|
nor a celibate car
|
# ? Jan 2, 2021 20:17 |
|
|
# ? Jun 8, 2024 21:05 |
|
All of my cars are fuckin' cars
|
# ? Jan 2, 2021 20:19 |
|
they should just make a car that brakes when you push the brakes, op
|
# ? Jan 2, 2021 20:36 |
|
instead of one that breaks when you brake
|
# ? Jan 2, 2021 20:37 |
|
nice one, punchy
|
# ? Jan 2, 2021 20:40 |
|
Sagebrush posted:All of my cars are fuckin' cars do not have sex with your car, sagebrush. I can't believe we're having to go over this again
|
# ? Jan 2, 2021 21:18 |
Doc Block posted:hell, just speeding up to pass somebody on the highway is a common example of needing an extra burst of speed for safety when im in my infiniti(w/ skyline guts) i just need to nudge the gas and i’m gonna overtake you anyway without much effort unless you’re driving a hellcat demon. smooth and stress free with fountains of wayne on xm radio. when im in my ancient mx5 im gonna overtake you even though it has 75hp because it weighs 90lbs., less if i’ve just taken a dump. but it’s like the millennium falcon doing the kessel run as i dramatically shift up. so you really don’t need a bunch of power but you’ll be glad you have it when you do. being on the right side of power to weight ratio really helps the mazda. i was on the wrong side of it with a pt cruiser for a while and it suuuucked.
|
|
# ? Jan 2, 2021 21:52 |
|
Laslow posted:yeah and they inexplicably speed up while you’re in the process. I want to believe this is a real post, but it’s probably a cp from some car meme
|
# ? Jan 2, 2021 22:37 |
|
kitten emergency posted:hot take, i dont think they're gonna release a fuckin car "What, like the back of a Lomarf!?"
|
# ? Jan 2, 2021 23:31 |
The Management posted:I want to believe this is a real post, but it’s probably a cp from some car meme my wife ended up getting a hot pink hayabusa because of this phenomenon.
|
|
# ? Jan 2, 2021 23:59 |
|
i don't think it's the cars he's amazed by
|
# ? Jan 3, 2021 00:00 |
|
Laslow posted:my wife ended up getting a hot pink hayabusa because of this phenomenon. I’ll get my motorcycle back one day
|
# ? Jan 3, 2021 00:57 |
|
if you're worried about having power when you need it you should get a Tesla. no more response curve and torque or gears to make high speed avoidance a dangerous and unpredictable move in traditional ICE cars.
|
# ? Jan 3, 2021 02:56 |
|
do the exact opposite, imo. get a car with a two stroke engine.
|
# ? Jan 3, 2021 03:01 |
|
President Beep posted:do the exact opposite, imo. get a car with a two stroke engine. https://www.youtube.com/watch?v=dXVRdSbKDT8
|
# ? Jan 3, 2021 03:07 |
President Beep posted:I’ll get my motorcycle back one day
|
|
# ? Jan 3, 2021 03:45 |
|
please, neicie j is my father.
|
# ? Jan 3, 2021 03:52 |
|
hmmm i like two-stroke engine better than two-pump chump ; i'll insist that my wife's boyfriend call me that instead
|
# ? Jan 3, 2021 03:55 |
|
electric cars are so boring, they should make bespoke wood gas cars instead https://www.lowtechmagazine.com/2010/01/wood-gas-cars.html
|
# ? Jan 3, 2021 04:02 |
|
maybe a modern day ford nucleon, just to keep things interesting
|
# ? Jan 3, 2021 04:04 |
|
how likely are full self-driving cars in 5-10 years? I haven't looked at it in awhile and I have a family member who insists it's just around the corner and based on what I know about IoT, 𝕞𝕒𝕔𝕙𝕚𝕟𝕖 𝕝𝕖𝕒𝕣𝕟𝕚𝕟𝕘 and the tech industry in general I'm incredibly sceptical. also he thinks tesla is good
|
# ? Jan 3, 2021 10:32 |
|
crepeface posted:how likely are full self-driving cars in 5-10 years? I haven't looked at it in awhile and I have a family member who insists it's just around the corner and based on what I know about IoT, 𝕞𝕒𝕔𝕙𝕚𝕟𝕖 𝕝𝕖𝕒𝕣𝕟𝕚𝕟𝕘 and the tech industry in general I'm incredibly sceptical. the batteries of self driving cars will be charged by fusion power plants.
|
# ? Jan 3, 2021 10:34 |
|
crepeface posted:how likely are full self-driving cars in 5-10 years? I haven't looked at it in awhile and I have a family member who insists it's just around the corner and based on what I know about IoT, 𝕞𝕒𝕔𝕙𝕚𝕟𝕖 𝕝𝕖𝕒𝕣𝕟𝕚𝕟𝕘 and the tech industry in general I'm incredibly sceptical. "Full self driving" in the sense of a car that no longer needs driver's controls because it can go door to door without human input? That will not happen until we invent hard AI. In 5 to 10 years we might have NHTSA level 3 autonomy (get on the highway and zone out, the car handles everything until it gets confused, at which point it safely pulls over). I do not believe we will have level 4 (get on the highway and go to sleep, car can handle every normal situation until you exit the highway) within the decade. The current state of the art is level 2: the car can drive itself but requires full-time driver attention, as the car may need the driver to suddenly take over at any moment to avoid a collision. Tesla uses weasel words to claim their cars are level 3 or 4, depending on how much adderall elon has snorted, but by the NHTSA definition to meet those levels a car must safely and gracefully handle all foreseeable situations without expecting the driver to save it.
|
# ? Jan 3, 2021 10:48 |
|
just jam an orange into the steering wheel. bam. two levels at least.
|
# ? Jan 3, 2021 14:12 |
|
even if you discount tesla, and you should, because they're the absolute worst out of anyone who even pretends to have autonomous features, the state of the art is failing to make unprotected left hand turns in perfect conditions in no weather whatsoever on roads you have mm precision lidar maps of. no one is anywhere near actual autonomy. the serious players who have been dumping tens of millions into it for the last decade are still only capable of driving in entirely optimal scenarios where nothing unexpected happens and they have extensive mapping data, and even then interacting with actual traffic is questionable. this isn't a "more data" scenario that's going to be solved by just having cars drive circles around a phoenix suburb for another five years. by comparison, tesla's "autonomy" is a child's toy. it's cruise control and lane keeping connected to a gps and a couple webcams, being allowed to control a one and a half ton vehicle. this also is not something that's going to be solved by having idiots plow into highway gores and firetrucks for the next decade because their model has the intelligence of a fruit fly, and failing to accurately identify objects is only the beginning of their troubles.
|
# ? Jan 3, 2021 14:55 |
|
achieving full autonomy with current techniques is very much in the 'trying to go to the moon by climbing higher in the tree' realm, well demonstrated by a lot of the best people being well-funded trying to make it happen for a long time. a breakthrough is still possible, but i don't think anyone has any idea where to even look for it.
|
# ? Jan 3, 2021 15:09 |
|
the russians used a chauffeur
|
# ? Jan 3, 2021 16:38 |
|
Cybernetic Vermin posted:'trying to go to the moon by climbing higher in the tree' That's such a good metaphor for the current state of things.
|
# ? Jan 3, 2021 16:43 |
|
I think the trap people fall into is that computer vision and classifiers are pretty amazing now compared to 10-15 years ago. but it's not obvious that they probably aren't the tools we need to solve this problem.
|
# ? Jan 3, 2021 16:45 |
|
jre posted:I think the trap people fall into is that computer vision and classifiers are pretty amazing now compared to 10-15 years ago. it's pretty endemic to a lot of gee-whillikers tech. you can do some really cool poo poo in very specific and tailored situations with ML, but making it general purpose is many orders of magnitude more difficult. people read the specific and it gets generalized by disingenuous marketing people and here we are
|
# ? Jan 3, 2021 16:50 |
|
crepeface posted:how likely are full self-driving cars in 5-10 years? I haven't looked at it in awhile and I have a family member who insists it's just around the corner and based on what I know about IoT, 𝕞𝕒𝕔𝕙𝕚𝕟𝕖 𝕝𝕖𝕒𝕣𝕟𝕚𝕟𝕘 and the tech industry in general I'm incredibly sceptical. they were saying it was around the corner in 2012. the cars we have now are not significantly better than they were back then. i forget who but someone i think from google gave an interview on it essentially saying, 'the first 90-95% of the problem was difficult to solve but doable, that last 5% is basically impossible' and the project was scrapped. as sagebrush said- we won't have full autonomous driving until we have a revolution in artificial intelligence where a computer is able to infer the intent, empathize if you will, of other drivers. right now teslas don't even have object permanence
|
# ? Jan 3, 2021 18:24 |
|
jre posted:I think the trap people fall into is that computer vision and classifiers are pretty amazing now compared to 10-15 years ago. i'd say they're necessary tools, but they aren't the only necessary tools. driving is an inherently social activity that requires a lot of contextual knowledge to understand and anticipate the environment and drivers around you, and i don't know that we have the ability to quantify that and codify it into a ruleset for a machine everything we've done so far is hacky abstractions that can certainly look like driving in the right scenarios, the problem is that people forget that and think it means a computer can drive
|
# ? Jan 3, 2021 18:26 |
|
kitten emergency posted:it's pretty endemic to a lot of gee-whillikers tech. you can do some really cool poo poo in very specific and tailored situations with ML, but making it general purpose is many orders of magnitude more difficult. people read the specific and it gets generalized by disingenuous marketing people and here we are yeah exactly. when you're dealing with anything regarding ai there's a general rule of thumb you can follow: is it a narrowly defined problem where we already have a massive data set of the correct answer to train the machines on? if yes then it will probably work. is the problem complicated where the answer in any given situation could be ambiguous? will it require the ai to come up with novel solutions? will it encounter previously unencountered situations? then it's a grift or the people running the project have no idea what they're doing, probably both. the definition, in my opinion, of intelligence is the ability to encounter a novel situation and piece together two previously unrelated ideas in order to come up with a novel solution. currently such capabilities aren't even on the horizon. they aren't even theoretically possible. until that's a possibility then full autonomous driving will be a pipe dream
|
# ? Jan 3, 2021 18:29 |
|
we can have fully autonomous cars today in urban and suburban areas if we are willing to make the following changes: no human-driven cars are allowed on roads. the speed limit on non-highway roads is reduced to 30 mph. all cars are connected to a central traffic management system. chances of this happening are probably zero.
|
# ? Jan 3, 2021 18:38 |
|
The Management posted:we can have fully autonomous cars today in urban and suburban areas if we are willing to make the following changes: tbf this all sounds pretty good though
|
# ? Jan 3, 2021 18:42 |
|
modern futurism forgets the older ideals of plentiful mass transit allowing for more beautiful cities. we need fewer blade runners
|
# ? Jan 3, 2021 18:46 |
|
who framed roger rabbit is needed now more than ever.
|
# ? Jan 3, 2021 18:47 |
|
The Management posted:we can have fully autonomous cars today in urban and suburban areas if we are willing to make the following changes: And once again we're back to trains
|
# ? Jan 3, 2021 18:50 |
|
The Management posted:we can have fully autonomous cars today in urban and suburban areas if we are willing to make the following changes: i could see this happening in cordoned off downtown zones in highly regimented places like singapore. but pretty much nowhere else.
|
# ? Jan 3, 2021 19:08 |
|
|
# ? Jun 8, 2024 21:05 |
|
sure, but if you have that level of state control, you'd probably just build more public transit it's a problem that you will never need to solve assuming you are capable of solving it.
|
# ? Jan 3, 2021 20:53 |