1k
499
0
199
388
3471Elon Musk is doubling down on his long-held belief that Tesla doesn’t need LiDAR (Light Detection and Ranging) sensors for its self-driving cars, even after a viral YouTube video showed the technology struggling in a basic obstacle test.
“People don’t shoot lasers out of their eyes to drive,” Musk posted on X, sharing a clip from Tesla’s 2019 Autonomy Day. “Just try Tesla self-driving today, which just uses cameras and AI, and you will understand.”
His comment probably came in response to a video from popular YouTuber and former NASA engineer Mark Rober. In the video, Rober tested Tesla’s Autopilot system against a LiDAR-equipped setup by setting up a fake wall that looked like a continuation of the road — a classic “Wile E. Coyote” trap. The Tesla, relying only on cameras, plowed straight into it.
The internet went wild.
While Musk criticized LiDAR as an expensive and unnecessary tool, calling it a “fool’s errand” back in 2019, the real story is more complicated. According to a 2024 earnings report, Tesla quietly became LiDAR manufacturer Luminar’s biggest customer, accounting for more than $2 million in revenue in the first quarter alone. That’s 10% of Luminar’s revenue.
A Tale of Two Teslas
YouTuber Kyle Paul decided to run the same test, but this time using Tesla’s Full Self-Driving (FSD) system instead of Autopilot. His 2022 Model Y, which runs on Tesla’s older Hardware 3 (HW3), failed every time. The car only noticed the wall when it was inches away — and likely only because of its parking sensors.
Then came the twist: Paul tried the same test with a Cybertruck running Tesla’s latest FSD version and newer Hardware 4 (HW4). This time, the vehicle passed with flying colors, slowing down and stopping on its own before hitting the fake wall.
The results suggest that not all FSD systems are equal, and newer hardware makes a real difference.
Still, some commenters pointed out that neither Paul nor Rober tested the system in tougher conditions.
Ten Years of Promises
Musk has long painted a rosy picture of autonomous driving. Speaking at an Nvidia conference in 2015, he said, “I view it as a solved problem. We know exactly what to do and we will be there in a few years.”
By 2019, he even claimed that buying a Tesla was like investing: “If you buy a Tesla today, I believe you are buying an appreciating asset, not a depreciating asset.”
Fast forward to today, and that hasn’t panned out. Used Teslas are depreciating faster than average, and the FSD software has seen multiple price cuts.
Musk recently admitted Tesla’s Hardware 3 isn’t capable of full autonomy, calling the realization “absolutely painful and difficult.” He added that some drivers even turn off self-driving to check their phones without being nagged by alerts.
“Right now, there is this perverse situation where people actually go to manual driving to check their text messages so the computer doesn’t yell at them,” Musk said.
Despite all the delays and criticism, Musk now says Tesla is rolling out a supervised robotaxi fleet in Austin, Texas later this year.
Competitors Gaining Ground
While Tesla sticks to its camera-only approach, others are racing ahead with LiDAR. Companies like Waymo and GM Cruise rely on the technology, and it shows in performance. One longtime Tesla investor, Ross Gerber, was very direct, saying, “My Tesla FSD is like a 12-year-old driving next to a Waymo,” he said.
Chinese companies are also dominating the LiDAR space. Hesai Technology, for example, now holds 47% of the global LiDAR market and is partnering with BYD to add sensors to over 10 new EV models launching in 2025.
Back in February, Tesla’s sales in China dropped to their lowest level since July 2022, with just 30,688 units sold.
As the debate over self-driving tech rages on, one thing is obvious: Tesla’s road to autonomy is still filled with speed bumps—and maybe a few fake walls.
1k4990199
3883471
