8 ms·

"I have had a completely autonomous waymo come to my location, pick me up, and take me to another location." - Waymo uses 3d mapping, limited geofencing, remote

"I have had a completely autonomous waymo come to my location, pick me up, and take me to another location." - Waymo uses 3d mapping, limited geofencing, remote operators and mobile roadside assistance teams because those cars are not even close to any type of autonomy. Those cars are "mice" in a well designed and designated (inch mapped) maze. The car without a driver in the driver's seat is like David Copperfield flying on the stage in a cheap magic show, in front of a few hundred people that paid $50 for the tickets - see https://youtu.be/qZS9maIq_Zc https://youtu.be/qZS9maIq_Zc

melony 4y ago
Does it matter? They are functional and safe enough for most sunbelt cities. We may not have FSD from day one but what we do have is leagues ahead of what's possible 5 years ago.
- fuckstick 4y ago
  Waymo in its current form in Arizona launched about 5 years ago.
- jocker12 4y ago
  What these failing companies are doing for almost 15 years now is 1 step forward, 3 steps back while promising they are 6 months away from the impossible. A.I. is only pattern recognition software statistical tool, that has zero capability of learning by itself from previous experience, and that shows you how any business designed around updating the constantly changing environmental data required to make the robots operate at a decent level, is prohibitively expensive.
Dylan16807 4y ago
Anyone can use 3d mapping and geofencing. That's not a disqualifier. As long as the remote operators and assistance teams are an order of magnitude smaller than putting a driver in every car, then it's close enough to autonomy to count as "closer" and to be useful.
- jocker12 4y ago
  "Anyone can use 3d mapping and geofencing" - that shows you their limitations and also doesn't qualify for "completely autonomous" standard. - Completely means anytime (regarding weather conditions or time of the day), anywhere (no geofencing) and completely adaptive behavior to the permanently and randomly driving conditions humans deal with while driving. Pattern recognition software alone (A.I.) would never be able to match human driving performances. "As long as the remote operators and assistance teams are an order of magnitude smaller than putting a driver in every car" - the entire gig is way to expensive and requires "time travel" level of scientific achievements, which is 100% fiction and 0% reality.
  - Dylan16807 4y ago
    > doesn't qualify for "completely autonomous" standard No, but it does qualify for "closer to reality today than they were 5 years ago" > Pattern recognition software alone (A.I.) would never be able to match human driving performances. That's okay. A trained human can do much better than necessary, and geofenced pattern recognition software doesn't have to be as good, especially because it should have better reaction times and braking force than a human. > "As long as the remote operators and assistance teams are an order of magnitude smaller than putting a driver in every car" - the entire gig is way to expensive and requires "time travel" level of scientific achievements, which is 100% fiction and 0% reality. Why? If you can run a fleet of 300 cars with 30 people, that's already enough to make tons of money once you get well-established. You don't need any scientific improvements for that, let alone the ones you're exaggerating.
    - jocker12 4y ago
      "No, but it does qualify for" - please check the statement my comment was responding to. The "1 step forward, 3 steps back" way the automation sector does R&D is not moving towards reality, is moving towards confusing the public to justify their pitch to eventual investors. "That's okay." - Maybe for you, but not for investors and for the market. "Why?" - It's unsustainable, requiring resources (provided at this point by naïve investors) that commercialization can't provide. Just look at the over $100 billion wasted on this hallucination with zero actual returns. Investors expect palpable returns, not promises and delays.

melony 4y ago

Does it matter? They are functional and safe enough for most sunbelt cities. We may not have FSD from day one but what we do have is leagues ahead of what's possible 5 years ago.

fuckstick 4y ago
Waymo in its current form in Arizona launched about 5 years ago.
jocker12 4y ago
What these failing companies are doing for almost 15 years now is 1 step forward, 3 steps back while promising they are 6 months away from the impossible. A.I. is only pattern recognition software statistical tool, that has zero capability of learning by itself from previous experience, and that shows you how any business designed around updating the constantly changing environmental data required to make the robots operate at a decent level, is prohibitively expensive.

Dylan16807 4y ago

Anyone can use 3d mapping and geofencing. That's not a disqualifier. As long as the remote operators and assistance teams are an order of magnitude smaller than putting a driver in every car, then it's close enough to autonomy to count as "closer" and to be useful.

jocker12 4y ago
"Anyone can use 3d mapping and geofencing" - that shows you their limitations and also doesn't qualify for "completely autonomous" standard. - Completely means anytime (regarding weather conditions or time of the day), anywhere (no geofencing) and completely adaptive behavior to the permanently and randomly driving conditions humans deal with while driving. Pattern recognition software alone (A.I.) would never be able to match human driving performances. "As long as the remote operators and assistance teams are an order of magnitude smaller than putting a driver in every car" - the entire gig is way to expensive and requires "time travel" level of scientific achievements, which is 100% fiction and 0% reality.
- Dylan16807 4y ago
  > doesn't qualify for "completely autonomous" standard No, but it does qualify for "closer to reality today than they were 5 years ago" > Pattern recognition software alone (A.I.) would never be able to match human driving performances. That's okay. A trained human can do much better than necessary, and geofenced pattern recognition software doesn't have to be as good, especially because it should have better reaction times and braking force than a human. > "As long as the remote operators and assistance teams are an order of magnitude smaller than putting a driver in every car" - the entire gig is way to expensive and requires "time travel" level of scientific achievements, which is 100% fiction and 0% reality. Why? If you can run a fleet of 300 cars with 30 people, that's already enough to make tons of money once you get well-established. You don't need any scientific improvements for that, let alone the ones you're exaggerating.
  - jocker12 4y ago
    "No, but it does qualify for" - please check the statement my comment was responding to. The "1 step forward, 3 steps back" way the automation sector does R&D is not moving towards reality, is moving towards confusing the public to justify their pitch to eventual investors. "That's okay." - Maybe for you, but not for investors and for the market. "Why?" - It's unsustainable, requiring resources (provided at this point by naïve investors) that commercialization can't provide. Just look at the over $100 billion wasted on this hallucination with zero actual returns. Investors expect palpable returns, not promises and delays.