Nope, not autonomous. They have remote pilots. Good try though and thanks for playing.
1 Like
In the form of twiddling the steering wheel, which serves to confirm the “driver” is not asleep, but provides no input to the actual driving.
1 Like
irrational absurd statements
Ah. I think I see the issue. You’re ignoring the “regulating and monitoring” aspect of driving. An L2 system doesn’t just require a driver when the system doesn’t know what to do. An L2 system requires a driver all the time to double-check everything that the system is doing. Every single thing that an L2 system does is required to be, and supposed to be, confirmed by a human in real-time to make sure it’s safe - 100% of the time, all the time. The “twiddling” is to make sure that this essential and required part of the system is actually there and working.
L2 systems don’t just need a human when the car encounters a situation that it doesn’t know how to handle. It needs a human 100% of the time, even when the car is “certain” what it’s supposed to do.
An L4 system is fundamentally different, because it doesn’t require that. It doesn’t require a human to be doing this all the time.
2 Likes
The impact may change, but it is ultimately the same system.
Just to be pedantic, and because this is the internet:
Overview The First Automatic Airplane LandingThe first commercial passenger aircraft to make a fully automatic landing was a British European Airways Hawker Siddeley Trident 1C, on June 10, 1965. The flight, designated BE343, traveled from Paris to London and landed using Smiths Industries’ Autoland system, guided by the Instrument Landing System.
Taking off can be automated. Flying is already often automated. So yeah, I’d say “it’s possible”. But it looks like it’s “level 2” 
Flight BE343, operating between Paris and London, was the first instance of a commercial passenger flight landing automatically.
Interestingly, the airline chose not to tell the passengers they were executing an automated landing either before, during, or after. It became public somewhat later when the system designer announced it.
1 Like
Sorry, but no need to argue this point. The twiddling of the steering wheel requirement went away about a year ago. Tesla now uses the in cabin camera to see if the driver is (sort of) paying attention. You can drive with hands completely off the steering wheel 100% of the time. If you do close your eyes or look away from the road for too long (maybe 3-5 seconds) you have to twiddle the steering wheel when warned to prove you are still alive.
Rambling threads like this tend to keep repeating erroneous info, discussing out of date “facts.”
Mike
1 Like
Are they? I’ve heard these types of statements before in regards to Tesla. For a long time, a popular narrative was that Tesla was so far ahead in EV development and manufacturing that no other manufacturer would be able to catch up. Especially since Tesla had amazing margins and the legacy manufacturers had, well, car company margins. The saying went that legacy manufacturers were years behind (and they were) so race had already been won by Tesla.
Of course today, Tesla is losing sales and market share, and very importantly, margin. The story that Tesla won the EV race is going away. Similar story with Tesla’s driver assist (whatever you want to call it). Tesla had the best system (and they did) and could train it up with millions of miles of actual data and no other company could do anything like that. The saying went that legacy manufacturers were years behind (and they were) so race had already been won by Tesla.
But today, many driver assist systems are rated better than Tesla’s. The results vary on how the metrics are weighted, but today Tesla appears to be among the leaders, not the leader.
Which brings us to L4. Tesla doesn’t have L4 and other manufacturers are years ahead in this space. But Tesla has already won? Bold statement, for sure.
3 Likes
Ah, I just was taking it from Tamhas. Whatever form it takes, the L2 system has to make sure that there’s a human driver in the car doing double-checking everything that’s going on in real time - because that’s an essential component of the system. At all times. Which is why L4 systems are so very different from L2 systems. Indeed, that real-time human operator is the fundamental difference between L4 and L2/L3 systems.
2 Likes
ChatGPT says:
{. Most modern drones can land autonomously, but whether they do depends on the drone’s type, software, and how it’s being used. Here’s a breakdown:
Consumer & Prosumer Drones (like DJI)
Yes, they can land autonomously.
Features like auto-landing, return-to-home (RTH), and precision landing are standard.
These drones use GPS and visual sensors to safely land without human input.
Human pilots can override the process or land manually if needed.
Commercial & Enterprise Drones
Usually autonomous landing is supported and often preferred.
In missions like delivery, inspection, or mapping, autonomous takeoff and landing are routine.
Advanced versions use RTK-GPS, LiDAR, or visual markers to land accurately.
Military Drones
High-end military UAVs can land completely autonomously using complex systems.
Human operators may monitor or intervene, but full autonomous landings are common.
Hobbyist / DIY Drones
May or may not have autonomous landing depending on flight controller and sensors.
Manual landing is more common unless customized for automation.
Summary
Most capable drones today can land autonomously.
Human input is often optional, used for safety, precision, or regulation compliance. }
Plaaa-yuh
ralph
Oh. Right. ChatGPT hallucinates.
So I got a second opinion.
Grok says fully autonomous takeoffs n landing are { “nearly ubiquitous in mid-to-high-end consumer drones, with millions of such drones in use globally.”}
“Millions” seems … A lot?
A third opinion- Perplexity says:
Paraphrase: Completely autonomous landings and take offs are common.
All 3 opinions mentioned “depends regulatory permission”.
FWIW
That is sort of the problem with believing what you read on the Internet and not having real life experience with what you are discussing.
You can easily drive and not be double checking AT ALL TIMES…you can be staring out the windshield at a bird and it is OK with that. The legal thing you clicked can say anything. You can even look at the nav screen (for a while) and it is OK with that.
But you are in the seat and if, for some reason, the car decides you must take over a red steering wheel will appear with an audio alert that you must take over. This very rarely happens. For me it is usually after I have already taken over, such as when I enter a parking garage and must drive up close to an RFID scanner I need to wave my phone at before the gate goes up. Has never happened on open road driving, IIRC.
Mike
2 Likes
In the most recent versions, after the warning to pay attention to the road, you just have to properly look at the road again and the warning goes away after a few seconds. I got the warning once a few weeks ago and did not have to touch the steering wheel.
As far as fully automated flying goes, I needed my roof pressure washed a few months ago. The guy had a cute little drone thing that he used to take before and after photos/videos. That thing appeared to be fully autonomous. He took it out of the case, put it on his truck bed. Then pressed a single button, and it took off by itself, then flew up above my house by itself, and then took photos and videos of my roof (I think he must have had some sort of software where it uses the address to find the target house, similarly to the way insurance companies do for inspections, or local permitting offices do for detecting un-permitted projects, etc), and then it landed by itself right onto his truck bed where it began.
1 Like
For me sometimes that works and sometimes it doesn’t. Might have something to do with how quickly it can sense your eyes. I wear glasses – maybe it takes longer to lock on the proper eye gaze. Of course, while looking attentive and looking at the road I tend to glance at the screen as well to see that the blue warning goes away.
In any case I looked it up. Tesla changed this in May 2024. This shows us how long the mental inertia lasts for those non-FSDers.
Mike
Just bookmarking this here for future reference:
5 Likes
Yeah, replaced by the story that after winning, Tesla went after new things instead. The recent Tesla shareholder letter actually says:
the beginning of our transition from leading the electric vehicle and renewable energy industries to also becoming a leader in AI, robotics and related services.
In Walter Isaacson’s book, there’s quotes from Elon about not being excited by doing low cost human-driven EVs, and that he’s betting the entire company on autonomy and robotics. So, yeah, self-inflicted pain and bad transition management.
The race for autonomy is kind of like a triathlon in that it takes different skills and capabilities to do well in the different aspects. Someone may take an early lead in swimming, only to lose that when cycling. Or cycling to running. Tesla is taking a different approach than almost everyone else, so it’s hard to compare actual progress towards the end goal.
For instance, most robotaxi companies rely on both LiDAR and extensively detailed pre-mapping. That not only makes the vehicles themselves more expensive, it hampers their ability to make enough robotaxi capable vehicles to matter. Waymo has less than 2,000 vehicles today, and their plans are to make another 100 a month for the next 20 months. VW/Moia/Mobileye are similarly retro-fitting sensors into ID.Buzz vehicles for trials, but if that were to expand, how many could they build, and at what price? Sure, they could eventually solve that problem, but as we’re seeing from Waymo and its expansion plans, that takes years. Tesla won’t have that production problem.
Tesla’s approach has been to iterate on their software until the required interventions are far enough apart to be considered safe enough. They’ve made some hardware changes (front bumper camera, uppped CPU capabilities), but those have been relatively minor and that hardware is already in anything off the assembly line.
I understand why people may doubt that Tesla can achieve L4 at a good enough safety (which has not been officially defined) with its current hardware, or even with slight modifications. I personally view LiDAR and other sensors that others use as something that was needed for companies who approached L4 directly, not as an L2 upgrade. The co-CEO of Google admitted to that on a recent CNBC interview. When asked about not needing LiDAR, she not only said “not yet” (implying eventually), she also said that LiDAR was needed for their path to today. Tesla’s L2 approach removed that need - with its own set of downsides, of course.
BTW, Tesla was recently granted a robotaxi permit for Austin, well ahead of the new law going into effect next month.
3 Likes
Waymo and Mobileye might not have that production problem either. The end game is to incorporate their sensor suite into vehicles on the factory assembly line. The Hyundai Iconic 5 will be delivered from the factory to Waymo configured for installation. Right now, Waymo’s robotaxi operations don’t require nearly enough vehicles to justify factory built systems. Each city utilizes about 100-300 vehicles.
That said, earlier this year Waymo announced plans to double their robotaxi fleet by 2026, from 1500 to 3000. That doesn’t sound like a lot perhaps, but it is exponential. That rate doesn’t have to continue for very long before the number of vehicles becomes very large.
Clarification: Tesla recently received a ride hailing permit from the Texas Department of Licensing and Regulation (TDLR). This permit is required for AV riding hailing, but not specific to it. Uber and Lyft have the same permit.
Fully autonomous vehicles used for ride-hailing require obtaining a separate permit from the Texas Department of Motor Vehicles certifying the vehicle meets autonomy standards. Tesla has not yet applied for this permit. In short, it is incorrect to say Tesla has been granted a robotaxi permit.
I think the regulatory status gives us good indication of the state of AV technology. Tesla has not applied for AV ride hailing permits in any jurisdiction I’m aware of. The only rational explanation is Tesla doesn’t believe they can demonstrate their AV technology is safe.
Which brings us to the first post in this thread. Tesla’s claim was that they would have probably have autonomous ride-hailing in probably half the population of the US by the end of the year. But if they don’t start applying for permits by about (looks at watch) right now, they will have AV ride hailing in zero percent of the population by the end of the year.
Waymo’s approach might be hard to scale, but Tesla’s looks even harder.
3 Likes
Clarification: Elon said that they’d “technically be able to do it.” Subject to regulatory approvals. Here’s a fuller quote:
I think we will probably have autonomous ride-hailing in probably half the population of the US by the end of the year. That’s at least our goal, subject to regulatory approvals. I think we will technically be able to do it. Assuming we have regulatory approvals, it’s probably addressing half the population of the US by the end of the year.
The operative words there are “configure for.”
From the Press Release:
The Hyundai IONIQ 5 will be delivered to Waymo with specific autonomous-ready modifications like redundant hardware and power doors.
Sensors and such will stiff have to be added. I haven’t seen anything about body modifications to support the added sensors. And there’s no time period stated for this, just “over time.”
Hardware doesn’t scale like software. Right now Waymo is going to build about 100 iPace’s a month for 20 month with the assembly line they purchased from Magna. We know Tesla’s production capabilities are orders of magnitude above that.
1 Like
That depends a lot on whether one interprets “rate” as 2X or +1500 in a year or so.
1 Like
I read somewhere that they purchased all the remaining i-Pace vehicles (something around 1500 of them). Presumably they are all parked somewhere waiting to be retrofitted into Waymo vehicles.
1 Like
As @Smorgasbord1 said, HW and SW scale differently. Once one has been through a few permitting processes, one is going to have a packet that just needs a new cover sheet for the next application. For cars, one has to actually build them, which Tesla is doing in volume.
1 Like