Tesla is the worst performing stock in the S&P 500. Analysts say it has further to fall

[Leongsam]

So Elon was right after all?

TESLARATI​

Ford CEO favors Waymo’s LiDAR approach over Tesla’s vision-only self-driving

June 28, 2025

Ford CEO Jim Farley shared some skepticism about Tesla’s camera-only approach to self-driving during a recent appearance at the Aspen Ideas Festival. When asked to compare Waymo and Tesla’s autonomous driving systems, Farley stated that Waymo’s LiDAR-based approach made “more sense,” citing safety, consumer trust, and the limitations of camera-based models.

Waymo’s LiDAR vs. Tesla’s Vision-Only Approach​

Farley was speaking with author Walter Isaacson when he made his comments about Tesla and Waymo’s self-driving systems. As they were conversing about autonomous cars, Isaacson asked Farley which approach to self-driving he preferred.

“To us, Waymo,” Farley said, though he also stated that both Tesla and Waymo have “ made a lot of of progress” on self-driving, as noted in a Fortune report. He also confirmed that he has had conversations about the matter with Tesla CEO Elon Musk. Despite this, he said that Ford still considers LiDAR as a pivotal part of autonomous driving.

“When you have a brand like Ford, when there’s a new technology, you have to be really careful. We really believe that LiDAR is mission critical… Where the camera will be completely blinded, the LiDAR system will see exactly what’s in front of you,” the Ford CEO stated.

Tesla and Ford’s self-driving plans​

Tesla recently launched a limited Robotaxi service in Austin, which uses autonomous cars with safety monitors in the front passenger seat. While controversial, Musk has maintained that Tesla’s vision-only approach will ultimately prove safer and more cost-effective in the long term. Tesla seems to be making headway towards this goal, with Musk stating recently that the first Model Y has been delivered autonomously to a customer in Austin.

Ford, for his part, is not pursuing its own fully autonomous, urban-driving system anymore. Instead, the company is focusing on “high-speed, eyes-off” experiences like BlueCruise. Ford does plan to partner with a company that has achieved true autonomous driving in the future, as soon as the technology is available.

 

[Hightech88]

So Elon was right after all?

TESLARATI​

Ford CEO favors Waymo’s LiDAR approach over Tesla’s vision-only self-driving

June 28, 2025

Ford CEO Jim Farley shared some skepticism about Tesla’s camera-only approach to self-driving during a recent appearance at the Aspen Ideas Festival. When asked to compare Waymo and Tesla’s autonomous driving systems, Farley stated that Waymo’s LiDAR-based approach made “more sense,” citing safety, consumer trust, and the limitations of camera-based models.

Waymo’s LiDAR vs. Tesla’s Vision-Only Approach​

Farley was speaking with author Walter Isaacson when he made his comments about Tesla and Waymo’s self-driving systems. As they were conversing about autonomous cars, Isaacson asked Farley which approach to self-driving he preferred.

“To us, Waymo,” Farley said, though he also stated that both Tesla and Waymo have “ made a lot of of progress” on self-driving, as noted in a Fortune report. He also confirmed that he has had conversations about the matter with Tesla CEO Elon Musk. Despite this, he said that Ford still considers LiDAR as a pivotal part of autonomous driving.

“When you have a brand like Ford, when there’s a new technology, you have to be really careful. We really believe that LiDAR is mission critical… Where the camera will be completely blinded, the LiDAR system will see exactly what’s in front of you,” the Ford CEO stated.

Tesla and Ford’s self-driving plans​

Tesla recently launched a limited Robotaxi service in Austin, which uses autonomous cars with safety monitors in the front passenger seat. While controversial, Musk has maintained that Tesla’s vision-only approach will ultimately prove safer and more cost-effective in the long term. Tesla seems to be making headway towards this goal, with Musk stating recently that the first Model Y has been delivered autonomously to a customer in Austin.

Ford, for his part, is not pursuing its own fully autonomous, urban-driving system anymore. Instead, the company is focusing on “high-speed, eyes-off” experiences like BlueCruise. Ford does plan to partner with a company that has achieved true autonomous driving in the future, as soon as the technology is available.

Yes, Elon is right. He is not the world's richest man for nothing. Lidar is a ranjiao tech which has been overhyped for autonomous driving as it is more suitable for military, aerospace (e.g. in missiles for bombing missions), and industrial applications, where cost is less of an issue and the need for precision is critical, rather than that for consumer cars. LiDAR (Light Detection and Ranging), despite its precision, hasn’t lived up to the hype compared to vision-based systems (like Tesla’s camera + AI approach). It also explains why the ranjiao overhyped Luminar stock (Lidar maker for cars) plunged like crazy for the past few years and has never recovered ever since LOL.

To samurai: Want to bomb Iran? Use Lidar.

------------------

What LiDAR Does Well​

• Shoots out laser pulses and measures reflections → produces a 3D map (point cloud) of the environment.
• Very accurate in measuring distance to objects.
• Works in low light or at night (since it doesn’t depend on visible light).

The Problems With LiDAR in Practice​

1. High-end automotive LiDAR sensors used to cost thousands of dollars each.​

• Even with falling prices, it’s still more expensive than cameras.
• Multiple LiDAR units needed for 360° coverage → adds cost, power consumption, and design challenges.

2. Resolution (detail in the image) is much lower than cameras.​

• A LiDAR map can detect “something is there” but struggles with classification (is that object a child, a dog, a bag of trash?).
• Cameras + neural networks are far better at understanding what the object is.

3. Rain, fog, snow, dust scatter LiDAR beams and reduce accuracy.​

• Cameras suffer too, but advanced vision algorithms can sometimes compensate better (pattern recognition, temporal smoothing).

4. LiDAR provides geometry, but no semantic understanding.​

• Vision systems, trained on millions of labeled images, can recognize stop signs, lane markings, pedestrians, etc.
• Tesla’s bet: neural networks + vision replicate human driving (since humans drive using eyes, not lasers).

5. LiDAR units are bulky, moving-part devices (though solid-state LiDAR is emerging).​

• Mounting multiple LiDARs on consumer cars is harder to mass-produce and maintain than camera systems.

Why Vision-Only Has Pulled Ahead​

1. Human Analogy: Humans drive with two eyes (cameras) + a brain (AI). If AI gets good enough, LiDAR may be redundant.
2. Data Advantage: Tesla has millions of cars on the road collecting video → huge training set for its neural nets. LiDAR companies don’t have this scale of real-world data.
3. Cost Efficiency: Cameras are cheap, tiny, and already integrated in cars for ADAS.
4. Software-Centric Approach: Vision-only relies on software updates and AI improvements, which scale faster than rolling out new hardware.

So Why Do Some Still Believe in LiDAR?​

• LiDAR excels at precise depth measurement (important for mapping, robotics, and slow-speed navigation like warehouses, mining trucks, robo-taxis in geo-fenced cities).
• Companies like Waymo still use LiDAR because they want redundancy and precision in controlled environments.
• The “sensor fusion” camp argues best safety comes from combining LiDAR + radar + vision.

In short:
LiDAR didn’t “fail” as a technology — it just hasn’t scaled well for consumer cars. The hardware is costly, less versatile in classification, and hard to mass-deploy compared to cameras. Vision-only systems are cheaper, more scalable, and increasingly capable as AI improves. That’s why Tesla’s approach (controversial as it is) has pulled ahead in commercial viability.
--Source: ChatGPT------

Why Lidar Struggled in Consumer Autonomous Driving

1. Early automotive Lidar units cost tens of thousands of dollars per sensor.​

• Even with cheaper solid-state Lidar today, unit costs are still hundreds to a few thousand dollars each, versus a few dollars for cameras.
• For mass-market cars, this makes scaling expensive.

2. Lidar gives a precise 3D point cloud but low-resolution “dot maps” of the environment.​

• Vision (cameras) provides high-resolution, semantic-rich images — color, texture, road signs, traffic lights.
• Neural networks today are extremely good at extracting meaning from visual data — much harder with sparse point clouds.

3. Cameras can do both object recognition and depth estimation with stereo vision + AI.​

• Lidar adds another layer of hardware + software + calibration complexity.
• Rain, fog, or snow can scatter Lidar beams, reducing reliability.

4. Tesla argues that since humans drive using only eyes (vision) and brains (AI processing), autonomous cars should be trained the same way.​

• Tesla invested billions into AI models that rely purely on cameras + radar (earlier) and now increasingly camera-only systems.
• This reduces costs and creates a single, unified sensor framework.

Why Lidar Still Matters (Other Sectors)​

Even though consumer autonomous driving shifted to vision-first, Lidar is far from useless — it thrives in sectors where cost isn’t the main concern, but precision and safety are critical:

1. Autonomous drones, missile guidance, terrain mapping.​

• Works in GPS-denied or jammed environments.
• High-precision detection of objects in low-visibility conditions.

2. Used in docking systems, planetary landers (NASA’s Mars missions use Lidar for landing).​

• Aircraft obstacle avoidance and navigation.

3. Mining trucks, warehouse robots, port automation.​

• Surveying, construction, and mapping with centimeter accuracy.
• Railway safety systems.

Here, the ROI of precision outweighs cost.

Why Investors Fled Auto Lidar​

• Initial hype assumed Lidar would be “standard” in every autonomous car.
• When big OEMs like Tesla went vision-first and others (like Mercedes, GM) slowed L3/L4 autonomy rollout, demand forecasts for Lidar collapsed.
• The total addressable market shrank drastically — leaving Lidar more niche than mainstream.

In summary:

• Lidar struggled in cars because it’s costly, harder to scale, and provides less useful semantic data compared to camera-based AI.
• Vision-first approaches (like Tesla’s) are winning in mass-market autonomy.
• But Lidar still has a bright future in defense, aerospace, and industrial sectors, where precision, reliability, and safety matter more than cost.

-------Source: ChatGPT-------------