9 ms·

More people need to see your comment because there seems to be a belief that full self driving is just around the corner. However, IMO, it will never happen bec

More people need to see your comment because there seems to be a belief that full self driving is just around the corner. However, IMO, it will never happen because it needs a full artificial general intelligence. In your case, the new lane was clearly marked, which just shows how weak the system really is. Imagine if there was snow covering the lane markings. Or sunlight reflecting off of the road confuses the vision. People bought the self-driving hype because the first 80% was easy. They forgot that the next 20%, as usual, would be incredibly hard, if not impossible. Personally, I will never be okay riding in a car that is on autopilot. Heck, I'm not even okay with the amount of automation in airplane cockpits. [1] [2] [1]: https://www.nbcbayarea.com/news/local/faa-report-pilots-addicted-to-automation/2044684/ https://www.nbcbayarea.com/news/local/faa-report-pilots-addi... [2]: https://hbr.org/2017/09/the-tragic-crash-of-flight-af447-shows-the-unlikely-but-catastrophic-consequences-of-automation https://hbr.org/2017/09/the-tragic-crash-of-flight-af447-sho...

xnx 5y ago
Self driving is really hard, but don't confuse Tesla's failings for the state of the art. Waymo's Driver is both much more advanced and much more cautious.
Closi 5y ago
> However, IMO, it will never happen because it needs a full artificial general intelligence. I don’t think that’s true - but it certainly needs more than a couple of poor resolution cameras and some underpowered hardware powered by hype. You can see how poor the vision of these systems is by just watching a few Tesla autopilot videos and seeing cars flash into/out of recognition on the screen. The first actual self driving car won’t be the one that does it with bargain basement hardware prioritising low cost over efficacy and safety.
- henrikschroder 5y ago
  > I don’t think that’s true - but it certainly needs more than a couple of poor resolution cameras How would your imagined autopilot system handle traffic flaggers, construction workers with signs, or various people directing traffic with handsignals? How will your system see the difference between a police officer making hand signs that you have to obey, and some random person making hand signals? How will it solve turn taking at stop signals, and how will it solve people waving you on, thereby changing the turn-taking order? There are tons of driving scenarios that require you to look at other humans, figure out who they are, and figure out what they want. For humans, this is dead simple. For computers... Nope.
  - Closi 5y ago
    The answer is quite simple in all these cases - it doesn't. A self driving car shouldn't be operating on some mysterious black-box AI interpretation of the intention of (often ambiguous) hand signals from drivers and police officers. Any situation with hand-waving should be passed back to a human driver until the hand-waving can be eliminated with future laws and road rules. We aren't going to have full level-5 autonomy that's reliant on cameras interpreting people waggling their hands around. The solution is that you architect a full system to allow for self driving, which will inevitably include changing how people interact with cars.
    - henrikschroder 5y ago
      Right, so you're saying that actual reality is too hard for self-driving systems to deal with, so we should change reality in order to coddle the self-driving cars so that they can handle it. ...which reinforces the point me and the grandparent post was making: You can't solve self-driving without AGI.
      - Closi 5y ago
        We change reality all the time to adapt to new technology - this is why we interacted with the first computers via punch cards rather than microphones and cameras. Were we coddling computers? Kind of, but solutions have to be a combination of what is possible. I guess there is a sliding scale with self driving cars: * Low technology, high societal or infrastructure change (for example we could build a totally different road network and dig cables into it, not practical but possible with technology 20 years ago!) * Medium technology, medium societal or infrastructure change. * High technology, low societal or infrastructure change. * Very high technology, no/little societal or infrastructure change (eg AGI that can interpret hand signals). All of these are possible versions of self driving cars - there is no “one” solution. Having AGI and going for the last bullet point isn’t the only option. As an example, when the 1950s imagined robotic vacuum cleaning they pictured a standing robot that would use your existing appliances - so they might have said “we need to work out how to pick up and use a household vaccum cleaner in order to make a useful household robot, which requires AGI!”. We now know that a better solution is the roomba, and AGI isn’t required. I personally think we will inevitably need some level of societal change, legislation change and possibly even infrastructure change to make self driving cars happen sooner.
- creckic 5y ago
  > You can see how poor the vision of these systems is by just watching a few Tesla autopilot videos and seeing cars flash into/out of recognition on the screen. Autopilot is not FSD. Autopilot is using the old system, FSD is using an architectural rewrite.
dmm 5y ago
> However, IMO, it will never happen because it needs a full artificial general intelligence. I think it will happen but only by making many expensive, intrusive changes to our cities and roads to accommodate "self-driving" cars.
- nathanaldensr 5y ago
  This problem is, this was never the marketed vision. The implicit and sometimes explicit message everyone seemed to swallow was "it's like human driving, but better!" when it's actually not. It's just another set of fundamentally-flawed algorithms running on binary computers that only know about zeroes and ones.
- jonplackett 5y ago
  I think we still must be missing something very fundamental with the way we’re doing AI. When you look at nature and see a tiny bee or fly navigating perfectly around all sorts of obstacles, including the newspaper I’m trying to swat it with, using just that teeny tiny brain running of mili-watts of power with just compound eyes to guide it. And in 3D space not a nice flat surface. But we still can’t make a car not drive into a massive concrete barrier? Something is off. But that gives me hope that there’s a discovery waiting to be made that can fix it.
  - wpietri 5y ago
    We surely are missing a lot. But nature has picked easier problems with a fly. They are small enough to be very robust; flying into things is no problem. They're also cheap to build and incredibly numerous. Flies, well, drop like flies. If we had a similar tolerance for car crashes, existing levels of automation would be fine.
    - wearywanderer 5y ago
      Even human brains, with all their capabilities, get the job done for 20 watts (or something in that order of magnitude.) Who even knows how much power we'd need for equivalent results from the best specialized silicon we have? I think you might burn a hundred gigawatts and still not get comparable results. Something is clearly wrong with our approach. It should be possible to do much more with much less, but there is some piece of the puzzle we're missing.
masklinn 5y ago
> More people need to see your comment because there seems to be a belief that full self driving is just around the corner. However, IMO, it will never happen because it needs a full artificial general intelligence. People who were not convinced by autopilot shoving cars under trucks are not going to get convinced by one more anecdote.
- ghoward 5y ago
  Actually, you are right. Guess my naivete showed through.
  - ghoward 5y ago
    To the downvoters, I wasn't being sarcastic. GP _is_ right, and I was wrong.
wpietri 5y ago
Yup. Rodney Brooks, famed robot scientist and iRobot founder, says not to expect driverless taxis in cities earlier than 2035: https://rodneybrooks.com/my-dated-predictions/ https://rodneybrooks.com/my-dated-predictions/ I don't think we quite need AGI to make something like a commercial urban taxi service work. For any given place, we can substitute a very large amount of training data. And I think we can get there through an approach like Waymo, where the taxi's number one job is to ask for central human help if it's in the slightest confused or at risk of causing harm. But even that will be a very long slog. And for the supposed goal of "full self-driving", I agree that is basically AGI, because it requires levels of judgment that even humans struggle with. The other day I came across some Covid-related road changes and I had to reason about human intent as expressed in roads while exchanging human social signals with the people around me. Brooks suggests we instead may see people trying to simplify the problem by banning humans and/or human drivers from areas where we want automated driving. That's anthema to me, but given the history of building to favor cars vs other modes and given the vast amounts of money behind these efforts, it wouldn't shock me that it turns out that way.
- jfrunyon 5y ago
  We've already seen several instances of Waymo's approach not working. I agree that ultimately they need a significant amount of local training data. But I believe they would be better off having a human drive a car around to map things out, run their 'AI' on the results, and then manually review/correct lower confidence results. Or, better yet, make a database and have cities provide feeds of changes/permits/closures.
  - what_ever 5y ago
    > We've already seen several instances of Waymo's approach not working. Define - 1. Several. 2. Not working. Disc: Googler but I don't know if that's necessary since I don't work at Waymo.
    - p1esk 5y ago
      Here's one: https://www.theverge.com/2021/5/14/22436584/waymo-driverless-stuck-traffic-roadside-assistance-video https://www.theverge.com/2021/5/14/22436584/waymo-driverless...
simias 5y ago
There's also the very common argument (even here on HN) that it doesn't matter because "humans have plenty of accidents too". I think it's not a good rebuttal because self-driving cars will have to be orders of magnitude safer than humans if order to be successful. If I'm drunk and I decide to drive and I end up crashing into a pedestrian, then it's my own damn fault and I'll face the consequences. If a self-driving car bugs out and kills someone, who's responsible? All it'll take in a few dozen stories of self-driving cars diving into oncoming traffic like in the parent's story and that'll destroy any trust in the tech for the foreseeable future.
- ffggvv 5y ago
  theres actually no proof of the claim that its safer or even as safe as human driving. The only "data" is from tesla themselves. Which will hurtle you 200mph at a wall and then disengage autopilot 100ms before hitting the wall and not count it as an autopilot death.
  - arcticfox 5y ago
    > Which will hurtle you 200mph at a wall and then disengage autopilot 100ms before hitting the wall and not count it as an autopilot death. This is false: "To ensure our statistics are conservative, we count any crash in which Autopilot was deactivated within 5 seconds before a crash." https://www.tesla.com/VehicleSafetyReport https://www.tesla.com/VehicleSafetyReport It would be more helpful to focus on actually deceptive issues in how they report data, namely the conflation of easy highway miles (Autopilot enabled) with more dangerous city driving (Autopilot unavailable). For example their Q1 2021 report: "In the 1st quarter, we registered one accident for every 4.19 million miles driven in which drivers had Autopilot engaged. For those driving without Autopilot but with our active safety features, we registered one accident for every 2.05 million miles driven. For those driving without Autopilot and without our active safety features, we registered one accident for every 978 thousand miles driven. By comparison, NHTSA’s most recent data shows that in the United States there is an automobile crash every 484,000 miles."
    - 5y ago
gher-shyu3i 5y ago
> it needs a full artificial general intelligence. How do you define full AGI? On a side note, another approach would be to have sensors in the roads that would transmit boundary information. That might be too expensive and difficult to maintain though.
- ghoward 5y ago
  I can't, but like Justice Potter Stewart, "I know it when I see it." In fact, I believe that the difficulty of even _defining_ AGI is yet more evidence that we are not anywhere close to it.
tibbydudeza 5y ago
There will always be an edge case that could result in a crash leading to a loss of life - don't think 99% is good enough that is why we won't have aircraft flying on autopilot all the time.
- osmarks 5y ago
  Humans are hardly perfect drivers either.
  - tibbydudeza 5y ago
    NHTSA says the leading cause of accidents is driver distraction , speeding and drunk/drugs , would rather solve those problems which are far more attainable than an actual level 5 car.
systemvoltage 5y ago
Curiously - Are Humans are AGI benchmark? Or is there more upper room at AGI+ level intelligence? I just think that we're on a continuum from virus to Humans, Chimps being significantly lower on the intelligence scale compared to Humans. There is nothing that says continuum stops at Humans. So, I am curious if AGI is defined as something profound (Turing complete, beats humans at every imaginable task by a large margin or mathematically max intelligence?) or just someone who can converse in a Turing test with humans without giving it away or drive as good as humans.
- ghoward 5y ago
  I don't think humans are the AGI benchmark, but smart mammals probably are. What I think an AGI needs to have is the _ability to learn_. To me, that's what intelligence really is. For many people, it's having a lot of knowledge. However, that is merely lack of ignorance. The reason that the ability to learn is important for AGI is because the true thing we miss with self-driving cars is _adaptability_. When roads change, humans adapt. But a self-driving car does not, unless a human interferes and tells it, "You learned about cones today."
gmadsen 5y ago
It is disingenuous to equate general self driving with tesla autopilot. The idea the full AGI is required is silly and incorrect. Waymo does not have these issues at all. It couldn't happen either. You don't need AGI to have a detailed map, you don't need AGI to properly localize to a detailed map. these are problems Tesla is giving itself
- clouddrover 5y ago
  > Waymo does not have these issues at all. False: https://jalopnik.com/watch-a-waymo-av-get-freaked-out-by-traffic-cones-bloc-1846897182 https://jalopnik.com/watch-a-waymo-av-get-freaked-out-by-tra...
  - gmadsen 5y ago
    did you read the article? that is not related to the tesla issue at all. Furthermore, this article just described bad decision making with regards to traffic cones, however was not dangerous. It was aware of drivable surfaces and direction of traffic... because it has a map Do you see how these are different situations? 1. Telsa swerving into an oncoming lane 2. Waymo coming to a controlled stop in front of a traffic cone, then thinks it is blocked and needs help
    - clouddrover 5y ago
      > Waymo coming to a controlled stop in front of a traffic cone And then driving off again, and then blocking traffic, and then evading Waymo's roadside assistance, etc. Do you see how these aren't different situations?
      - gmadsen 5y ago
        again, nothing high danger. That is not a neural net, it could have easily been programmed to not try and maneuver around and just stay stopped, so sounds like a behavior that needs some additional improvement. It didn't lose sight of the cones. It was an issue with generating an appropriate path around the cones. Notice how it didn't hit the cones or anything else That is fundamentally different than trusting a neural net to tell you where the lanes are, where the objects are, and what kind of objects they are. you do not drive into an oncoming lane if you have a hd map. you do not drive into a concrete wall if you have lidar.
jfrunyon 5y ago
It doesn't actually need full general AI, but it does need to know and understand all of the roads in its "service area". IMHO the real solution will be building a database of roads (and lanes, speed limits, traffic lights, stop signs, etc). BTW, the average driver gets in a crash every 18 years. The best humans aren't perfect either, and a self-driving car really only needs to be better than the worst humans to be an improvement (although it will have to be much better to succeed).
- imoverclocked 5y ago
  > a self-driving car really only needs to be better than the worst humans to be an improvement I think this is false because varying abilities of people will use the autopilot, not just the worst. The autopilot needs to be better than average for an overall improvement. I think it needs to be much better than average to be compelling enough for most people to adopt it. Most people consider themselves to be good drivers, even with sufficient evidence to the contrary.
  - jfrunyon 5y ago
    > I think this is false because varying abilities of people will use the autopilot, not just the worst. That depends entirely on how effective it is and what incentives/disincentives there are (i.e. subsidies, taxes, insurance), which will itself depend on how effective it is.
- ghoward 5y ago
  But what happens when the road is suddenly closed for an accident? Or new construction appears? What will database-based self-driving cars do then?
  - jfrunyon 5y ago
    They would handle it the exact same way they handle any obstacle in front of them. Which is a lot easier than figuring out the speed limit, and the lane demarcations, and which lanes go where, and oh the stripes are covered by snow, and oh the stripes are worn away, and ... No one ever said that they could get all of their info from a database. But there is a significant amount of info which they could get from a database, which they might not be able to get at all otherwise.
nobodylikeme 5y ago
It won't happen until cars can talk to each other, think of how ants walk in a line and follow each other around an obstacle. Car A (driven in manual mode) tells Car B (FSD) that tells Car C (FSD) that tells... Car Z (FSD): "I stopped at a stop sign in stop & go traffic at Lat/Long xyz while I spotted a human in the middle of the road. This is marked on my maps data (updated overnight) as not being a normal stop. No new data was detected except the human standing and a human laying on the ground." Car B, C, D... all confirm the same until Z passes the area and sees no more humans (they've been moved). Car Z tells Car AA all good here, proceed as normally. I'm just rambling now, but it's truly how ants operate. Draw a line of chalk in the middle of their "line" or walking patch and watched them avoid it and find a way around it. The difference is all of these cars will feed a true neural net of public data of live road conditions, accidents, etc. A possible privacy nightmare, but these are different times.
dec0dedab0de 5y ago
I think self driving will catch on for special roads or lanes that are automatic only. Mostly on long highways that are mostly straight.
- ggreer 5y ago
  That's the reality that Tesla owners already live in. Last week I visited my family for my grandmother's birthday. Of the 700+ miles traveled, over 98% were on autopilot. I entered my destination and hit "Navigate on Autopilot". It drove from onramp to exit, obeying speed limits and switching lanes to get around slow vehicles. It even handled the interchanges.
zelon88 5y ago
I agree with most of your post, but I'd like to point out that as more and more traditional cars are replaced with robots the possibility opens up for creating communication protocols between highway infrastructure controllers and vehicles. Vehicles communicating with each other and their surrounding infrastructure. Imagine coordinated intersections that don't need traffic lights. The capability of the system would increase as more nodes are added. Kind of like the Geth from the Mass Effect series.
sneak 5y ago
> Personally, I will never be okay riding in a car that is on autopilot. FWIW, I'm not okay riding in a car that's driven by ~50% of nonprofessional human drivers; people are, in general, really bad at the task.
creckic 5y ago
> which just shows how weak the system really is The old system. Current Autopilot is using the old system. Not relevant to the new FSD beta.

xnx 5y ago

Self driving is really hard, but don't confuse Tesla's failings for the state of the art. Waymo's Driver is both much more advanced and much more cautious.

Closi 5y ago

> However, IMO, it will never happen because it needs a full artificial general intelligence. I don’t think that’s true - but it certainly needs more than a couple of poor resolution cameras and some underpowered hardware powered by hype. You can see how poor the vision of these systems is by just watching a few Tesla autopilot videos and seeing cars flash into/out of recognition on the screen. The first actual self driving car won’t be the one that does it with bargain basement hardware prioritising low cost over efficacy and safety.

henrikschroder 5y ago
> I don’t think that’s true - but it certainly needs more than a couple of poor resolution cameras How would your imagined autopilot system handle traffic flaggers, construction workers with signs, or various people directing traffic with handsignals? How will your system see the difference between a police officer making hand signs that you have to obey, and some random person making hand signals? How will it solve turn taking at stop signals, and how will it solve people waving you on, thereby changing the turn-taking order? There are tons of driving scenarios that require you to look at other humans, figure out who they are, and figure out what they want. For humans, this is dead simple. For computers... Nope.
- Closi 5y ago
  The answer is quite simple in all these cases - it doesn't. A self driving car shouldn't be operating on some mysterious black-box AI interpretation of the intention of (often ambiguous) hand signals from drivers and police officers. Any situation with hand-waving should be passed back to a human driver until the hand-waving can be eliminated with future laws and road rules. We aren't going to have full level-5 autonomy that's reliant on cameras interpreting people waggling their hands around. The solution is that you architect a full system to allow for self driving, which will inevitably include changing how people interact with cars.
  - henrikschroder 5y ago
    Right, so you're saying that actual reality is too hard for self-driving systems to deal with, so we should change reality in order to coddle the self-driving cars so that they can handle it. ...which reinforces the point me and the grandparent post was making: You can't solve self-driving without AGI.
    - Closi 5y ago
      We change reality all the time to adapt to new technology - this is why we interacted with the first computers via punch cards rather than microphones and cameras. Were we coddling computers? Kind of, but solutions have to be a combination of what is possible. I guess there is a sliding scale with self driving cars: * Low technology, high societal or infrastructure change (for example we could build a totally different road network and dig cables into it, not practical but possible with technology 20 years ago!) * Medium technology, medium societal or infrastructure change. * High technology, low societal or infrastructure change. * Very high technology, no/little societal or infrastructure change (eg AGI that can interpret hand signals). All of these are possible versions of self driving cars - there is no “one” solution. Having AGI and going for the last bullet point isn’t the only option. As an example, when the 1950s imagined robotic vacuum cleaning they pictured a standing robot that would use your existing appliances - so they might have said “we need to work out how to pick up and use a household vaccum cleaner in order to make a useful household robot, which requires AGI!”. We now know that a better solution is the roomba, and AGI isn’t required. I personally think we will inevitably need some level of societal change, legislation change and possibly even infrastructure change to make self driving cars happen sooner.
creckic 5y ago
> You can see how poor the vision of these systems is by just watching a few Tesla autopilot videos and seeing cars flash into/out of recognition on the screen. Autopilot is not FSD. Autopilot is using the old system, FSD is using an architectural rewrite.

dmm 5y ago

> However, IMO, it will never happen because it needs a full artificial general intelligence. I think it will happen but only by making many expensive, intrusive changes to our cities and roads to accommodate "self-driving" cars.

nathanaldensr 5y ago
This problem is, this was never the marketed vision. The implicit and sometimes explicit message everyone seemed to swallow was "it's like human driving, but better!" when it's actually not. It's just another set of fundamentally-flawed algorithms running on binary computers that only know about zeroes and ones.
jonplackett 5y ago
I think we still must be missing something very fundamental with the way we’re doing AI. When you look at nature and see a tiny bee or fly navigating perfectly around all sorts of obstacles, including the newspaper I’m trying to swat it with, using just that teeny tiny brain running of mili-watts of power with just compound eyes to guide it. And in 3D space not a nice flat surface. But we still can’t make a car not drive into a massive concrete barrier? Something is off. But that gives me hope that there’s a discovery waiting to be made that can fix it.
- wpietri 5y ago
  We surely are missing a lot. But nature has picked easier problems with a fly. They are small enough to be very robust; flying into things is no problem. They're also cheap to build and incredibly numerous. Flies, well, drop like flies. If we had a similar tolerance for car crashes, existing levels of automation would be fine.
  - wearywanderer 5y ago
    Even human brains, with all their capabilities, get the job done for 20 watts (or something in that order of magnitude.) Who even knows how much power we'd need for equivalent results from the best specialized silicon we have? I think you might burn a hundred gigawatts and still not get comparable results. Something is clearly wrong with our approach. It should be possible to do much more with much less, but there is some piece of the puzzle we're missing.

masklinn 5y ago

> More people need to see your comment because there seems to be a belief that full self driving is just around the corner. However, IMO, it will never happen because it needs a full artificial general intelligence. People who were not convinced by autopilot shoving cars under trucks are not going to get convinced by one more anecdote.

ghoward 5y ago
Actually, you are right. Guess my naivete showed through.
- ghoward 5y ago
  To the downvoters, I wasn't being sarcastic. GP _is_ right, and I was wrong.

wpietri 5y ago

Yup. Rodney Brooks, famed robot scientist and iRobot founder, says not to expect driverless taxis in cities earlier than 2035: https://rodneybrooks.com/my-dated-predictions/ https://rodneybrooks.com/my-dated-predictions/ I don't think we quite need AGI to make something like a commercial urban taxi service work. For any given place, we can substitute a very large amount of training data. And I think we can get there through an approach like Waymo, where the taxi's number one job is to ask for central human help if it's in the slightest confused or at risk of causing harm. But even that will be a very long slog. And for the supposed goal of "full self-driving", I agree that is basically AGI, because it requires levels of judgment that even humans struggle with. The other day I came across some Covid-related road changes and I had to reason about human intent as expressed in roads while exchanging human social signals with the people around me. Brooks suggests we instead may see people trying to simplify the problem by banning humans and/or human drivers from areas where we want automated driving. That's anthema to me, but given the history of building to favor cars vs other modes and given the vast amounts of money behind these efforts, it wouldn't shock me that it turns out that way.

jfrunyon 5y ago
We've already seen several instances of Waymo's approach not working. I agree that ultimately they need a significant amount of local training data. But I believe they would be better off having a human drive a car around to map things out, run their 'AI' on the results, and then manually review/correct lower confidence results. Or, better yet, make a database and have cities provide feeds of changes/permits/closures.
- what_ever 5y ago
  > We've already seen several instances of Waymo's approach not working. Define - 1. Several. 2. Not working. Disc: Googler but I don't know if that's necessary since I don't work at Waymo.
  - p1esk 5y ago
    Here's one: https://www.theverge.com/2021/5/14/22436584/waymo-driverless-stuck-traffic-roadside-assistance-video https://www.theverge.com/2021/5/14/22436584/waymo-driverless...

simias 5y ago

There's also the very common argument (even here on HN) that it doesn't matter because "humans have plenty of accidents too". I think it's not a good rebuttal because self-driving cars will have to be orders of magnitude safer than humans if order to be successful. If I'm drunk and I decide to drive and I end up crashing into a pedestrian, then it's my own damn fault and I'll face the consequences. If a self-driving car bugs out and kills someone, who's responsible? All it'll take in a few dozen stories of self-driving cars diving into oncoming traffic like in the parent's story and that'll destroy any trust in the tech for the foreseeable future.

ffggvv 5y ago
theres actually no proof of the claim that its safer or even as safe as human driving. The only "data" is from tesla themselves. Which will hurtle you 200mph at a wall and then disengage autopilot 100ms before hitting the wall and not count it as an autopilot death.
- arcticfox 5y ago
  > Which will hurtle you 200mph at a wall and then disengage autopilot 100ms before hitting the wall and not count it as an autopilot death. This is false: "To ensure our statistics are conservative, we count any crash in which Autopilot was deactivated within 5 seconds before a crash." https://www.tesla.com/VehicleSafetyReport https://www.tesla.com/VehicleSafetyReport It would be more helpful to focus on actually deceptive issues in how they report data, namely the conflation of easy highway miles (Autopilot enabled) with more dangerous city driving (Autopilot unavailable). For example their Q1 2021 report: "In the 1st quarter, we registered one accident for every 4.19 million miles driven in which drivers had Autopilot engaged. For those driving without Autopilot but with our active safety features, we registered one accident for every 2.05 million miles driven. For those driving without Autopilot and without our active safety features, we registered one accident for every 978 thousand miles driven. By comparison, NHTSA’s most recent data shows that in the United States there is an automobile crash every 484,000 miles."
  - 5y ago

gher-shyu3i 5y ago

> it needs a full artificial general intelligence. How do you define full AGI? On a side note, another approach would be to have sensors in the roads that would transmit boundary information. That might be too expensive and difficult to maintain though.

ghoward 5y ago
I can't, but like Justice Potter Stewart, "I know it when I see it." In fact, I believe that the difficulty of even _defining_ AGI is yet more evidence that we are not anywhere close to it.

tibbydudeza 5y ago

There will always be an edge case that could result in a crash leading to a loss of life - don't think 99% is good enough that is why we won't have aircraft flying on autopilot all the time.

osmarks 5y ago
Humans are hardly perfect drivers either.
- tibbydudeza 5y ago
  NHTSA says the leading cause of accidents is driver distraction , speeding and drunk/drugs , would rather solve those problems which are far more attainable than an actual level 5 car.

systemvoltage 5y ago

Curiously - Are Humans are AGI benchmark? Or is there more upper room at AGI+ level intelligence? I just think that we're on a continuum from virus to Humans, Chimps being significantly lower on the intelligence scale compared to Humans. There is nothing that says continuum stops at Humans. So, I am curious if AGI is defined as something profound (Turing complete, beats humans at every imaginable task by a large margin or mathematically max intelligence?) or just someone who can converse in a Turing test with humans without giving it away or drive as good as humans.

ghoward 5y ago
I don't think humans are the AGI benchmark, but smart mammals probably are. What I think an AGI needs to have is the _ability to learn_. To me, that's what intelligence really is. For many people, it's having a lot of knowledge. However, that is merely lack of ignorance. The reason that the ability to learn is important for AGI is because the true thing we miss with self-driving cars is _adaptability_. When roads change, humans adapt. But a self-driving car does not, unless a human interferes and tells it, "You learned about cones today."

gmadsen 5y ago

It is disingenuous to equate general self driving with tesla autopilot. The idea the full AGI is required is silly and incorrect. Waymo does not have these issues at all. It couldn't happen either. You don't need AGI to have a detailed map, you don't need AGI to properly localize to a detailed map. these are problems Tesla is giving itself

clouddrover 5y ago
> Waymo does not have these issues at all. False: https://jalopnik.com/watch-a-waymo-av-get-freaked-out-by-traffic-cones-bloc-1846897182 https://jalopnik.com/watch-a-waymo-av-get-freaked-out-by-tra...
- gmadsen 5y ago
  did you read the article? that is not related to the tesla issue at all. Furthermore, this article just described bad decision making with regards to traffic cones, however was not dangerous. It was aware of drivable surfaces and direction of traffic... because it has a map Do you see how these are different situations? 1. Telsa swerving into an oncoming lane 2. Waymo coming to a controlled stop in front of a traffic cone, then thinks it is blocked and needs help
  - clouddrover 5y ago
    > Waymo coming to a controlled stop in front of a traffic cone And then driving off again, and then blocking traffic, and then evading Waymo's roadside assistance, etc. Do you see how these aren't different situations?
    - gmadsen 5y ago
      again, nothing high danger. That is not a neural net, it could have easily been programmed to not try and maneuver around and just stay stopped, so sounds like a behavior that needs some additional improvement. It didn't lose sight of the cones. It was an issue with generating an appropriate path around the cones. Notice how it didn't hit the cones or anything else That is fundamentally different than trusting a neural net to tell you where the lanes are, where the objects are, and what kind of objects they are. you do not drive into an oncoming lane if you have a hd map. you do not drive into a concrete wall if you have lidar.

jfrunyon 5y ago

It doesn't actually need full general AI, but it does need to know and understand all of the roads in its "service area". IMHO the real solution will be building a database of roads (and lanes, speed limits, traffic lights, stop signs, etc). BTW, the average driver gets in a crash every 18 years. The best humans aren't perfect either, and a self-driving car really only needs to be better than the worst humans to be an improvement (although it will have to be much better to succeed).

imoverclocked 5y ago
> a self-driving car really only needs to be better than the worst humans to be an improvement I think this is false because varying abilities of people will use the autopilot, not just the worst. The autopilot needs to be better than average for an overall improvement. I think it needs to be much better than average to be compelling enough for most people to adopt it. Most people consider themselves to be good drivers, even with sufficient evidence to the contrary.
- jfrunyon 5y ago
  > I think this is false because varying abilities of people will use the autopilot, not just the worst. That depends entirely on how effective it is and what incentives/disincentives there are (i.e. subsidies, taxes, insurance), which will itself depend on how effective it is.
ghoward 5y ago
But what happens when the road is suddenly closed for an accident? Or new construction appears? What will database-based self-driving cars do then?
- jfrunyon 5y ago
  They would handle it the exact same way they handle any obstacle in front of them. Which is a lot easier than figuring out the speed limit, and the lane demarcations, and which lanes go where, and oh the stripes are covered by snow, and oh the stripes are worn away, and ... No one ever said that they could get all of their info from a database. But there is a significant amount of info which they could get from a database, which they might not be able to get at all otherwise.

nobodylikeme 5y ago

It won't happen until cars can talk to each other, think of how ants walk in a line and follow each other around an obstacle. Car A (driven in manual mode) tells Car B (FSD) that tells Car C (FSD) that tells... Car Z (FSD): "I stopped at a stop sign in stop & go traffic at Lat/Long xyz while I spotted a human in the middle of the road. This is marked on my maps data (updated overnight) as not being a normal stop. No new data was detected except the human standing and a human laying on the ground." Car B, C, D... all confirm the same until Z passes the area and sees no more humans (they've been moved). Car Z tells Car AA all good here, proceed as normally. I'm just rambling now, but it's truly how ants operate. Draw a line of chalk in the middle of their "line" or walking patch and watched them avoid it and find a way around it. The difference is all of these cars will feed a true neural net of public data of live road conditions, accidents, etc. A possible privacy nightmare, but these are different times.

dec0dedab0de 5y ago

I think self driving will catch on for special roads or lanes that are automatic only. Mostly on long highways that are mostly straight.

ggreer 5y ago
That's the reality that Tesla owners already live in. Last week I visited my family for my grandmother's birthday. Of the 700+ miles traveled, over 98% were on autopilot. I entered my destination and hit "Navigate on Autopilot". It drove from onramp to exit, obeying speed limits and switching lanes to get around slow vehicles. It even handled the interchanges.

zelon88 5y ago

I agree with most of your post, but I'd like to point out that as more and more traditional cars are replaced with robots the possibility opens up for creating communication protocols between highway infrastructure controllers and vehicles. Vehicles communicating with each other and their surrounding infrastructure. Imagine coordinated intersections that don't need traffic lights. The capability of the system would increase as more nodes are added. Kind of like the Geth from the Mass Effect series.

sneak 5y ago

> Personally, I will never be okay riding in a car that is on autopilot. FWIW, I'm not okay riding in a car that's driven by ~50% of nonprofessional human drivers; people are, in general, really bad at the task.

creckic 5y ago

> which just shows how weak the system really is The old system. Current Autopilot is using the old system. Not relevant to the new FSD beta.