← SnapRecaps

My Tesla Backed Into My Garage on FSD… Didn’t Expect This

► 2,532 views ⏲ 19:04 Watch on YouTube ↗

Summary

Tesla's HW4 FSD vastly outperforms HW3 in tricky maneuvers, leaving older owners disadvantaged, while the video clarifies Robotaxi limits and questions FSD's true safety metrics.

Executive Summary

This video compares Tesla’s Full Self-Driving on Hardware 3 versus Hardware 4, showing a night-and-day difference: Hardware 4 confidently reverses into a tight garage in the rain—even adjusting its mirror—while Hardware 3 repeatedly fails, lacks the “brain power” to start from a standstill or back up, and even misses the garage entirely. The host argues that owners who paid for FSD years ago may already be stuck on inferior hardware, since the system improves unevenly across platforms and regions. Another key point is the misconception that personal FSD equals a Robotaxi: a Robotaxi only masters a controlled slice of the environment, whereas personal FSD must handle the entire world under driver supervision. Tesla is moving toward removing driver monitoring within a year or two, and the broader debate is whether FSD is truly safer than humans—though Tesla and regulators may not even measure collisions the same way.

Key Points

  • ▶ 0:00 The hardest FSD test yet: Tesla reverses itself into a garage around another car with essentially no room to spare—even in rain, adjusting its mirror to avoid clipping—something FSD couldn't do months earlier.
  • ▶ 1:00 The video's key questions: Can Hardware 4 beat Hardware 3 at this? Does rain or sun affect the cameras? What happens with more obstacles? And why is Robotaxi considering a completely different approach if FSD is getting this good?
  • ▶ 1:49 The uncomfortable question for owners: if you paid for FSD years ago, your car may already be behind—Hardware 3 and Hardware 4 aren't improving at the same pace, meaning the hardware in your Tesla may decide how far into the future it can go.
  • ▶ 2:31 The host asks viewers to subscribe as the biggest thank-you, and notes that sponsors like this make the videos possible.
  • ▶ 3:06 Monarch Money unifies bank accounts, loans, investments, and credit cards into one dashboard, giving a quick financial overview, spending awareness, and plain-English AI insights.
  • ▶ 4:01 The host recommends Monarch Money for keeping finances simple and private (no data selling, bank-level encryption), with the link in the description.
  • ▶ 4:23 The Model X runs FSD 12.6.4.4 while the Model Y has the latest 14.3.3.3, and Hardware 3 can't start FSD from a standstill like Hardware 4.
  • ▶ 5:16 The car misses the garage entirely and stops with no take-over prompt; the driver attributes this to Hardware 3 lacking "the brain power to pull it off."
  • ▶ 7:11 FSD cannot reverse on Hardware 3 — the car pulls into a neighbor's driveway and later still can't figure out the angle to back into the garage.
  • ▶ 10:51 Hardware 3 vs Hardware 4 is a night-and-day difference; Hardware 4 is "on a different planet" in effectiveness, explaining why many HW3 owners wouldn't pay $99/month for FSD.
  • ▶ 11:16 The street retry failure appears caused by FSD's neural net confusion on Hardware 3—it couldn't decide whether to back in or nose in, when it simply should have nosed into the garage.
  • ▶ 12:00 FSD is a "giant training machine," but not all data is equal: regional data laws (e.g., China) and Europe's regulatory limits mean FSD behaves differently by region and hardware, despite being marketed as one unified system.
  • ▶ 12:36 The biggest misconception: your personal Tesla with FSD is not a Robo Taxi — personal FSD must handle “the entire world” under your supervision, while a Robo Taxi only masters a controlled slice of the environment.
  • ▶ 13:20 Tesla aims to completely remove driver monitoring within a year or two; FSD is already confident enough that brief phone glances aren’t nagged, signaling a major shift in supervision expectations.
  • ▶ 15:56 In the bicycle obstacle-course test, FSD drives forward into a tight garage with confidence—no reversing needed—navigating with only about two inches of clearance.
  • ▶ 17:16 The driver disengaged due to nervousness during the garage reverse, even though the steering wheel had begun to turn — making it unclear whether the car would have hit or cleared.
  • ▶ 17:44 The key assessment: if the car had continued straight back it would have hit, but because it was starting to turn, it might have cleared.
  • ▶ 18:04 The bigger debate is whether FSD is safer than a human driver — with the core issue being that Tesla and regulators may not even be measuring collisions the same way.

Video Sections

  • ▶ 0:00 Introduction and Key Questions (0:00 - 2:31) - - The intro frames the hardest FSD reverse-parking garage test and raises the Hardware 3 vs Hardware 4 uncertainty.
  • ▶ 2:31 Sponsor: Monarch Money (2:31 - 4:23) - - A sponsor break for Monarch Money with a quick reminder to support the channel.
  • ▶ 4:23 Hardware 3 FSD Parking Struggles (4:23 - 9:56) - - Hardware 3 repeatedly fails to back into the garage, gets stuck in FSD, and draws a neighbor interaction.
  • ▶ 9:57 Hardware 3 Retry and Hardware 4 Comparison (9:57 - 12:35) - - Street retry still shows Hardware 3 limits, and the narrator compares Hardware 4, FSD data, and regional training.
  • ▶ 12:36 FSD Misconceptions, Obstacle Course, and Incentives (12:36 - 16:37) - - Robo taxi misconceptions are cleared up, then a bicycle obstacle course shows FSD driving forward with confidence.
  • ▶ 16:37 Sun Glare, Safety Data, and Closing (16:37 - 19:05) - - A sun-glare close call leads to FSD safety-data discussion and final comments.

Exact Transcript

Load the full timestamped transcript on demand and click any time to jump in the video.