Caution: Do not try this FSD test at home, kids! Testing FSD’s response to an elevated obstacle: a 2x4 sticking out directly into its path. At 30 mph, FSD does not react ❌ At 20 mph, FSD swerves to avoid it ✅ Shoutout to Muzeïshen for trusting his life... in my hands and AJ for being a witness that no harm was done This test was a bit spontaneous and will be improved in the future. I know that FSD still has trouble with suspended obstacles. But to be fair the wood was not clearly visible to me until about a couple of seconds before I intervenedshow more

Balls of steel man pov

@OGMarsLife @Muzeishen wow great test!

@OGMarsLife @Muzeishen Some people would say it’s stupid lol Thx for the vote of confidence!

@OGMarsLife @Muzeishen we know this still gives trouble to FSD. As long as the testers were vigilant and expecting the worst then this is fine. I’ve been using FSD since v10.3 and NoA before that so I’ve seen a million times worse things than this personally

@edgecase411 @OGMarsLife @Muzeishen At 30MPH it was turning to avoid and you disengaged after it started to turn

@edgecase411 @OGMarsLife @Muzeishen I wasn’t the one who did the test man. Idk

@Muzeishen Real world, not a test - but at low speeds it did well.

@Muzeishen That’s exactly the test I wanted to do but didn’t have the space to do it from the back of the trunk on this street. Thx for sharing! Will plan for it better next time for sure

@Muzeishen Not a smart test. If you hit that 2x4 it could have done severe damage and potentially kill the person holding it

@Muzeishen He volunteered! But yes I wouldn’t do that again

@Muzeishen If someone asked you if they should jump off a roof… If you need to test it again, just ensure nobody is in the car or closed to it. You can always secure a 2x4 with some bungee cords

@Muzeishen Like I said the test was spontaneous. Fwiw if you knew what FSD10.9 was doing you’d be appalled that it was even legal to be used on the streets. Just food for thought.

@Muzeishen Honestly, FSD at 33mph has started to slow, ie chevrons in the blue path, and the steering has turned, before the manual takeover. I think the car would have cleared the 2x4.

@Muzeishen The steering wheel turning before the FSD state disengages is a byproduct of system latency. There is a time between manual torque applied to steering to when the system disengages and this is what you’re seeing.

@Muzeishen Where's @IfindRetards... this one's certifiable

@Muzeishen @IfindRetards Oh please. You must not have used FSD V10.9 🙄

@Muzeishen @IfindRetards Talking about trying to get a 2x4 wedged into the guy's neck. This can be tested 1k other ways not on the list of 1k ways to die

@Muzeishen That's a good way to kill your friend. Interesting test, but you really need to rethink your setup if your friend wants to keep his head attached

@Muzeishen If you read the caption you’d see that I said it wasn’t a well planned test

@Muzeishen Very interesting how it only responded at a slower speed. @Muzeishen you are bold man lol

@Muzeishen Perhaps it has more time to “see” it at slower speed

@Muzeishen No excuse for a robot lol

@Muzeishen 😭

@Muzeishen Very cool

@Muzeishen why not try with pool noodles first so, it doesn't damage the testing car.

@Muzeishen Good test before v15

@Muzeishen Nice test bro, but keep 2x4s away, wtf.

@Muzeishen just wondering if it hit the pole, is it gonna chop the guy's head off?

@Muzeishen Yes

@Muzeishen Love it! Thx for the sacrifices you guys made!

@Muzeishen 🫶

@Muzeishen That is why is countries with some brains they actually have penants hanging from such a bar, and they don’t even have FSD.

@Muzeishen If the system hadn’t worked, the plastic surgeon would have had some serious work to do....if he survived.

@Muzeishen @Muzeishen had already one lined up just in case

@Muzeishen Well thats illegal and 99 out of 100 human drivers would miss it

@Muzeishen What’s illegal? It’s ok. Just have it as a point of reference when FSD is able to avoid it one day since it will be super human 😏

@Muzeishen Having any sort of material sitcking out of your vehicle without proper flagging

@Muzeishen Same idea as Chuck’s thin chain test, though FSD still hasn’t (I think) ever successfully avoided it.

It sees my gate bar that lift up all the time no bigger than a 2x4. In fact it tracks it all the way to the top and always accidentally lurches to a an almost stop momentarily to make sure it doesn’t descend again before it crosses. Maybe a color scheme thing. Hard to see your 2x4. Most humans would drive right through that bar at 15 mph I think

@Muzeishen 흥미로운 비교지만 공공도로에서 장애물을 이용한 테스트는 위험합니다. 승인된 폐쇄 코스와 안전 요원이 있는 환경에서만 진행하고, 일반 사용자는 절대 따라 하지 않도록 강조해야겠네요.

@Muzeishen The 30→20 mph split is the real result: a ten-mph change turning “no reaction” into a swerve needs controlled repeats before it says much about FSD.

@Muzeishen 위험한 테스트는 절대 재현하지 말아야 하지만, suspended obstacle 같은 엣지 케이스를 공개하는 건 개선에 큰 도움이 되죠. 30→20mph 차이를 데이터로 남긴 점이 특히 중요해 보입니다.

@Muzeishen Not surprised

@Muzeishen Nice test!!

@Muzeishen Sentient

@Muzeishen V14 drove wider for me in suburban street once because the tree was overgrown over the road - rather wide low growing tree. Was at probably 20km at the time but I was like wow when it did it.

@Muzeishen I am not sure I would even have reacted to that one. Pretty uncommon scenario, even for my 40+ year old brain, lol.

@Muzeishen Strange with it's reaction time it should be able to swerve at 30 mph just fine.

@Muzeishen Lmao yeah because this situation happens once every 350 years....

@Muzeishen Does it really?

@Muzeishen Lmao yeah from behind... Not sticking out the fucking driver window lmao

@Muzeishen Who gives a shit, it’s a suspended object. Fsd either sees it or not.

@Muzeishen I mean I guess you need to worry about it more if you live near a home depot in the ghetto.... Thats the only edge case that's even remotely even in the realm of possibility.

@Muzeishen Oh so it doesn’t happen at the Home Depot in bel air

@Muzeishen Doubtful lol

@Muzeishen What version of FSD is this?

@Muzeishen still got work to do

@Muzeishen 실제 도로에서 속도별 반응을 비교한 점이 유익하네요. 장애물 유형과 시야·노면 조건까지 함께 기록하면 FSD 개선에 더 큰 도움이 될 것 같습니다.

@Muzeishen Now add a baby on the left for this test to be real and useful

@Muzeishen

@Muzeishen Obviously a doll. What will dumbo do?

@Muzeishen 위험한 시험을 공개할수록 한계도 함께 드러나야 한다는 점이 중요합니다. 20·30mph 결과 차이는 유용한 신호지만, 일반 도로에서 재현하지 말고 폐쇄 환경의 반복 데이터와 안전 기준으로 검증해야겠네요.