car on a monthly basis? The only thing they can do is hook your car and run diagnostic and turn off the faults. This should tell us all something. There are so many lawsuits against Tesla for FSD and the promises made but not kept that will keep it rolling in and out of courtrooms for years to come. I've come to the conclusion to not download any updates as who knows what's in those updates that just might shut off parts of what you have and activate others you may not want. For those who paid 10K they have a legitimate beef and child play games should come to an end. All that said, I love my car the way it is.

Elon should give us stocks to compensate. He should pay us interest for these years!

 

[DaveG_NJ]

I've driven on full regen & one-pedal driving and am being heavily dinged. I don't even touch the brake pedal. I feel that is a flaw in the algorithm .

I feel Full regen should be encouraged, not penalized.

In warm weather and a max 80% charge, I use one-pedal driving exclusively. Yes, you will get dinged for "hard-braking" if you take your foot off the accelerator pedal too quickly. "Hard braking" is a misnomer - rapid deceleration is a better term since it's measuring negative G-forces. And why that matters is debatable. I wish they would have been more precise in their description before I took a 100 mile drive, drove like a grandma, and only scored an 86 because of regen.

 

[levi.peters193]

Then you shouldn't have describe your driving trips as "on FSD". Only beta users can say that.

I drive around mainly on FSD (not beta) im 90% freeway driver. my score is mostly based on driving FSD. So if the FSD freeway driving is hard braking or following too close even with setting distance to 7 I still get peeped at. The score should be against the FSD not myself as a driver. thats where the score system is flawed.

 

[EVNow]

I drive around mainly on FSD (not beta) im 90% freeway driver. my score is mostly based on driving FSD. So if the FSD freeway driving is hard braking or following too close even with setting distance to 7 I still get peeped at. The score should be against the FSD not myself as a driver. thats where the score system is flawed.

We don't use "FSD" terminology for that - we use AP or NOA. That is how Tesla describes them too.

 

[T3M3T]

We don't use "FSD" terminology for that - we use AP or NOA. That is how Tesla describes them too.

I agree with you... the score should be based on AP and how much people pay attention when using it!

 

[chamshaw]

First road I ever drove I New Zealand, and first time driving on the left, was Highcliff Road on the Otago Peninsula. It seemed like the narrowest road I'd ever been on. I took Portobello Road back that night - just as narrow and winding, but with the ocean literally washing over the road in a storm. I ended up living on the North Island in Reikorangi just off the infamous Akatawara Road. I've never driven in Britain, but I can't imagine it could be narrower.
Never mind me. I'm just sitting in Texas, nostalgic as Hell now, missing NZ more every day. Enjoy a Speight's Old Dark for me.

G'day kavyboy; there are a few narrower roads to terrify you in good old Aotearoa if you ever come back. In Britain you'll have brambles and stone walls hemming you in. Eight years in Texas (Kenedy, yes, one N, and Runge) with my new girl that I met in London and moved to US with before seven in UK before returning home here. Domestically we're going full electric and the Tesla is the first move, wall charger done, solar next. Ain't life exciting?