Jump to content
  • Announcements


  • Content count

  • Joined

  • Last visited

Community Reputation


About Ghost_MH

  • Rank

  • Birthday 08/24/1982

Profile Information

  • Gender
  1. Waymo's car went to the severe other end of the spectrum and just stopped when it couldn't figure out what the person was doing. That is a lot safer than what Uber did here. Then again, that was at an intersection where the car was already stopped and wasn't in the middle of the night and in the middle of a highway with a 45MPH speed limit. Uber's cars clearly need a lot of work. I don't think anybody is arguing that, but this whole story is more about the car failing to predict someone behaving in an illogical manner than of the car recklessly barreling down the street without a care in the world.
  2. We already know Uber is behind Waymo in all of this. That's the reason Waymo can have driverless cars with no safety driver on public streets and Uber still needs someone behind the wheel monitoring everything. Still, how do you differentiate between someone walking up the side of the car and then stopping and someone that doesn't stop? We sometimes take for granted how easy it is to tell the difference, but that's a lot harder for an AI to accomplish. Uber's cars seem to have a history of following the letter of the law and finding themselves on the legally right side of accidents. It's the same place Waymo found themselves a couple of years or whenever it was back when one of their cars found itself in a showdown with a cyclist.
  3. What you missed here is what I think Legend was pointing toward. How does a car know tell the difference between a pedestrian walking on the side of the road and a pedestrian walking on the side of the road that plans to randomly walk into the car's path? I think we can all agree that it's probably not advisable to have the car slam on its brakes every time there's a nearby pedestrian walking on the side of the road or even on a sidewalk that isn't any higher than the street. From our perspective, you have the logic to think "Yeah, this person is going to throw themselves in front of me" and don't even think twice about the guy that's walking on the side of the road just trying to get to work because there's a mountain of snow on the sidewalk. You also don't think twice about the person that jumped into the street to walk around some crew fixing a street lamp versus the kid who looks like he's on the verge of running away from his parents and into the street. AI has a much harder time differentiating between these types of events. Uber's car doesn't seem sophisticated enough to take into account the direction the woman is driving, the time of the day, and the lack of a change in direction or speed. If all was working as intended, it likely assumed she was walking to the "edge of the sidewalk" and would then wait for her turn to cross. Rather, she just kept going and if the car doesn't have collision detection and leaves that to the driver, it would just keep going and never bother trying to apply the brake. These things are hard and people are random and thoughtlessly suicidal.
  4. I was coming here to post just that. Like I said earlier in this thread, Uber is really behind the competition when it comes to driverless cars. It really does seem that if their lidar was working properly, then there's a huge chance that Uber just hasn't programmed its cars to look for humans acting in erratic and illegal ways. It also brings up another good point. As these driverless cars become more the norm, where does that take us as far as culpability. Should cars shoot for always following the letter of the law and, thus never being at fault in these accidents, or should they be programmed to expect to skirt around laws here and there to perform in ways human drivers would expect other humans to drive?
  5. That doesn't really mean anything because there are so many variables are play here. The position and state of the Moon, the headlight pattern and strength of the car. Fog. So many different things. Sent from my SM-G960U using Tapatalk
  6. Consider Uber's program to be in alpha, I have no idea where they consider it, but for this conversation's sake, we'll call it in alpha. It's road ready, but with close supervision. This isn't like Waymo that already has fully autonomous cars on the road without safety drivers behind the wheel. That program we can consider in beta. They're closer to what you're thinking, but I don't know if they'd consider it production ready. We're not fully there yet, but we're on the way.
  7. What? These cars are in testing. I don't think Uber's work here as being in beta is even accurate as it's still too early even for that. Uber's end game here is obvious. They own a huge chunk of the taxi game and now they can drop their one liability which are their drivers. You understand that the endgame is to get to the point where these cars are safer than manually operated cars and that the attention of a driver is entirely unnecessary, right? The car here was a test of the tech, not an actual production vehicle.
  8. Exactly. I see a lot of people online complaining that in real life, it would have been easier to see her. Counter to that, I also know a TON of people and animals get hit by cars in exactly this manner every year, so I don't buy that. All this tells me is that Uber's cars are probably not as good or as far along as some of the competition. Lidar should have picked this woman up easily and it didn't. I get that the whole reason a safety driver is in the car is to pick up when systems malfunction, but that doesn't really help us in situations in which people in regular cars get into accidents all the damn time.
  9. Police have released footage, though they do cut out the moment of impact... Driver wasn't paying all that much attention, but I don't think it would have made much of a difference. A human in a regular car would have likely hit her as well. There's nothing obstructing the car's view of that woman, but she really does just pop out of the shadows.
  10. Smash Bros. for Switch Announced

    I was going to point out exactly this. In Goku's defense, Dragon Ball has spawned a massive videogame franchise as well. The series has pumped out games since the mid-80s in a way few other franchises can claim. Probably Spider-Man, but that may be it. Batman? Not as many, but those have been coming strong for a good thirty years. Sent from my SM-G960U using Tapatalk
  11. Anybody going to see Tomb Raider?

    Vikander is a great fit for the movie and she got ripped for this film. Insecure body shamers online are crazy to me. The older Tomb Raider movies were also horrible, but they did well in theaters. It's a shame they couldn't do this one right and with a $9m Friday, it's unlikely they'll get another chance. Sent from my SM-G960U using Tapatalk
  12. Smash Bros. for Switch Announced

    I would be down for a few not traditionally videogame characters being brought into the fold. There is absolutely no reason to believe that will actually happen, but it still would be really cool. Cover the whole gamut. Go with some American animation too. I'd take Finn from Adventure Time with Fionna as an alternate skin. Sent from my SM-G960U using Tapatalk
  13. Until you start wearing a mask because you're tired of all the bugs in your teeth. Sent from my SM-G960U using Tapatalk
  14. No, it's teleportation. Do you really want to deal with another airplane ride when you could just pop up wherever you need to go instantly? This is assuming the power works like Jumper. Also, why suffer through awkward social obligations if you can be in like Thailand or something. Sent from my SM-G960U using Tapatalk
  15. That's the thing about these very reasonable and moderate Republicans. They still vote with Trump 90+% of the time. https://projects.fivethirtyeight.com/congress-trump-score/ryan-a-costello/ Though, I guess 93.9% Trump is better than 100%. Sent from my SM-G930T using Tapatalk