Tesla proves technology is not always the solution

Elon Musk’s conviction that more tech is always better is hurting Tesla — and killing people

Tesla vehicle crashed while under Autopilot control on March 23, 2018. Source: KTVU

It would be hard to find someone who better embodies the ideology of Silicon Valley than Elon Musk — even though he lives in Los Angeles. The tech messiah who founded a range of future-focused companies doesn’t back down from his belief that technology is always the solution to the problem he identifies, be it to reduce the cost of tunneling, liberate drivers from having to drive, or make car manufacturing more efficient.

But even though Musk is unwavering in his belief that more tech is always better, reality is catching up to him. The cost savings he promises to attain through the Boring Company have already been achieved with better contracting practices in several other countries, his initial cost estimates for Hyperloop have been shown to be completely out to lunch, and now the troubles at Tesla are mounting, mainly because his tech-dependent approach is failing.

There’s no denying that Tesla made electric cars cool and desirable with its Roadster before slowly expanding its lineup to include slightly more affordable and similarly appealing models. However, the technological side of Musk’s master plan isn’t working out as planned, resulting in missed deadlines on Model 3 production and Autopilot features, which call the company’s future into question. In order to solve these problems, Musk may finally have to accept that more technology isn’t always best; yet whether he’s capable of such a realization remains an open question.

When automation isn’t more efficient

In true Musk fashion, when the electric-car baron built his Tesla factory in Fremont, California, he promised it would be more efficient than other vehicle factories because its production would make greater use of automation. Musk asserted this would allow Tesla to churn out more vehicles in less time and at a lower cost by reducing the necessary human labor.

However, ignoring what other companies have already learned about the projects he undertakes has become a dependable Musk trait. Just as he didn’t bother to see if other jurisdictions were building tunnels much cheaper than the US before rushing into the Boring Company, he also doesn’t seem to have looked at GM and Volkwagen’s previous failed attempts at automating final assembly; thus recreating the same problems.

Tesla is currently under a massive crunch. After years of missing production deadlines, the company is buried in debt and its credit rating has been downgraded “six levels below investment grade” by Moody’s, so Musk needs to show investors he’s making progress. Tesla produced 2,020 Model 3s in the final week of March, below the 2,500 per week goal, which Musk assured investors will rise to 5,000 by June. Whether current numbers can be sustained, let alone increased, remains to be seen as workers had to be taken off the Model S and Model X production lines just to get to 2,020. It’s also worth remembering that Tesla’s workers are paid less than the industry standard and injuries at its Fremont factory are 31 percent higher than the industry average — that’s before the increased pressure that management must have placed on them to churn out more Model 3s.

This was never supposed to be a problem, of course, because robots were expected to meet the production targets. Musk tried to aggressively automate the final assembly process — putting all the components into the vehicle — but, like other car companies before Tesla, he failed spectacularly. There have been stories for years about workers having to fix Tesla vehicles once they roll off the assembly line; a November 2017 report claimed more than 90 percent of Model S and Model X vehicles had defects which had to be repaired after assembly, compared to less than 10 percent of vehicles at Toyota.

The failure of Tesla’s hyper-automation strategy has resulted in higher costs for the company. According to a Bernstein report, “automation in final assembly doesn’t work.” The robots are not able to perform the intricate tasks necessary to install all of the components, and even if some assembly workers can be eliminated, engineers with higher salaries need to be hired to manage and maintain the robots; plus there’s the cost of the machines themselves.

Roger Bohn, a professor at UC Davis and expert on high-tech manufacturing, agrees with this assessment:

Fundamentally, Tesla has a product design and production process that are “not manufacturable.” That is, the product tolerances are considerably tighter than the process variation. The result is that they produce lots of junk that must be scrapped or reworked. They can partially reduce process variation by stopping more often to adjust machines, but this causes downtime and creates “bottlenecks.”

Musk went all in on automating final production and failed. He burned a ton of money in the process, hurt his workers, and missed his production targets time and time again. Now investors are angry and looking for evidence of progress — and that may be the only thing that can force Musk to break out of his bubble and recognize that his heavy reliance on robots has failed and he may have to rely on humans to build his cars after all.

But the assembly line isn’t the only place where Musk’s insistence on automation is failing. While factory issues are costing the company money and credibility, the ongoing issues with the Autopilot system is costing lives.

Self-driving Teslas are killing their drivers

Since a self-driving Uber test vehicle killed a pedestrian in Arizona a few weeks ago after its sensors failed to detect her and its “safety” driver was looking at her phone instead of at the road, there’s been increased attention paid to the question of whether driverless vehicles are really as safe as tech titans have claimed — and it looks very unlikely.

In that discussion, the fact that a Tesla vehicle killed its own driver for the second time less than a week after Uber’s fatal collision hasn’t gotten nearly the same level of attention — but it should. Musk and Tesla are selling a vision of self-driving vehicles that doesn’t yet exist, and it makes drivers believe that the autonomous systems in their vehicles have more functionality than is actually available.

On its website, Tesla claims that all vehicles “have the hardware needed for full self-driving capability at a safety level substantially greater than that of a human driver,” but the company has no actual evidence to back up that claim. It’s just one of Musk’s platitudes. It further describes a bunch of features on its Autopilot page, even though the capabilities of Enhanced Autopilot are still being rolled out and, under Full Self-Driving Capability, the company states it “is not possible to know exactly when each element of the functionality described above will be available.”

Autopilot suffers from the same issue as nearly everything Musk touches: it keeps missing its deadlines. Even though Musk loves to show off autonomous features, they haven’t been rolled out. In October 2017, Andrew J. Hawkins reported that “Tesla vehicles built since October 2016 have many fewer safety and convenience features enabled than in older models” because the company cut ties with its previous partner, Mobileye, and hasn’t been able to replicate the capabilities it lost as a result.

Navigant Research

Tesla is also the only company working on self-driving vehicles that has abandoned LiDAR sensors, which provide a three-dimensional picture of the surroundings (even in low light), mapping the area in a way radar and cameras cannot. Uber’s choice to reduce the number of LiDAR sensors on its vehicles from seven to just one on the roof is being considered as a possible contributor to its vehicle’s fatal collision. It should probably come as no surprise that Navigant Research, which does an annual ranking of autonomous vehicle systems, places Tesla in last place, further calling into question the claim that Autopilot is safer than human drivers.

When it investigated the first instance of a Tesla vehicle killing its driver in May 2016, the National Transportation Board found that Autopilot allowed the driver to use the system in areas it was not designed for and didn’t require him to keep his hands on the wheel, which is generally required for the level of functionality it had. Initial reports about the most recent incident suggest the driver also didn’t have his hands on the wheel and wasn’t paying attention to the road; the family even told the media that the driver “had complained to his Tesla dealership that his SUV would swerve toward the same median where he was later killed.” How could the vehicle’s sensors not have picked up the median in its path?

In the aftermath, Tesla defended Autopilot and suggested that the system is only “semiautonomous” and that drivers still need to remain alert and should keep their hands on the wheel. That’s the truth, but it’s not at all how the company’s website frames the system, nor how Musk promotes it in public appearances.

In 2015, Musk said that “in the distant future, people may outlaw driving cars because it’s too dangerous. You can’t have a person driving a two-ton death machine.” Yet again, his faith in technology has blinded him from seeing how he may have created just such a death machine. Musk promotes his vision of a self-driving future, while ignoring the flaws that exist in the system he’s selling to Tesla customers. This is incredibly irresponsible and is endangering unsuspecting drivers who believe his assurances about the safety of Autopilot.

Elon Musk needs to wake up

There’s nothing wrong with using technology to improve productivity or make life easier for people, but when technological solutions are pursued with the implementation of that technology as the goal, rather than socially beneficial outcomes, one’s priorities need to be reexamined.

In the cases of Model 3 production and the implementation of Autopilot, Musk is guilty of pushing automation for the sake of using more technology, regardless of whether it’s delivering the desired outcomes. He’s so preoccupied with automation that the production problems and Autopilot’s growing death toll have been brushed off as necessary sacrifices on the path to full implementation, but that cannot be allowed to continue.

It seems that Tesla’s investors are finally getting Musk to address the production issues, but whether he’s actually finding a solution for the high defect rate or is simply pushing workers harder to churn out more vehicles is currently unknown. That hasn’t changed anything with Autopilot, however, where Tesla continues to assert its technology is safer than human drivers, even though its system is the worst in the industry and has a higher body count than any other.

Just as Elon Musk is the best representation of Silicon Valley’s ideology, his approach to automation at Tesla is also representative of how the industry approaches technology. Companies pursue digitization and automation as their primary goal, often failing to examine whether those actions are actually producing positive social outcomes. As more people turn against Big Tech and support for antitrust action grows, tech leaders must act to address the negative aspects of their “disruptions,” but whether they’re capable of escaping their bubble to do so seems very unlikely.

Radical Urbanist

Cities have more power than ever to shape the future. In Radical Urbanist, @parismarx explores their approaches to transportation, climate change, and monopolistic tech companies to see if they’re doing enough.

3.1K

3.1K claps
Paris Marx

Written by

Socialist, traveller, urbanist. MA Geog, McGill. I write critically about tech and cities, and curate the Radical Urbanist newsletter: http://bit.ly/radurb

Radical Urbanist

Cities have more power than ever to shape the future. In Radical Urbanist, @parismarx explores their approaches to transportation, climate change, and monopolistic tech companies to see if they’re doing enough.