RoboTrucks, Robofactory and RoboEverything Interview at Brighter with Herbert

by

I, Brian Wang, was interviewed at the Brighter with Herbert Youtube channel on Robotaxi, Robotrucks, robofactories and roboEverything.

Tesla will succeed with Full Self Driving and Teslabots. It may still be two or three more years until the full robotaxi level is reached. It could even be five years but the progress is and will speed up this year and beyond. Robotaxi could require the system being 10 to 100 times safer than the average human driver The economic impact of a very comfortable drive that is as safe or twice as safe as an average human driver will be very large.

The autopilot and FSD systems already improve the driving experience and improve safety while still requiring the human driver to pay attention and occasionally intervene. However, FSD and Autopilot are driven all over the USA, Canada, Europe and Asia. Autopilot has been used for over ten billion miles of driving and FSD beta has been used for over 320 million miles. Tesla FSD beta is now adding about 50 million miles every month.

The rate of miles driven and the compute power for AI training are both increasing exponentially.

Robotaxi could increase the value of car by 3 to 5 times based upon the hours it can be driven. However, many people will choose to continue to drive themselves even if it costs more. People buy luxury cars when they could buy and use an adequate basic economy car. Many people in New York still choose to drive cars despite a very good public transit system and taxis. This is in spite of high costs and difficulties parking.

Tesla’s AI team recently announced that the software can train with Youtube videos of driving. This means the system is independent of camera positions. This is critical for being able to license FSD and Autopilot to other car makers. The different cars can have cameras in other locations.

RoboTrucking

Robotrucking will be more impactful than Robotaxi in many ways. If truck drivers are replaced for all 30 million large trucks and perhaps another 100 million small and medium sized commercial trucks in the world this will transform the supply chain.

The transport costs in the supply chain could eventually be reduced by 80% or more. Driving speed for delivering cargo could increase from 55 mph to 110 mph or more. Also, the robotruck could drive 23 hours a day and only stop for recharging.

Truck fleet owners will not hesitate to replace human drivers once the systems are good enough. 70% of drivers could be replaced once there is a driving system that could safely platoon trucks.

As stated before, Tesla’s AI team recently announced that the software can train with Youtube videos of driving. This means the system is independent of camera positions. $150,000 semi trucks can get cameras and self driving computers added for about $20,000 to $30000 or $5,000 per year plus the hardware. This means the existing fleet can be rapidly converted to remove the driver for a 4-12 month return on investment.

$5,000 per year made on all 30 million trucks would be $150 billion per year. This would be over ten times the current profits at Tesla.

$150 billion per year in profits would mean a 10X in value. This would not need as many batteries. It would only need 30 million robotruck conversion kits (hardware 4 or 5 chips, cameras etc…).

RoboFactory

Teslabots are making rapid progress and are leveraging the training data and AI training systems of FSD. Tesla is building the factory to make the custom actuators and mass produce Teslabots. Teslabots have already been shown to walk, move objects and handle tools and wires.

If one to four Teslabots help every human factory worker, this could reduce the number of human factory workers. It can mean the ability to shift the human workers and spread 100,000 human factory workers over 8 factories instead of 4 factories. Production would rapidly increase.

The safety and regulatory issues with Teslabot in the factory would be less than for driving cars or trucks. This could mean the more rapid large scale deployment of Teslabots in Tesla factories than robotaxi or robotrucks.

The Tesla GigaMexico factory will use the unboxed process. This will make factory processes Teslabot friendly. Teslabots will not need to climb inside a partially complete car frame. They can work on all of the parts until they are finally assembled.

Tesla sells Model Ys for about $50000, but it currently costs them about $30000 to make them. A Teslabot is 1/30th of the mass of a Model Y. It would use 1/30th of the batteries. The software is an overall cost of development. If billions of bots are produced then the cost would trend toward the cost of the hardware plus Apple iPhone-like margins including the software (say 40% gross margin). At Model Y cost of $30k then the hardware cost for Teslabot will go to $1000. $2000 with margins and software. A bot can work for 8000 hours in a year. 8760 hours in a year. $2000 divided by 8000 hours is $0.25. If you add 10 cents per hour for electricity then it is $0.35 per hour. Going beyond that is bots can work in the factory and work cheaper than humans. Currently 15000 workers in Tesla China factory. Replace all of them with $0.35 per hour bots. Reduce labor cost component. If a lot of bots can increase production rates. by 2X then all costs spread over more units. Bot-produced solar and batteries can lower the cost of energy by vastly increasing the supply. Those trends could get us to $500-1000 per bot costs and lower energy costs.

RoboEverything Means RoboTaxi is Just the Beginning

Teslabots can be proven safe for residential and office situations. Teslabots and robotic machines can be involved in all aspects of the supply chain like mining and construction. This would enable bots to build factories, building and more bots.

Teslabots would have 100 to 1000 times the compute and power of a smartphone. They could hold more powerful LLM, sensors and communication systems. They will be able to listen and communicate via voice commands like Google Home.

Humanoid robots will be more common than smartphones and will lead to colonization of the Solar System and a Singularity.

6 thoughts on “RoboTrucks, Robofactory and RoboEverything Interview at Brighter with Herbert”

  1. Judging safety merely by miles traveled alone is deeply misleading. Judged that way, the crewed Apollo missions to the Moon would have been the safest mode of travel ever: they went millions of miles without a single fatality (aside from the renamed Apollo 1 test firing that killed 3 astronauts). Not even the ill-fated Apollo 13 mission injured a single astronaut. But does anyone really believe this?
    A much better measure is Unique Vehicle Conditions Encountered, though UVCE is hard to define.
    More importantly, autopilot cars tend to have unpredictable accidents. While humans have more accidents at night, during storms, etc. and can adjust their driving accordingly, autopilot cars seem to malfunction almost at random, at least from a human perspective. Tests show cars veering out of lane in bright sunny days, confused by simple things humans barely have to think about.
    Most importantly, who is liable for a crash involving FSD? Until that is resolved, there can’t be any FSD without a human operator to take over, and to take personal liability in an accident. Tesla has been very consistent in deflecting blame from its products back to the traditional driver, no matter what iteration of autopilot its cars have.

  2. I don’t know if full self driving will be achieved in the next 5 years. There are too many unexpected situations.

    Full self driving would save a lot. Just look at the trucks. Convoy of self driving truck would be way more effective, because drag is reduced if you drive behind truck. Then less electricity would be used for same mileage. Not to mention more effective driving habits and no cost for human driver,…. Logistics would get way cheaper, faster(no need to rest). Goods will get cheaper because more effective logistics.

    As for the personal robotaxis. 10 cent per km can be achieved with ridesharing. If you want to somewhere and pick a few people along the way. Would take a little longer but would still be way faster than public transport.

    To use cars for a fraction of the day is not efficient. There will be less old cars on the road. Robotaxis will get more mileage in a short period than normal cars. The peak hours could be a problem, because most people need car to get to work and back at specific hours. Then it is not used as often.

    Benefits would be big. If it would be applied at massive scale, we would not longer need traffic lights and reduce congestions, get faster and cheaper to destination.
    I think that people will still want to drive, so traffic lights won’t be abolished for quite some time.

      • Do you invest in Tesla, Brian?

        This has been news for a week, now. Were you still unaware?

        Range calculations cannot possibly be *that* complicated, if Tesla was able to write fraudulent estimation software so sophisticated that they overestimated the range of full batteries, then dropped back to more accurate ones after they were less than 50% capacity.

        This is a decade long pattern of fraud, and it is obvious that Tesla *knows* it is fraud because they conspired to keep this out of the public eye, and even conspired to keep current owners from understanding the problem.

        I will quote the article directly:

        “Tesla years ago began exaggerating its vehicles’ potential driving distance – by rigging their range-estimating software. The company decided about a decade ago, for marketing purposes, to write algorithms for its range meter that would show drivers “rosy” projections for the distance it could travel on a full battery, according to a person familiar with an early design of the software for its in-dash readouts.”

        “Then, when the battery fell below 50% of its maximum charge, the algorithm would show drivers more realistic projections for their remaining driving range, this person said. To prevent drivers from getting stranded as their predicted range started declining more quickly, Teslas were designed with a “safety buffer,” allowing about 15 miles (24 km) of additional range even after the dash readout showed an empty battery, the source said.”

        Elsewhere:

        “What Ponsin didn’t know was that Tesla employees had been instructed to thwart any customers complaining about poor driving range from bringing their vehicles in for service. Last summer, the company quietly created a “Diversion Team” in Las Vegas to cancel as many range-related appointments as possible.”

        “The Austin, Texas-based electric carmaker deployed the team because its service centers were inundated with appointments from owners who had expected better performance based on the company’s advertised estimates and the projections displayed by the in-dash range meters of the cars themselves, according to several people familiar with the matter.”

Comments are closed.