Brad Ideas

Robocars & More

The IIHS (Insurance Institute) released a study claiming that in spite of claims that self-driving cars could prevent 90% of accidents (the ones where the driver is at fault) the number was only 1/3rd, namely the perception errors and impairment cases. I am not sure they could have got it more wrong, and outline this in a new Forbes.com article:

IIHS is wrong in claim that only 1/3rd of crashes can be prevented

So, yesterday in Taiwan a Tesla, claimed to be on Autopilot, smashed in broad daylight on a mostly empty road into a truck lying on its side, almost hitting the driver standing in front of it. While Tesla Autopilot is just driver assist, and not meant to catch everything, it's not great to miss a giant truck and a human. I explore why in my new Forbes site article found in at Tesla in Taiwan crashes

I reported last week about Zoox shopping for a buyer. Now reports have surfaced that Amazon is in negotiations. It's a strange matchup but would still have big consequences. A big push into self-driving by Amazon could upend logistics and retail, but Zoox's efforts at a vehicle custom designed for taxi service might be wasted.

Read about it on Forbes.com at Match and Mismatch of Amazon buying Zoox

In a year when several other companies have slowed development and plans for full robocars, MobilEye's CEO this week indicated they were on track to deploy Robotaxi service in Israel in early 2022, and will follow on with France, Korea and China.

Analysis is at MobilEye reaffirms prediction of robotaxi service by 2022

On May 11, a major event took place in the bitcoin world, yet it had no negative effect on the price of the coin and much less effect on mining than it would seem it should. This event is known as the "halving," and it means the reward for mining bitcoins was suddenly cut in half.

I didn't pick the topic, in spite of having written a bunch on in yesterday, but tomorrow we will do another debate in our "Zoom Tank" on the topic of the good and bad news for roadway public transit.

The main debaters are Jarrett Walker from Human Transit, and Randal O'Toole, the "antiplanner" from Cato Institute who is always full of amazing data.

After the short debate, the four sharks, including yours truly, tear into the debaters and their arguments, and then the audience has a go.

Earlier I wrote about how EasyMile had to stop operations after a sudden stop in one of their vehicles gave minor injuries to a passenger.

Today they announced their plan to resume operations. It includes seatbelts and education. It's a start, but I wonder if the whole idea of "stops" is the problem. Stops are inherent in the 20th century thinking that surrounds public transit. Big vehicles need to make lots of stops picking up and dropping off passengers. But that's a problem if you expect the vehicle not to start until everybody has their seatbelt on!

Everybody shut down testing during the Covid crisis. It's not over but now Waymo and others are getting back on the road, testing vehicles with nobody inside, with one safety driver, and sometimes with two. Yes, delivery and rides are essential services, but what are the issues around this?

Read some thoughts at Forbes's site in Waymo and others resume testing

I've known Zoox since before it began and their vision has always been bold. In a possible hiccup, the downturn has led them to shop around for a potential buyer if they can't get more investment.

I analyse what this means in my story Zoox searches for a buyer on the Forbes site.

Many people opposed to lockdowns feel they are an improper state interference with our liberty. Possibly an unconstitutional one, particularly in the case of religious gatherings.

Zoox is the $1B funded startup trying to build a radical design self-driving car. Last week they released a video of an hour long drive through Las Vegas, going through pick-up zones in hotels and the airport. The car does a number of impressive things, but at the same time, showing these things and an hour of driving are only the tip of the iceberg of what you need to do, making the video unimpressive at the same time.

In this new article, I go into what goes on in the video, and what it means.

Ready to get on a 10 hour overseas flight, wearing an N95 mask, sharing bathrooms, in the middle seat between two coughing passengers? I didn't think so. With all the idle cruise ships out there, would you sail across the ocean like your grandparents in a private cabin if they followed good virus procedure? 4 days stuck in a room (kinda like now) to prove that your virus test is accurate to the country you're heading to.

Last year Tesla released "smart summon" which let you (very slowly) call your car to you from across a parking lot. It was cute but a bit of a dud, as it's not just very useful. Now Elon Musk promises "reverse summon" that will valet park your car for you. But if you have to watch it, it's not going to be very useful either.

Eventually, though, we'll get a robotic valet park that works without supervision. That will be very useful, allowing cheaper parking and better charging. Even today, the basic summon could allow slightly denser parking for cars that have it.

Next Monday, we will be doing an online version of a popular panel session we have done every year at the "Automated Vehicle Summit" -- the oldest self-driving car conference.

In this session we have a speaker propose a controversial idea, and then, a bit like a "shark tank" our panel (myself included) tear into the concepts and discuss them, and the audience asks questions too.

If you read my earlier report on efforts to convert CPAP machines into ventilators with new firmwware the good news is that the feared massive ventilator shortage seems (for now) to have been avoided.

Tesla doesn't want to use LIDAR. So they are hoping for success in a technique known as pseudo-LIDAR, where you train neural networks to look at images and calculate the distance to everything in the scene, as though you had a LIDAR. It's not here yet, but an interesting question is, should this succeed, is it better for Tesla or for their LIDAR using competitors who already have tons of experience using 3D point clouds?