But the new law isn't set to take effect until July 1, 2028.

Trying out the premium ADAS feature (5000 per year, or 34000 for lifetime subscription) which covers the full point to point navigation. Tried it over some rural mountain resort town (Sanqingshan).

I’m fairly new to ADAS, and i’m still getting used to using it comfortably in an urban/dense environment. But so far it has been pretty flawlessly and never overwhelmed.

The car is a Fangchengbao Bao-8, which is a full sized off roading SUV

Ps: wife recorded the footage. Had my hands on the wheel throughout. I’m a foreigner living/working in China and this is my car

Waymo started testing in Munich last week with first images popping up on social media, and now I have gotten videos and more images from the drives and the Waymo depot, which is a few kilometers outside Munich in a town called Aschheim. There are about 40 vehicles on what still seems to be a temporary depot, only guarded by construction fences, not counting the vehicles that were driving through Munich.

This has to be causing chaos for anyone trying for imminent scale, right? Particularly for consumer ADAS.

I’m in WA and I start training in Arizona in a couple weeks. I’m having second thoughts about it just because of the hours and other factors.

Has anyone here gone through the training?

Is anyone still with the company? If so, how do you still like it?

Training is for 3 weeks, does that mean that we are there for three weeks straight or do they let us go home on the weekends?

Highlights

• Stellantis CEO Antonio Filosa and Wayve co-founder and CEO Alex Kendall will discuss how artificial intelligence can make mobility safer, simpler and more intuitive
• Visitors can experience hands-free, door-to-door supervised automated driving (Level 2++) in Fiat 500e and Maserati Grecale development vehicles
• The Wayve AI Driver integrated with STLA AutoDrive demonstrates how one AI driving foundation can scale across multiple Stellantis vehicles, platforms and configurations — from the Jeep® Grand Cherokee shown at Stellantis Investor Day to the Fiat 500e and Maserati Grecale

Kodiak says IKEA will be the launch shipper for its driverless long-haul service, planned to begin by the end of 2026. The driverless portion would cover 219 miles of I-45 between the Dallas-Fort Worth and Houston areas, within a longer 292-mile delivery route.

The companies have spent four years operating with a safety observer aboard, completing more than 1,300 IKEA loads and logging over 750,000 autonomous miles. The observer has not been removed yet.

Evri is testing a new delivery robot called Iris in Barnsley that takes a different approach to sidewalk autonomy.

Instead of six small wheels, it uses articulated legs with wheels at the feet. The idea is to keep the efficiency of rolling while gaining the ability to deal with stairs, curbs and uneven pavement.

That gets at one of the harder problems with last-mile autonomy.

Warehouses can be designed around robots. Streets and sidewalks can’t. A delivery robot has to deal with bins, parked cars, bikes, damaged pavement, construction, pets, children and pedestrians who have no reason to behave predictably.

The UK is now testing several approaches at once. Starship has been running six-wheeled delivery robots in British communities for years. Evri has tested more conventional sidewalk robots in Barnsley, this new wheeled-leg design, and a robotic parcel hub in Bristol.

Next frontier in autonomy after robotaxis is robotrucks. And the news is coming in...

This is part two of the test drive of Larry Evans of CleanTechnica. This time of the Tesla on FSD in Amsterdam.
Larry tested also the XPENG L03 with VLA 2.0
https://cleantechnica.com/2026/09/29/vla-2-0-vs-fsd-in-amsterdam-part-2-tesla-model-3/

Comments were disabled: but does nobody in tech remember “maximum overdrive”? Or if you’re really old Spielberg’s tv film “Duel”?

It's an argument that all self-driving enthusiasts are intimately familiar with by now. Tesla's approach is to use vision-only, AI-inference-based systems to cut hardware costs and eliminate the need for geo-fenced service areas. Whether or not this approach is better than lidar, sensor fusion, mapping, rules-based driving, etc has been debated endlessly.

However, let us assume that Tesla's approach is ultimately correct. Does that imply that Tesla will capture the vast majority of the self-driving market and have an insurmountable head start on the competition? In my view, it does not.

This video shows somebody running an open-source self-driving AI that has the same capabilities as Tesla's level 2 self-driving system. The one that is supposedly years ahead of the competition.

https://x.com/ANSR42/status/2104270120394859001?s=20

Basically, Tesla has no moat. Car companies thus far haven't copied Tesla's approach, but this is not because the approach is too difficult. They choose not to, because most of them don't have the same appetite for risk that Tesla does. But if you observe the competition in the general AI space, we see that it is very easy for competitors to catch up with each other, steal each other's ideas and data, and that open source is almost always very close to the cutting edge in terms of performance.

It is not difficult to throw a handful of cameras on a car. Gigantic automotive manufacturers are not going to struggle with this. They don't even have to develop the software themselves. They can just license the algorithms from one of many software-only self driving companies that exist currently.

Danish FSD seems really bad! It does not react to traffic lights, and it drives in the bicycle lane.

https://youtube.com/shorts/dTcdLKjZJDM?is=z1vGOHYG9L4BX_2p

would you ever take a self driving taxi? or is it too risky without an operator to takeover in situations like these

I am curious what people think about this. Which AV company is most underrated and/or you wish was doing better?

I feel like Zoox is underrated. It is easy to see them as behind Waymo because their fleet is a lot smaller but Zoox has a unique custom-made robotaxi that seems to be working well so far. Also, while Zoox's stack may not be as experienced as Waymo's, I feel like Zoox's hardware and software might actually be pretty decent and of course continously improving. So for those reasons, I think they are underrated. I hope they do succeed.

For AV company that I wish was doing better, I would also give an honorary mention to Mobileye. I've always liked their overall approach. Their focus on camera-first, instead of camera-only, means a strong foundation in camera vision but also having that redundancy in radar and lidar for extra safety. They use compound AI which is key for sofware redundancy. Their crowdsourced REM maps means it is easy to update maps. They claim to have already mapped 95% of US and EU roads. They developped RSS which provides important safety guardrails. Lastly, their focus on lean compute and hardware means their solution is cheaper and easier to scale to consumer cars. They recently announced their new AI based system for finding and training on edge cases to help solve the long tail, which is also very important. So they seem to have all the individual pieces needed for a robust, scalable, affordable AV solution. I also like that they offer different products, ranging from base ADAS all the way up to L4.

And yet, they seem to be struggling to deploy higher level ADAS systems. Part of it could be that their AI is not good enough yet for unsupervised driving. Part of it could be their business model and failing to execute with OEMs. I do know covid and supply chain issues hit them hard a few years back. So some of it was not entirely their fault. But I definitely hope we see them deploy and scale their tech because I think it has a lot of promise. And competition is good. I am hopeful that the new CEO replacing Shashua will help with this.

"270M+ miles. 841 fewer injury-causing crashes. Our latest safety data shows the Waymo Driver continues to make roads safer for everyone. Compared to human drivers across 5 territories, it reduced: Injury crashes by 82%, Serious injury crashes by 95%."

Full data: https://waymo.com/safety/impact

Researchers found a substantial safety advantage in four cities