New FSD beta testers (June 2022) - First experiences

[DanCar]

I'll give it a D-. Took my kids to school 1.5 miles away. Got honked twice by the same driver. Another driver called me an idiot. FSD was stopping all over the place and I was trying to figure out why.

Way to work was better. Still slowing down / stopping unexpectedly, but no one seemed angry.

Overall I like it. Maybe that will wear off. Yes the tech is impressive even though it is a quarter baked. I had very low expectations.
Love the full screen visualization. Wish there was an option for 3/4 and 1/2 view. Only options are 1/4 and full view. Sometimes it is nice to see the navigation route.

I was on 2022.20 branch prior to receiving FSD. Score was 89.

 

[Proppilot]

I'll give it a D-. Took my kids to school 1.5 miles away. Got honked twice by the same driver. Another driver called me an idiot. FSD was stopping all over the place and I was trying to figure out why.

Way to work was better. Still slowing down / stopping unexpectedly, but no one seemed angry.

Overall I like it. Maybe that will wear off. Yes the tech is impressive even though it is a quarter baked. I had very low expectations.
Love the full screen visualization. Wish there was an option for 3/4 and 1/2 view. Only options are 1/4 and full view. Sometimes it is nice to see the navigation route.

I was on 2022.20 branch prior to receiving FSD. Score was 89.

So much more phantom braking with the latest update and HARD

 

[Z_Lynx]

I think we need another reaction icon. A sad face.

Really discouraged to hear that. Have not yet tried 69.2.2 but was not happy to find 69.2 had significantly more phantom braking than the previous version.

 

[Sharps97]

New beta tester as of this week, using 2022.20.17. Quick summary, grade is F minus.

Tried it briefly, but it so thoroughly tried to kill me and wreak havoc across multiple straight-forward, non edge case situations that it isn't something I'll be using again until the next major version release.

Issues:

• Left / Right turns are either too wide (go into a ditch) or to narrow (say hi to oncoming traffic)
• Random freezes mid turn (say hi to being rear ended)
• Maniac acceleration to traffic light, will it / won't it stop? Hard breaking insanity.
• Navigate attempt through one-way intersection the wrong way
• Not guaranteed to stay within the lines when / where it obviously should

General feelings of FSD otherwise (i.e., on high-way):

• Impressively poor understanding of when to change lanes
• Impressively indecisive follow through on executing lane changes at speed (very, very dangerous)
• 50/50 chance it can appropriately handle merging in traffic
• Unable to navigate high volume traffic situations when approaching backed up exit lanes

I think in general, FSD is probably still more an 'alpha' state where Tesla is just throwing things at the wall to see what sticks, and if it does stick, band-aiding over it repeatedly until it fails, and then we regress for a while and try again. If I had to guess, I'd say that Tesla's image recognition tech is probably not 'best in class', and that is probably more to do with the onboard hardware rather than anything else. ML isn't magic, just brute-force regression against an error minimization term, so garbage in, garbage out would apply.

I think the next major 'leap' in performance should come from significantly improved cameras (much higher resolution, much higher frame rate) along with the associated processing power to handle that. But, that's only for image recognition. Driving policy is garbage and I'm not confident that Software 1.0 is the correct approach to solve something that is effectively an NP hard problem.

It may be that human drivers are just 'meat-sacks' with eyeballs, but they are sentient. Until the driving policy is based on generalized sentient AI, I don't see speedy improvements ahead.

 

[LowlyOilBurner]

Well, at least it’s able to drive over 27mph (actually drives the speed limit, or whatever you set it for) on unmarked roads. That’s a plus. A few of my routes I take weekly for work showed vast improvement in lane choices, to the point they are intervention-less drives.

-The stopping mid turn is a new “feature” I experienced a few times yesterday. Fix this.

-The stopping on certain unmarked roads when an on coming vehicle is approaching. Fix this.

-The hypersensitivity to VRU’s (pedestrians & cyclists) in certain situations. This is where some of the phantom braking is tied to, I think. Fix this.

-“creep barrier”. For left turns from a stop sign/light, they still position the car too far to the left, where if someone was turning on the road I’m making a left out of, you might be in their way. When I drive, I approach the intersection perpendicular to the road I’m entering, creep out, then go. I don’t creep out at an angle like FSD does. It’s just not normal feeling. Fix this.

-they still can’t keep the car centered in the lane on certain curves, crosses the yellow. Proper lane keeping should be “intro to FSD 101”. Lane keeping worked great way back on 2020.48.5 (before FSDb). But people bitched to was too slow for curves, well, guess what? The car stayed in its own lane. Fix this.

Otherwise, it’s a pretty decent improvement over 10.11, which was the last time I had FSDb, despite the shortcomings. Like all FSD releases, learn the shortcomings, when you can and cannot use it, and enjoy it for what it is.

 

[DanCar]

I'm thinking of calling it: FSS = full self stop

 

[DanCar]

I'm starting to like it more. I know where it fails and I know where FSD tends to succeed, so it is pleasant in the good cases and I don't use it in the bad cases. On the bad side, I notice I'm not paying attention as much as I should, so perhaps FSD is not for me. I'm going to need the electrical shock enhancement for those that don't pay attention.

 

[drtimhill]

I'm starting to like it more. I know where it fails and I know where FSD tends to succeed, so it is pleasant in the good cases and I don't use it in the bad cases. On the bad side, I notice I'm not paying attention as much as I should, so perhaps FSD is not for me. I'm going to need the electrical shock enhancement for those that don't pay attention.

Way back with the much earlier betas it wasn't an issue .. the car was crazy enough that you were always paying attention! But yes, it is now getting so good its rather too easy now to relax your vigilance, which was the very thing that scared Waymo away from the L2/L3 space. Perhaps Tesla will need to switch from making scary noises to jolting drivers with a quick zap from the battery

 

[Silicon Desert]

I'm interested in knowing from the people as to where they are seeing unexpected braking occur when you say it has returned and in some cases hard braking. I simply have not experienced one single issue of that occurring in 69.2.2 or 69.2.3 Scratching my head as to why people are reporting it.

 

[drtimhill]

I'm interested in knowing from the people as to where they are seeing unexpected braking occur when you say it has returned and in some cases hard braking. I simply have not experienced one single issue of that occurring in 69.2.2 or 69.2.3 Scratching my head as to why people are reporting it.

This is indeed "the big mystery" .. many people here report repeated serious braking while others see almost none at all. I suspect (with no proof at all), that it is probably a combination of several things:

1. Follow distance setting too low. This has been suggested by a number of people, but I'm not aware of any testing to confirm this. Also, it doesn't explain PB events when there is no car ahead of you.
2. Variations in sensors. Again, no real information on if this can be the cause, and while possible it seems less likely given the large PB variance amongst owners compared to the manufacturing tolerances of sensors.
3. Different driving conditions. This includes weather, time of day, road conditions etc. This is certainly plausible, and is likely to be a factor imho.
4. Different road usage. Some drivers seem to be reporting PB events when using stuff like TACC on roads that it wasn't really intended for. This one is difficult to judge, as people get very defensive and/or annoyed about stuff like this when discussed here.
5. Different driving styles. We all have different styles, from timid to "get out of my way!". It's possible some habits of some drivers are more likely to trigger PB events.
6. False reports. Clearly, humans are not perfect, and some percentage of PB events could be REAL events where the car was responding to a real issue that the human did not notice.
7. Different definitions. What exactly IS a PB event anyway? Sure, a butt-clenching 60mph to 10mph sudden brake is PB (assuming it wasn't a false report), but I've seen people claiming 65mph to 61mph is PB. Where do you draw the line?
8. Driver sensitivity. It's quite possible that once you have seen a few events you learn to distrust the car and ANY slight braking is mentally escalated to a PB event (see #7).
9. Conformation bias. Once someone is convinced PB is evil and pervasive, there is a tendency to notice the PB events but not the absence of PB events (that is, a drive with no PB events is ignored). This tends to magnify the perceived frequency of events in the mind of the driver.

There are probably others, but these seem to be the main ones that have been posited in this and other threads.

 

[Silicon Desert]

This is indeed "the big mystery" .. many people here report repeated serious braking while others see almost none at all. I suspect (with no proof at all), that it is probably a combination of several things:

1. Follow distance setting too low. This has been suggested by a number of people, but I'm not aware of any testing to confirm this. Also, it doesn't explain PB events when there is no car ahead of you.
2. Variations in sensors. Again, no real information on if this can be the cause, and while possible it seems less likely given the large PB variance amongst owners compared to the manufacturing tolerances of sensors.
3. Different driving conditions. This includes weather, time of day, road conditions etc. This is certainly plausible, and is likely to be a factor imho.
4. Different road usage. Some drivers seem to be reporting PB events when using stuff like TACC on roads that it wasn't really intended for. This one is difficult to judge, as people get very defensive and/or annoyed about stuff like this when discussed here.
5. Different driving styles. We all have different styles, from timid to "get out of my way!". It's possible some habits of some drivers are more likely to trigger PB events.
6. False reports. Clearly, humans are not perfect, and some percentage of PB events could be REAL events where the car was responding to a real issue that the human did not notice.
7. Different definitions. What exactly IS a PB event anyway? Sure, a butt-clenching 60mph to 10mph sudden brake is PB (assuming it wasn't a false report), but I've seen people claiming 65mph to 61mph is PB. Where do you draw the line?
8. Driver sensitivity. It's quite possible that once you have seen a few events you learn to distrust the car and ANY slight braking is mentally escalated to a PB event (see #7).
9. Conformation bias. Once someone is convinced PB is evil and pervasive, there is a tendency to notice the PB events but not the absence of PB events (that is, a drive with no PB events is ignored). This tends to magnify the perceived frequency of events in the mind of the driver.

There are probably others, but these seem to be the main ones that have been posited in this and other threads.

nice post, and some potential combination of reasons. I hope more folks who experience this issue will explain the circumstances where braking occurred... not that it is going to explain the differences, but might give some common clues. Folks out at the GF seem to agree with several of your reasons bases on snapshots.

 

[Proppilot]

I experience it in climbing and descending turns on a two lane highway with oncoming traffic. Not sharp turns, just lazy sweeps in the road. It is very rare that it happens (but it does) when I cannot reconcile what may have triggered it.

Really good posting above however (and this does sound pedantic) point 5 is mutually exclusive. You cannot be in control of the car and it PB’s as it is on AP so it cannot be related to driver style unless you mean other drivers around us ?

 

[drtimhill]

I experience it in climbing and descending turns on a two lane highway with oncoming traffic. Not sharp turns, just lazy sweeps in the road. It is very rare that it happens (but it does) when I cannot reconcile what may have triggered it.

Really good posting above however (and this does sound pedantic) point 5 is mutually exclusive. You cannot be in control of the car and it PB’s as it is on AP so it cannot be related to driver style unless you mean other drivers around us ?

Some people are lumping bogus AEB type events when manually driving the car into the PB bucket, hence #5.

 

[Shopaholic]

When the speed limits change on the road, very slow to react - speed changed from 65 to 45 and 35 over he cause of say 1/4 mile and it does recognize (as it displays the left panel) but the actual speed does not change - have to break or take it off autopilot to slow down. This behavior is worse than before! Overall impression - "D"

 

[Proppilot]

Some people are lumping bogus AEB type events when manually driving the car into the PB bucket, hence #5.

Roger that.

 

[drtimhill]

When the speed limits change on the road, very slow to react - speed changed from 65 to 45 and 35 over he cause of say 1/4 mile and it does recognize (as it displays the left panel) but the actual speed does not change - have to break or take it off autopilot to slow down. This behavior is worse than before! Overall impression - "D"

yeah its been noted multiple times that the car is VERY sluggish to slow down (contrasted to the way it speeds up). Ping the beta team on this .. the more of us who point this out, the faster it will get escalated.