Waymo doesn't serve any snowy locales yet. But sure, years and years ago mapping was worse than it is today? The mapping used today is working quite well in warm weather locales.
2,943 karma · joined October 15, 2007
Mine doesn't even last as a glorified spotify controller mostly chilling in idle.
> for $200 I’m not expecting edge compute
What _are_ you expecting then? The hardware doesn't let you replace your phone and it isn't more convenient either. Pixel phones respond to "OK Google" faster than you can press this thing's button.
The vision is just what everyone is working on -- connecting LLMs to tools. Wake me up when it can do half of what Google Assistant can already do for me. I'll add that Google Assistant is free, and fast.
https://x.com/xyz3va/status/1805689140639408277
https://x.com/xyz3va/status/1805985433156759781 (this is the OP)
https://x.com/xyz3va/status/1805993239368860069
Doesn't help that he was shilling crypto right before rabbit. What people love about the form factor has mostly to do with Teenage Engineering's quirky design. Somehow Jesse is a board member there, it seems.
In general I don't recommend trusting their software. If you like the hardware, you can flash AOSP and use google assistant -- it's way more capable and AOSP isn't harvesting your credentials. https://github.com/RabbitHoleEscapeR1/r1_escape
I have a vested interest in the success of AVs (worked in the industry for the better part of a decade) but the companies building them are far from perfect in terms of how they conduct themselves wrt safety. There needs to be some development in terms of the supporting tech.
The way I think about this is: planes came before the FAA and key technologies that we use to manage flights, such as radar and radio being used in conjunction. AVs are in their "just learned to fly" era.
Robots aren't humans. You need accurate depth perception to maneuver a robot precisely, and you need ground truth depth measurements to train learned depth perceivers as well as to understand their overall performance. Humans learn it by combining their other senses and integrating over very long time using very powerful compute hardware (brain). To date, robots learn it best when you just get the raw supervision signal directly using LiDAR.
> Walk from a bright area to a dark area and it'll take seconds for your eyes to adjust
You do realize cameras have the same issue, and that HDR isn't free / is very computationally intensive?
Waymo is geofenced, but within its geofence it requires zero human intervention. Tesla on the other hand is famous for mistaking the moon for a traffic light. Saying "Tesla has so many miles on YouTube" is hilarious because first of all there are channels with lots of Cruise & Waymo footage too, and more importantly it's not the # of miles that matters, but the # of non-trivial scenarios you can handle.
Depth sensing is again, estimated or using time of flight sensors which is pretty much short-range lidar. HDR is used already in AV perception, but still loses to your eyeballs in dynamic range and processing time.
Also, "cheating" is just a necessary step to build baseline metrics. You need ground truth.
It may very well be the case that cheating is needed to generate the training data necessary to stop cheating...someday.
No, the depth is estimated. It's not accurate, at least not in the way you need for L4.
> I never said anything to the contrary. Animals get hit all the time, not just because a human wasn't paying attention.
I was just clarifying what the bar is. The bar is that avoidable accidents need to be avoided. Nobody will get mad if a plane crashes due to unavoidable circumstances (freak accident where two engines go out due to bird strikes or something). People will stop flying in the plane when it becomes clear that the airline is not doing everything it can to avoid fatalities.
Your brain is _really really_ good at surmounting challenges including many that you did not mention. We don't know how to get close to this in terms of reliability when using cameras and ML alone. Cameras and ML alone can go very far, but every roboticist understands the problem of compounding errors and catastrophic failure. Every ML person understands how slow our learning loops are.
Consider that ML models used in the field have to get by with a fixed amount of power and ram. If you want to process time context of say 5 seconds, and with temporal context 10Hz and with resolution 1080p, how much data bandwidth are you looking at? Comparing what you see with your eyes with a series of 1080p photos, which is better? Up it to 4k: how long does it take to even run object detection and tracking with a limited temporal context?
Your brain is working with more temporal context, more world context, and has a much more robust active learning loop than the artificial systems we're composing today. It's really impressive what we can achieve, but to those who've worked on the problem it feels laughable to say you can solve it with just cameras and compute.
There are plenty of well respected researchers who think only data and active learning loops are the bottlenecks. In my experience they're focused on treating the self driving task as a canned research problem and not a robotics problem. There are as many if not more respected researchers who've worked on the self-driving problem and see deeper seated issues -- ones that cannot be surmounted without technologies like high fidelity sensors grounded in physics and HD maps.
Even if breadth of data is the problem and Tesla's approach is supposedly yielding more data -- there is also the question of the fidelity of said data (e.g. the distances and velocities from camera-only systems are estimated and have noiser gaussians than ones generated with LiDAR). If you make what you measure, and your measurements are noisy, how can you convince yourself or your loss function for that matter that it's doing a good job of learning?
It's relatively straightforward to build toy systems where subsystems have something on the order of 95% reliability. But robotics requires you to cut the tail much further. https://wheretheroadmapends.com/game-of-9s.html
For one, frame rate and processing rate on human eyes is way higher than cameras. Dynamic range is another. Also, Cruise and Waymo are some of the only companies that have hard internal data / ability to simulate how well their safety drivers do, and in the very same scenario what their software driver will do. Without LiDAR you can't build that simulation, and once you have that data if you continue to use HD Maps and LiDAR there's probably a good reason.
> Dogs jump into the road in front of cars all the time and get killed, and kids get endangered at school bus crossings. That's a reality of life that robotaxis do not need to solve.
Robotaxis need to avoid any accident that a human would be able to avoid.
> IR and thermal sensing are unnecessary if the bar is human level
See, you could say this if you had some data that showed that incidents per X miles (when the vehicle is driving at night) is sufficiently low, + if the software passes some contrived scenarios to gut-check its ability to see in the dark with the necessary reliability. But you don't have that data, do you? Someone has it though :) and I'd argue regulators should have it too.
Sadly, our regulatory agencies are currently set up for very delayed decisionmaking. Since every company can release software at any time, you could imagine a regulatory platform that tracks software releases and their corresponding safety statistics in real time.
If you can make the unit economics work for a large quantity of individual cars, mapping is a small fixed cost.
I agree that it's not economical to map every city and road in the US, since you need to generate revenue from every mapped road and city. So you can think of HD maps as amounting to building roads. They will be built in lucrative places. Cruise and Waymo won't make money from putting taxis in nowhere Arkansas, so they don't need to map it.
> the current limiting factor is not what the car sees or knows but what it does with that information. It is unclear how or why HD mapping would help them at that point.
That's simply untrue. All the hard stuff continues to be reliability and sensor gated. Cruise and Waymo have amazing sensors and even they struggle with sensor range, sensor reliability, model performance on tail cases, etc. For example, at night these cars typically do not have IR or Thermal sensing. They are relying on the limited dynamic range of their cameras + active illumination + hoping laser gets enough points / your object is reflective enough. Laser perception also hits limits when lasers shine on small objects (think: skinny railroad arm). Cars also have limits with regard to interpreting written signs, which is a big part of driving.
Occlusions are still public enemy #1. Waymo killed a dog. Cruise crashed into a fire truck coming out of a blind intersection even though their sensors saw the truck within 100ms.
LiDAR and HD mapping together are supremely useful, even if you don't drive with it, for enabling you to simulate accurately. You cannot simulate reliably while guessing at distances and locations. HD maps let you use visual odometry to localize, and distance measurements grounded in physics backstop the realism of your simulation at least in terms of the world's shape.
Tesla lacks the ability to resim counterfactuals with confidence since they don't have HD ground truth. There are believers at the company that maybe you could make "good enough" ground truth from imagery alone but that in and of itself is a huge risk, and it's what skipping steps looks like. Most in the industry agree that barring a major change in strategy they just have no way to regression test their software to the level of reliability required for L4 / no human supervision.
The safety story is an interesting one. Companies like Cruise and Waymo are not forthcoming with their incident data. They share infrequently and through spreadsheets that do not capture every incident. It's pretty ass, and I'd be wary of trusting their self-reported data. I imagine their insurance companies have slightly better data than the gov't, but even then maybe not.
Furthermore: tailgater collisions are a thing. You can't just stop on a dime, especially on fast roads (even just 45mph roads).
This. Chopping down a stop sign with an angle grinder is an easy terrorist feat but not one you hear about because nobody stands to gain anything.
OP: “preventing unwanted pregnancies is good for environment” (because humans have big carbon footprint)
You: “having fewer children is bad: see these economies”
This reads as criticism of birth control and the viability of population reduction as an environmental strategy. But:
1. Do unwanted children typically contribute positively to GDP? Isn’t reducing that terrible situation just a win on all fronts?
2. How is collapsing population in developed countries related to the OP’s suggestion of widespread birth control? Are they really that related?
3. How is having fewer kids bad for environment? Afaict it is unambiguously good for that.
Rabbit on the other hand...don't give them your passwords!
Saying this as someone who won't waste money on the Humane pin but ordered an R1 for fun. Wish Humane got the price point (and a lot of other things) right. Like focusing on device latency and integrating Spotify instead of trying to boil the ocean and integrating only Tidal.
Definitely not hooking it up to Spotify or any personal accounts after hearing how they handle security.