I’m not an expert and if you are I’ll defer but I don’t buy that being first to market isn’t a huge lead. As I understand the AI system learns from real world experience so having tens of thousands of cars driving in every scenario all sharing data has to be a massive edge over a few test cars for the other companies.
Staying with some family while my mom is in hospital which has me watching Price is Right for the First time ever. I am amazed at the skill spread of the games. Some of the car games must have win percentages higher than 75% while others must be below 5%!
Of note is one where they give them 7 dollars and they have to guess all five numbers in the car and for every number off they give a dollar. To win they need to have one dollar at the end of the guessing.
Does anyone win this? Seems like huge odds to win this one. If I was better at math I’d calculate it.
I think it’s very clear that Tesla’s sensor suite is inferior to all of its competitors, but the tradeoff is that it has access to way more training data than anyone else can get their hands on, because it was commercially viable to outfit a bunch of production cars with a dozen cameras way before it would be plausible to do that with LIDAR sensors. This article says that Waymo logged 625k test miles in 2020, which seems like a pretty small number to teach a NN how to drive. In essence, Tesla signed up to take on a really hard CV problem, while everyone else has opted to use cheat mode for sensing and then figure out how to solve the policy problem without a billion miles of training data.
I would not be at all surprised to find out that neither approach is viable, but I would be mildly surprised if a non-Tesla competitor produces a commercially successful system without outfitting a bunch of production cars with LIDAR and a GPU to gather training data.
What’s stopping Tesla from adding lidar?
The price. Lidar can’t be produced cheap enough yet to put it in a commercially available car. Tesla made the decision they don’t want to bet on lidar getting cheap enough once it gets mass produced. There is also the chance that the company producing cheap lidar is the one that is going to make the most profit off self driving cars and not the companies that made the AI software.
That’s amazing. I feel like someone had to literally be insane to come up with that.
Tesla’s “full self-driving” feature has attempted to drive under a railroad crossing arm while a speeding train passes. It’s nearly driven head on into a concrete wall of a parking garage, attempted ill-advised left turns and sometimes swerved on highways around semi trucks, when there was no clear reason to do so.
But systemic faults like these are nothing to be concerned about because…a suicidal maniac might do the same things?
Lol Clovis
Man Jal I had an absolute shit of a day and it warms my heart to see that ludicrous criticism of Clovis. Thanks for that.
1 Like
The question I’ve always wanted to ask is how are so many beer snobs also piss drinkers. There are sooo many of them! At least a half dozen at every kegger I’ve ever been to. How the hell are these Venn circles overlapping so hard?
1 Like
I mean Tiger Woods drove 90 mph into a tree a month ago without touching his brakes. Just for my own fun to laugh at how much of an idiot you are I googled each of these and in seconds found dozens of examples of each involving human drivers.
Care to tell us all more about how PMP is actually PM and not a certification?
Raise your troll game. You are getting boring.
Ah, Lucky 7. That’s won a fair bit. Once you understand that the prices of the cars they bring out for that game aren’t exactly random, it becomes easier.
1 Like
The last three numbers seem pretty random no?
Interesting. I didn’t realize games were always tied to a certain dollar amount of prizes. I assumed they shifted around.
I’d be curious if someone has ever tracked win rate of each game and which is actually hardest. I might go on a Sunday morning Google fest…
Not really. The producers don’t want the game to be impossible, so the cars they put up will have prices that generally have at least a couple of numbers up the middle. If you guess 5 a lot, you’re likely to have a decent chance.
Why does Dr Pepper come in bottles?
Because his wife died.
4 Likes
That was a crappy game because the prizes you could win were so cheap. It would’ve been better if they added in a trip or a living room set or something.
It was pretty easy to win — keep your hand if you have three nines or better, otherwise pass it to the house. That strategy would win like 90% of the time.
Lucky 7 has been won about 30% of the time for the games they’ve actually tracked, which is close to what I would’ve guessed.
what