正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

Reasoning is what really changed the game with FSD V14, and as they implement more of it, it’s going to be game over. Here’s a great example. FSD exits the parking lot of this coffee shop. If you zoom into the nav at the top right corner, you can... show more

Mike P

19,255 subscribers

46,221 次观看 • 5 个月前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

FSD v14.3.7 doesn’t proceed on an advanced green This is the first time ever (?) that I have had FSD sit at a green light at not proceed. What’s also interesting is you can see the turn signal turn on for a second and then turn back off - but in the car the nav was 100% set to turn left. Any ideas? Has this ever happened to any of you.

FSD v14.3.7 doesn’t proceed on an advanced green This is the first time ever (?) that I have had FSD sit at a green light at not proceed. What’s also interesting is you can see the turn signal turn on for a second and then turn back off - but in the car the nav was 100% set to turn left. Any ideas? Has this ever happened to any of you.

Devin Olsen

27,706 次观看 • 13 天前

I’m really looking forward to these upcoming Tesla FSD updates: • FSD V14 “Lite” for HW3 owners: Can’t wait to try it, even as a daily HW4 user. It’s expected to release at the end of this month and include the same features as HW4 V14. • Pothole Avoidance: This will be very welcome, especially for Teslas with lower-profile tires. It will be a game changer for me. • FSD Remembering Your Parking Preferences: Elon Musk says future FSD releases will remember your parking preferences when arriving at a destination. Great for garages, work spots, etc. • Grok Voice Commands for FSD: Elon Musk says the ability to prompt FSD with language is coming within ~3 months. You’ll be able to use your voice to tell the car what to do or where to go. This will be incredibly cool. CANT WAIT.

Nic Cruz Patane

60,550 次观看 • 2 个月前

Sprinkle of Cybertruck FSD mishaps. FSD v14.2.1. FSD was getting ready to make a left turn. In fairness, I pushed the pedal as there still was time to make the turn before the oncoming car. However, I took the foot off and the truck came to a stop. Then when the turn was almost complete, FSD threw the red hands and the message said "Aborted with a system error." If you are wondering why I looked at the camera in between, I wanted to make sure it was recording the situation. I have hours of video footage I recorded while driving on FSD 14.2 and 14.2.1.

Sprinkle of Cybertruck FSD mishaps. FSD v14.2.1. FSD was getting ready to make a left turn. In fairness, I pushed the pedal as there still was time to make the turn before the oncoming car. However, I took the foot off and the truck came to a stop. Then when the turn was almost complete, FSD threw the red hands and the message said "Aborted with a system error." If you are wondering why I looked at the camera in between, I wanted to make sure it was recording the situation. I have hours of video footage I recorded while driving on FSD 14.2 and 14.2.1.

Andy ze German

14,174 次观看 • 8 个月前

Rapidfire epic moments on FSD V14.3.5 1) FSD turns right into a small alley that only fits one car at a time, sees oncoming car, reverses out of alley to make space, realizes oncoming car is actually parking, re-enters alley. 2) Insane speed to vehicle cues. As FSD approaches the intersection at the end of the small alley, it yields for traffic, but at the SECOND the car to our left slows, it realizes they’re letting us go and executes with zero hesitation. You can see the nose of the vehicle on the far left side of the shot here to gauge reaction speed. 3) FSD approaches railroad traffic, clears the vehicle intersection, but stops shy of first RR track instead of queueing behind lead car, ignoring the human who decided to foolishly stop in front of the track. This is why I’m bullish. This tech is just a constant mindblower when you’re out here using it.

Rapidfire epic moments on FSD V14.3.5 1) FSD turns right into a small alley that only fits one car at a time, sees oncoming car, reverses out of alley to make space, realizes oncoming car is actually parking, re-enters alley. 2) Insane speed to vehicle cues. As FSD approaches the intersection at the end of the small alley, it yields for traffic, but at the SECOND the car to our left slows, it realizes they’re letting us go and executes with zero hesitation. You can see the nose of the vehicle on the far left side of the shot here to gauge reaction speed. 3) FSD approaches railroad traffic, clears the vehicle intersection, but stops shy of first RR track instead of queueing behind lead car, ignoring the human who decided to foolishly stop in front of the track. This is why I’m bullish. This tech is just a constant mindblower when you’re out here using it.

Mike P

159,909 次观看 • 1 个月前

This is a tough UPL turn for fsd. Cross traffic is going 60mph & the sign blocks the view from the left but from the driver's seat, I could see the car coming at the 0:48 sec mark. Fsd tried to proceed & I had to disengage

This is a tough UPL turn for fsd. Cross traffic is going 60mph & the sign blocks the view from the left but from the driver's seat, I could see the car coming at the 0:48 sec mark. Fsd tried to proceed & I had to disengage

Elias Martinez

20,743 次观看 • 4 个月前

This major issue with the FSD 14.3 branch is still present in 14.3.4. FSD has completely forgotten how to handle this green left turn arrow FSD fights back if you try to hold the accelerator and in 2 of the 3 attempts this causes the wheel to fly around so much I have to grab it This is a major issue that generally results in a critical disengagement (either getting completely stuck at the light or becoming stranded in the intersection if the car goes the wrong way - either way confusing human traffic). Rarely the car does somewhat respond to the accelerator and can be coaxed through. FSD has never had this level of difficulty going through this intersection from this direction through 14.2.2.5

This major issue with the FSD 14.3 branch is still present in 14.3.4. FSD has completely forgotten how to handle this green left turn arrow FSD fights back if you try to hold the accelerator and in 2 of the 3 attempts this causes the wheel to fly around so much I have to grab it This is a major issue that generally results in a critical disengagement (either getting completely stuck at the light or becoming stranded in the intersection if the car goes the wrong way - either way confusing human traffic). Rarely the car does somewhat respond to the accelerator and can be coaxed through. FSD has never had this level of difficulty going through this intersection from this direction through 14.2.2.5

RBOTAXI DRIVER

13,941 次观看 • 2 个月前

Forget UPL, it's the Devin's simple right hand turn that FSD needs to master first. I am blown away that software as good as FSD can make mistakes as simple but devastating as this. This is the second time in less than two weeks that FSD has tried to pull infront of a vehicle on a VERY SIMPLE right hand turn. The car drove pretty much flawlessly all day today (3~ hours of driving) but it ends the day with a move like this and it just takes all the confidence I have with FSD away in a simple move. I understand that the version of FSD we have today is not what will be used in Austin in June, but the fact that FSD is making decisions like this and 'full autonomy' is less than 3 months away - it's a bit concerning to me. I would feel a bit better if a near flawless version of FSD existed 'in the wild' for several months before Austin happens, but that's just me. Tesla Model 3 HW4 FSD 13.2.8 Tesla AI Ashok Elluswamy

Forget UPL, it's the Devin's simple right hand turn that FSD needs to master first. I am blown away that software as good as FSD can make mistakes as simple but devastating as this. This is the second time in less than two weeks that FSD has tried to pull infront of a vehicle on a VERY SIMPLE right hand turn. The car drove pretty much flawlessly all day today (3~ hours of driving) but it ends the day with a move like this and it just takes all the confidence I have with FSD away in a simple move. I understand that the version of FSD we have today is not what will be used in Austin in June, but the fact that FSD is making decisions like this and 'full autonomy' is less than 3 months away - it's a bit concerning to me. I would feel a bit better if a near flawless version of FSD existed 'in the wild' for several months before Austin happens, but that's just me. Tesla Model 3 HW4 FSD 13.2.8 Tesla AI Ashok Elluswamy

Devin Olsen

137,423 次观看 • 1 年前

FSD's performance has once again surpassed humans. While waiting at an intersection for the green light, FSD suddenly moved forward. The driver thought FSD was about to run a red light and prepared to disengage. However, FSD only moved forward slightly and then stopped. it didn’t run the red light. It turned out that a car behind was trying to squeeze past from FSD’s left rear side. FSD judged that the space was too narrow and unsafe, as it might get hit, so it moved to avoid the situation. This is truly incredible. Source: FSD国产小萌萌 at DY

FSD's performance has once again surpassed humans. While waiting at an intersection for the green light, FSD suddenly moved forward. The driver thought FSD was about to run a red light and prepared to disengage. However, FSD only moved forward slightly and then stopped. it didn’t run the red light. It turned out that a car behind was trying to squeeze past from FSD’s left rear side. FSD judged that the space was too narrow and unsafe, as it might get hit, so it moved to avoid the situation. This is truly incredible. Source: FSD国产小萌萌 at DY

𝐶ℎ𝑎𝑟𝑙𝑒𝑠

52,678 次观看 • 1 年前

One of the greatest luxuries of FSD V14 is the fact that it handles this type of driving far better and more effortlessly than you. This is when constant 360 vision puts on a clinic. As a human, you have to scan your mirrors over and over, turn your head to check blinds, and even still, these little slides can be straining and annoying. FSD is so so good at flowing through dense traffic that you legitimately don’t even flinch. It’s as low stress as driving through an empty parking lot. Now imagine this in a foreign town you’ve never been to before? FSD will just cook and you won’t have a care in the world. Just sit there chillin, glancing at the screen and mirrors feeling like you really shouldn’t even have to do that anymore. Could never give this tech up after getting used to it.

One of the greatest luxuries of FSD V14 is the fact that it handles this type of driving far better and more effortlessly than you. This is when constant 360 vision puts on a clinic. As a human, you have to scan your mirrors over and over, turn your head to check blinds, and even still, these little slides can be straining and annoying. FSD is so so good at flowing through dense traffic that you legitimately don’t even flinch. It’s as low stress as driving through an empty parking lot. Now imagine this in a foreign town you’ve never been to before? FSD will just cook and you won’t have a care in the world. Just sit there chillin, glancing at the screen and mirrors feeling like you really shouldn’t even have to do that anymore. Could never give this tech up after getting used to it.

Mike P

12,031 次观看 • 2 个月前

When you mount a camera to your car and run it every time you FSD in your Tesla, you’re bound to catch those little moments of pure magic that give you chills. This might be my favorite I’ve ever captured. Here’s a very awkward edge case. I am approaching a railroad crossing from an adjacent road. At the very moment I approach the stop sign, the lights begin to flash, and the bell begins to ring. FSD V14.2 does its NHTSA stop, and is at the point when it would normally proceed (BEFORE the gate arms begin to fall). FSD is able to gather context from the flashing red lights and the sound of the bell. From an adjacent road. Without any gate arm movement as additional cue. I really wasn’t sure here, and thought I might need to take over because of our awkward positioning, and because the arms weren’t going to drop before FSD would typically go. But nope. No problem for the sentient AI. As a cherry on top… when the gate lifts, the silver car waves me on and FSD begins to engage the right turn, but it stops before executing the turn…. Wtf?… oh lol. Forgot my robot car has eyes on the side of its head. Unbelievable technology. This company is at the very beginning of something incredible. Sustainable abundance or die trying. ❤️

When you mount a camera to your car and run it every time you FSD in your Tesla, you’re bound to catch those little moments of pure magic that give you chills. This might be my favorite I’ve ever captured. Here’s a very awkward edge case. I am approaching a railroad crossing from an adjacent road. At the very moment I approach the stop sign, the lights begin to flash, and the bell begins to ring. FSD V14.2 does its NHTSA stop, and is at the point when it would normally proceed (BEFORE the gate arms begin to fall). FSD is able to gather context from the flashing red lights and the sound of the bell. From an adjacent road. Without any gate arm movement as additional cue. I really wasn’t sure here, and thought I might need to take over because of our awkward positioning, and because the arms weren’t going to drop before FSD would typically go. But nope. No problem for the sentient AI. As a cherry on top… when the gate lifts, the silver car waves me on and FSD begins to engage the right turn, but it stops before executing the turn…. Wtf?… oh lol. Forgot my robot car has eyes on the side of its head. Unbelievable technology. This company is at the very beginning of something incredible. Sustainable abundance or die trying. ❤️

Mike P

92,367 次观看 • 5 个月前

FSD slams on the brakes for a cat in the road that doesn’t even exist. This is FSD 12.6.4 on HW3, which is a lot more sensitive to shadows & objects in the road with the 2026.8.6 update. There are many reports of brake stabbing. It brings me back to the days when I was testing with v10.3 of FSD back in 2021. It’s interesting that it’s the same FSD version as before, and not something like 12.6.4.1, but I guess Tesla can change the weights without changing the version. And they’ve likely been doing this for a while with 12.6.4. 😒

FSD slams on the brakes for a cat in the road that doesn’t even exist. This is FSD 12.6.4 on HW3, which is a lot more sensitive to shadows & objects in the road with the 2026.8.6 update. There are many reports of brake stabbing. It brings me back to the days when I was testing with v10.3 of FSD back in 2021. It’s interesting that it’s the same FSD version as before, and not something like 12.6.4.1, but I guess Tesla can change the weights without changing the version. And they’ve likely been doing this for a while with 12.6.4. 😒

TechGeek Tesla 🔋⚡️

29,115 次观看 • 4 个月前

This traffic circle has been a long-standing pain point for me on FSD. As you can see, the car consistently drives in the bike lane here. I’ve documented this behavior before on FSD 12, 13, and now 14. The issue appears to stem from outdated map data. This section of the circle used to function as a two-way roadway. At the time, there was no bike lane and vehicles traveled in the space FSD still prefers today. Since the redesign, that area is now a dedicated bike lane, but FSD doesn’t reliably recognize the change. Depending on the version, it will occasionally handle this correctly, but most of the time it does not. This is currently the only issue I consistently experience with FSD. It’s not safety critical. The car will move if a cyclist, vehicle, or object is present, but reliance on outdated map data in locations like this needs to improve.

This traffic circle has been a long-standing pain point for me on FSD. As you can see, the car consistently drives in the bike lane here. I’ve documented this behavior before on FSD 12, 13, and now 14. The issue appears to stem from outdated map data. This section of the circle used to function as a two-way roadway. At the time, there was no bike lane and vehicles traveled in the space FSD still prefers today. Since the redesign, that area is now a dedicated bike lane, but FSD doesn’t reliably recognize the change. Depending on the version, it will occasionally handle this correctly, but most of the time it does not. This is currently the only issue I consistently experience with FSD. It’s not safety critical. The car will move if a cyclist, vehicle, or object is present, but reliance on outdated map data in locations like this needs to improve.

Arthur | TesCamStudio.com

46,580 次观看 • 7 个月前

I had one of those moments today that really showed me how much Tesla’s FSD has improved. We were heading to our next destination, and I turned on FSD in my Cybertruck. Instead of taking the obvious route out of the parking lot, it made a turn and drove straight into a parking garage. At that point, my significant other was yelling at me to disengage bc she didn’t want to waste time. Her argument was simple… FSD will probably get confused in a situation like this and we may end up going on an infinite loop FSD ride here. A parking garage is a pretty challenging environment if you think about it… multiple levels, ramps, tight turns, parked cars, limited visibility, and signs that can be easy to miss. But I told her I wanted to see what it would do. My thinking was that the latest version of FSD is much better at understanding their surroundings, reading signs, recognizing where they are, and making decisions based on what they see in real time. Sure enough, after realizing it wasn’t on the correct path, my Cybertruck kept moving, analyzed the situation, followed the garage layout, went all the way to the top floor, found the exit, and drove itself back out. Then it continued toward our destination without me touching the wheel. What impressed me wasn’t that it made a wrong turn… tbh human drivers do that all the time. What impressed me was that it recognized the situation, adapted, and successfully recovered on its own. Years ago, that probably would have been the end of the experiment bc it would go on an infinite loop, eventually resulting in a disengagement. Today, with FSD v14.3.4, it felt like watching the AI handle an unexpected problem and work its way through it. For me, that’s one of the biggest signs of progress in Tesla FSD, where it’s not 100% perfect, yet being able to make sense of a mistake, figuring out a solution, and moving forward.

I had one of those moments today that really showed me how much Tesla’s FSD has improved. We were heading to our next destination, and I turned on FSD in my Cybertruck. Instead of taking the obvious route out of the parking lot, it made a turn and drove straight into a parking garage. At that point, my significant other was yelling at me to disengage bc she didn’t want to waste time. Her argument was simple… FSD will probably get confused in a situation like this and we may end up going on an infinite loop FSD ride here. A parking garage is a pretty challenging environment if you think about it… multiple levels, ramps, tight turns, parked cars, limited visibility, and signs that can be easy to miss. But I told her I wanted to see what it would do. My thinking was that the latest version of FSD is much better at understanding their surroundings, reading signs, recognizing where they are, and making decisions based on what they see in real time. Sure enough, after realizing it wasn’t on the correct path, my Cybertruck kept moving, analyzed the situation, followed the garage layout, went all the way to the top floor, found the exit, and drove itself back out. Then it continued toward our destination without me touching the wheel. What impressed me wasn’t that it made a wrong turn… tbh human drivers do that all the time. What impressed me was that it recognized the situation, adapted, and successfully recovered on its own. Years ago, that probably would have been the end of the experiment bc it would go on an infinite loop, eventually resulting in a disengagement. Today, with FSD v14.3.4, it felt like watching the AI handle an unexpected problem and work its way through it. For me, that’s one of the biggest signs of progress in Tesla FSD, where it’s not 100% perfect, yet being able to make sense of a mistake, figuring out a solution, and moving forward.

Teslaconomics

15,826 次观看 • 2 个月前

FSD V14.2 Ambulance Fail As always I think it's healthy to show the good and the bad with FSD. Last night when I was out testing 14.2 I ran into an issue where an ambulance was coming from my right and FSD failed to notice it in time. I ended up having to disengage which marked my first and only (non-nav related) disengagement on V14.2. Let me know what you think in the comments.

FSD V14.2 Ambulance Fail As always I think it's healthy to show the good and the bad with FSD. Last night when I was out testing 14.2 I ran into an issue where an ambulance was coming from my right and FSD failed to notice it in time. I ended up having to disengage which marked my first and only (non-nav related) disengagement on V14.2. Let me know what you think in the comments.

Devin Olsen

138,762 次观看 • 9 个月前

Chinese NIO influencer Dong Maimai recently flew from China to California to try FSD 14.1. He was blown away by its capabilities. Here is a scenario he shared where FSD waited behind a car by the garage ticket dispenser and then FSD decided to back out and switched to the empty lane on the right to pick up a parking ticket before the access control arm was lifted up. This example shows the strong situational awareness of FSD and quick execution. No other systems can do this other than FSD.

Chinese NIO influencer Dong Maimai recently flew from China to California to try FSD 14.1. He was blown away by its capabilities. Here is a scenario he shared where FSD waited behind a car by the garage ticket dispenser and then FSD decided to back out and switched to the empty lane on the right to pick up a parking ticket before the access control arm was lifted up. This example shows the strong situational awareness of FSD and quick execution. No other systems can do this other than FSD.

Ray

86,120 次观看 • 8 个月前

Found a FSD V14.2 limit the other night in some of the worst fog I’ve ever driven through at night, while it was raining. In the early part of the video, you can see me take over. I did this because the car was going about 50 mph, and I completely lost my bearings. It’s funny, because cameras def see better in fog than humans. Even this video doesn’t look as bad as it looked to me IRL. But I was totally uncomfortable, and took over because I really had no clue where the road was at the moment I disengaged. FSD didn’t quit here and might’ve been fine, but I’d prefer to see it go slower in these moments. I wasn’t confident that it understood how bad the fog was, and I wasn’t familiar with the area. When I attempt to re-engage, FSD quits immediately. I think this was because of rain (you can see cam visibility pop ups). People will say this is a vision problem, but it really just comes down to needing better context and adjusting speed in these conditions. Conditions this severe aren’t super common, and the brain probably doesn’t have a ton of exp here yet. Idk if I’ll ever find myself in a condition this bad again but if I do I hope to see FSD signal greater awareness.

Found a FSD V14.2 limit the other night in some of the worst fog I’ve ever driven through at night, while it was raining. In the early part of the video, you can see me take over. I did this because the car was going about 50 mph, and I completely lost my bearings. It’s funny, because cameras def see better in fog than humans. Even this video doesn’t look as bad as it looked to me IRL. But I was totally uncomfortable, and took over because I really had no clue where the road was at the moment I disengaged. FSD didn’t quit here and might’ve been fine, but I’d prefer to see it go slower in these moments. I wasn’t confident that it understood how bad the fog was, and I wasn’t familiar with the area. When I attempt to re-engage, FSD quits immediately. I think this was because of rain (you can see cam visibility pop ups). People will say this is a vision problem, but it really just comes down to needing better context and adjusting speed in these conditions. Conditions this severe aren’t super common, and the brain probably doesn’t have a ton of exp here yet. Idk if I’ll ever find myself in a condition this bad again but if I do I hope to see FSD signal greater awareness.

Mike P

11,580 次观看 • 5 个月前

This is the coolest thing to happen to me so far with FSD. I asked Grok to not take a left on S Mcdill and instead go straight take take a left on Bayshore Blvd, and do my surprise…IT DID IT! I wake it last minute too; which you can see FSD/Grok change directions.

This is the coolest thing to happen to me so far with FSD. I asked Grok to not take a left on S Mcdill and instead go straight take take a left on Bayshore Blvd, and do my surprise…IT DID IT! I wake it last minute too; which you can see FSD/Grok change directions.

Vad3r

61,298 次观看 • 1 个月前

I chose to label this Navigation error as “Critical” while testing FSD v14.3.2 Here’s why: ✅ This intervention wasn’t “preference,” as the maneuver FSD routed was illegal ✅ If a police officer saw this maneuver, it would result in a ticket

I chose to label this Navigation error as “Critical” while testing FSD v14.3.2 Here’s why: ✅ This intervention wasn’t “preference,” as the maneuver FSD routed was illegal ✅ If a police officer saw this maneuver, it would result in a ticket

TESLARATI

43,968 次观看 • 4 个月前

FSD wants to make a left turn at this light but pauses twice for oncoming traffic. I applaud this. That truck was blocking a lot of the view and the system was fast to pause. I like the fast response to “danger”.

FSD wants to make a left turn at this light but pauses twice for oncoming traffic. I applaud this. That truck was blocking a lot of the view and the system was fast to pause. I like the fast response to “danger”.

Martha 🌑🚀

35,139 次观看 • 7 个月前