Even After $100 Billion, Self-Driving Cars Are Going Nowhere

Will self-driving cars solve our traffic and street safety problems? A tech historian says no

Ryan Fonesca

Norton argues there’s a much more straightforward, proven solution: Build more housing closer to where people need to travel. “Proximity, in short, is better than all the tech in the world in terms of making mobility safer, with less congestion,” he said. Then there’s the prevalence of unsafe speeding, which Norton also tied to urban and suburban sprawl. His suggestion: “Create environments where you don’t need speed, because your destinations are close.”

So what would it take to actually address those issues? Norton said first and foremost: “Not letting the people who sell us tech and sell us cars also be the people who tell us what our so-called solutions are.” “Tech companies and automakers should not be confused with policymakers,” he said. “We have made that confusion very pervasive, and I’m not that optimistic that we’ll resist that successfully.”

Cities, states and federal officials could also do more of the things that have proved to make streets safer and less congested in other, similarly developed nations, Norton suggests. He points to Western Europe, where streets have become safer due to meaningful efforts to bring down vehicle speeds, create more active transportation infrastructure and rethink the public spaces devoted to private cars.

Self-Driving Investment Crash: 58% Drop In Autonomous Vehicle Dollars

John Koestier

The self-driving vehicle has crashed. Not the car itself, although that happens, but certainly the investment craze that enveloped the venture capital world over the past few years. In 2021, investors poured a record $9.7 billion into autonomous vehicle development. Last year however, that dropped by almost 60% to just $4.1 billion. Overall investment in robotics, which have totaled $90 billion globally over the past five years, is also down, hitting $12 billion in total last year versus $18 billion in a much more optimistic 2021.

The big winner recently: vertical robotics that are purpose-built for specific industries such as warehousing, shipping, health, defense, manufacturing, and construction.

I think we are in danger of ignoring the main relevance of driverless cars to the issue of the true nature of the human mind.  I mean it could be possible that driverless cars are just too expensive to develop, but could exist in principle, or that people really want an alternative to cars altogether, so that this technology is too late in the day. (Even though to be honest, I don't really believe either of those suggestions).

I have been recently re-reading James Carpenter's book, "First Sight". It is a somewhat turgid read, but it proposes an interesting idea. He asserts that the endless variations and successful repetitions of Dean Radin's presentiment experiment are showing us that human consciousness (maybe all consciousness) enjoys some form of advance knowledge of what is about to happen next - even when that is based on a quantum-based stream of random numbers. He seems to suggest that every event that enters our consciousness starts off as an unconscious psychic alert that helps us to use the conventional information that may follow.

Presumably, this is completely lacking in any driverless car, and it may explain the problems engineers are having making these vehicles safe, even in a benign urban setting.

The development costs for these cars are phenomenal, and a healthy slice of that money (after bribes to the regulatory authorities) must have been spent on software (in its broadest sense). I'd suggest that there must be a law of diminishing returns here, and that safe, driverless cars are simply not possible if they are to mix with other vehicles and people.

It is also worth remembering that when humans are taught to drive, this starts in quiet backstreets, and gradually moves on to main roads and eventually to motorways. Those quiet back streets are going to present all kinds of hazards such as children playing, non-standard road layouts, pot holes, junk in the road, etc etc. The fact that the learning process is reversed in this way is rather suggestive.

David

Tesla’s Full Self Driving Isn’t The Only Technology With Speed Sign Detection Problems

Ramsey Faragher

Last week we woke up to the news that Tesla’s Full Self Driving product was being recalled due to problems, including ignoring the speed limit. Speed limiting technology has also been mandated over in Europe, where the EU’s Vehicle General Safety Regulation has now been in force for a little over 6 months. It introduced a range of mandatory advanced driver assistant systems to all new cars aiming to improve road safety, and establishes the legal framework for the approval of automated and fully driverless vehicles in the EU. The new safety measures will help to better protect passengers, pedestrians and cyclists across the EU, expectedly saving over 25,000 lives and avoid at least 140,000 serious injuries by 2038. The 2019/2044 regulation also mandates all new cars that have already launched be fitted with an Intelligent Speed Assist (ISA) by 7 July 2024. Tesla’s struggles though help to demonstrate how difficult it is to get this technology working in practise.

Machine vision is generally not as reliable as human vision when context and complex environments are important. For example small speed restriction signs on the back of trucks, indicating their maximum speed, could be misread by ISA cameras as the current road speed. Scientific studies have also demonstrated how small changes to speed signs, such as the application of relative benign stickers or marks, can result in machine learning systems mis-labelling road signs when humans would still recognise the correct sign. Similarly, GPS positioning accuracy can be poor in urban environments where signals are blocked and reflected, leading to an ISA system believing it is on, for example; a 30 mph side road when it is really on a 50 mph main road, or vice versa.

Self-driving trucks struggle to deliver

Joann Muller

There's been an industry shakeout over the past year, causing once-giddy investors to pull back, while survivors shed workers and struggle to fund continued development.

Aurora lost $1.7 billion last year and will need to raise more cash to fund its commercial rollout, co-founder Sterling Anderson tells Axios.

The bottom line: Self-driving trucks are stuck in low gear.

There's a lot of attempts at spin from the executives they quote but seems to be the facts are not looking good for driverless car/truck companies hoping to replace human workers with machine "learning" pipe-dreams...

Tesla’s Full Self Driving Isn’t The Only Technology With Speed Sign Detection Problems

Ramsey Faragher

You’d just use GPS for speed limits… that’s part of the reason for the EU’s Galileo satellite network.

You’d just use GPS for speed limits… that’s part of the reason for the EU’s Galileo satellite network.

In theory yes, but it seems to not work consistently. We'll see if it works with the satellite network.

Tesla car in Beta full self-driving mode almost rams cyclist (+ video)

Top Tesla engineer confesses he staged an inaccurate self-driving video at Elon Musk’s bidding

Tesla Officially Admits Its Self Driving Cars Cause Crashes With A Massive Recall

Tesla lawyers claim Elon Musk’s past statements about self-driving safety could just be deepfakes

James Vincent

The company made the argument to justify why Musk shouldn’t give a deposition as part of a lawsuit blaming Tesla’s Autopilot software for a fatal crash in 2018. A judge gave a tentative order for Musk to be deposed.

Lawyers for automaker Tesla have argued that statements by Elon Musk about the capabilities of the company’s Autopilot software can’t be trusted as they could be deepfakes, according to reports from Reuters and Bloomberg.

Tesla presented this argument as part of its justification as to why Musk shouldn’t be interviewed under oath for a lawsuit blaming the company for the death of Apple engineer Walter Huang in a fatal crash in 2018.

Your post fully supports my earlier claim in this thread that the problem with selfdriving cars is mainly a legal issue.

We are proud of Autopilot’s performance and its impact on reducing traffic collisions,” the company states with a January 2023 (https://www.tesla.com/VehicleSafetyReport) Vehicle Safety Report update. The numbers are staggering. In the 3rd quarter, one crash was recorded for every 6.26 million miles driven using Autopilot. Teslas not using Autopilot technology logged one crash for every 1.71 million miles driven.

For comparison, the NHTSA estimated one automobile crash every 652,000 miles. That equates to Tesla drivers without Autopilot engaged being 2.5 times safer and with Autopilot being used ten times safer than the national average.

https://www.notateslaapp.com/news/1144/t...ety-report

Your post fully supports my earlier claim in this thread that the problem with selfdriving cars is mainly a legal issue.


https://www.notateslaapp.com/news/1144/t...ety-report

Seems like nothing more than corporate hype.

edit:

Tesla Again Paints A Crash Data Story That Misleads Many Readers

Brad Templeton

Every quarter, Tesla releases crash data on their cars in various modes. Recently it also released its annual “impact report,” which for the first time included some data on drivers using their prototype “full self driving” system, which has been pre-purchased by several hundred thousand owners. Much of the coverage of the report has described it as presenting an incredibly positive story of Tesla safety. Their raw numbers do seem very good, but the reality is disturbingly different.

I have done analysis of these reports before, generally concluding that what they actually show is that users of Autopilot have a roughly similar number of crashes to those not using it. The Impact report did not cite a number for non-Autopilot use with ADAS active safety, but the quarterly safety report for Q4/2022 reported about 0.71 airbag events per million miles, a bit worse than the annual summary.

Tesla’s number give a very incorrect impression — so incorrect that it is baffling why they publish them when this has been pointed out many times by many writers and researchers. Oddly, Tesla has the real data — they have the best data in the world about what happens to their vehicles. The fact that they could publish the truth but decline to, and instead publish numbers which get widely misinterpreted raises the question of why they are not revealing the full truth, and what it is that they don’t reveal.