Skip to content
Do you have adblock enabled?
 
If you can read this, either the style sheet didn't load or you have an older browser that doesn't support style sheets. Try clearing your browser cache and refreshing the page.

(CBS San Francisco)   The machines are learning... how to do three-point turns   (sanfrancisco.cbslocal.com) divider line
    More: Amusing, English-language films, SAN FRANCISCO, San Mateo County, California, My Chemical Romance, Grateful Dead, Automobile, Gerard Way, San Francisco International Airport  
•       •       •

3174 clicks; posted to Main » on 14 Oct 2021 at 12:17 AM (9 days ago)   |   Favorite    |   share:  Share on Twitter share via Email Share on Facebook



51 Comments     (+0 »)
View Voting Results: Smartest and Funniest


Oldest | « | 1 | 2 | » | Newest | Show all

 
2021-10-14 12:21:59 AM  
So if I come up with a way to round up, corral and break wild self driving cars do I get to keep them?

/no?
//grand larceny convictions?
///that doesn't sound very "law of the west" to me
 
2021-10-14 12:28:21 AM  
Could they at least be programmed to bring pizza?
 
2021-10-14 12:28:21 AM  
Fark user imageView Full Size
 
2021-10-14 1:11:48 AM  
Jared gets stuck in driverless car - Silicon Valley
Youtube -trd_f6j3eI
 
2021-10-14 1:18:00 AM  
It is universally regarded as a dumb idea, will eventually become universally accepted, and anyone who criticizes them will be regarded as an old fart.
 
2021-10-14 1:18:14 AM  
Well obviously self driving cars are never going to work, we should give up now.
 
xCh
2021-10-14 1:28:07 AM  
It seems like a self-driving car trap:
Fark user imageView Full Size
 
2021-10-14 1:28:14 AM  
Must be using Apple Maps. That's tried to have me drive down a bike path.
 
2021-10-14 1:28:39 AM  
To be fair, this ain't about AI or self driving cars per se. They don't decide to go this route of free will. This is about mapping software.
 
2021-10-14 1:32:42 AM  

wildcardjack: Must be using Apple Maps. That's tried to have me drive down a bike path.


Somehow, I doubt Google's self-driving car division is using Apple Maps.
 
2021-10-14 1:32:46 AM  
Fark user imageView Full Size
 
2021-10-14 2:02:56 AM  

wax_on: Well obviously self driving cars are never going to work, we should give up now.


It would be for the best
 
2021-10-14 2:12:02 AM  
I turned it off at "A little speculation here ..."

Why are you speculating? Talk to Waymo. At least make an attempt. And go with that attempt early in the story. If you get no reply or a "no comment", THEN you can speculate.

Christ, go back to journalism school.
 
2021-10-14 2:18:26 AM  

Danger Avoid Death: I turned it off at "A little speculation here ..."

Why are you speculating? Talk to Waymo. At least make an attempt. And go with that attempt early in the story. If you get no reply or a "no comment", THEN you can speculate.

Christ, go back to journalism school.


Yeah, because trillion dollar companies like Google are known for being easy to contact and forthright about their embarrassing failures.
 
2021-10-14 2:34:01 AM  
"I awoke to a strange hum and I thought there was a spacecraft outside my bedroom window."

She sounds...interesting.
 
2021-10-14 2:34:02 AM  
"I noticed it while I was sleeping," says Jennifer King. "I awoke to a strange hum and I thought there was a spacecraft outside my bedroom window."

Well, yeah, a spacecraft wold be my first thought too.
 
2021-10-14 2:35:28 AM  

Danger Avoid Death: I turned it off at "A little speculation here ..."

Why are you speculating? Talk to Waymo. At least make an attempt. And go with that attempt early in the story. If you get no reply or a "no comment", THEN you can speculate.

Christ, go back to journalism school.


You should have stuck with it for the very next sentence....
 
2021-10-14 2:42:06 AM  

Danger Avoid Death: I turned it off at "A little speculation here ..."

Why are you speculating? Talk to Waymo. At least make an attempt. And go with that attempt early in the story. If you get no reply or a "no comment", THEN you can speculate.

Christ, go back to journalism school.


they... literally did that.

"KPIX 5 reached out to WayMo who responded by asking for more information about the neighbors' concerns and the specific location of the circling cars. As of airtime, they had not yet provided any explanation. "
 
2021-10-14 3:17:05 AM  
Why do I suddenly want a Hot Wheels Waymo Jaguar?

/It's because it looks completely ridiculous and should cost about a dollar.
 
2021-10-14 4:11:52 AM  

fragMasterFlash: So if I come up with a way to round up, corral and break wild self driving cars do I get to keep them?

/no?
//grand larceny convictions?
///that doesn't sound very "law of the west" to me


Let them go . . . and if they come back to you they're yours.

//or some damn thing
/farkin' 1970s
 
2021-10-14 4:14:27 AM  

Enigmamf: Danger Avoid Death: I turned it off at "A little speculation here ..."

Why are you speculating? Talk to Waymo. At least make an attempt. And go with that attempt early in the story. If you get no reply or a "no comment", THEN you can speculate.

Christ, go back to journalism school.

Yeah, because trillion dollar companies like Google are known for being easy to contact and forthright about their embarrassing failures.


Oh, I didn't realise it wasn't easy. If it's not easy, a journalist definitely shouldn't bother doing his job!
 
2021-10-14 4:16:52 AM  

BafflerMeal: To be fair, this ain't about AI or self driving cars per se. They don't decide to go this route of free will. This is about mapping software.


Yup. The map they are following thinks that street goes somewhere. It doesn't.

In the SF market they are still using human drivers in the cars, but the drivers are letting the computer navigate because that's how it learns.
 
2021-10-14 6:32:12 AM  
I watched a semi driver take over 25 minutes to back into a dock yesterday. Probably the worst 37 point turn I've ever seen.
 
2021-10-14 6:39:10 AM  

wax_on: Well obviously self driving cars are never going to work, we should give up now.


Actually, yes.

/Spoken as a guy who actually writes AI software
 
2021-10-14 7:52:34 AM  

Evil Twin Skippy: wax_on: Well obviously self driving cars are never going to work, we should give up now.

Actually, yes.

/Spoken as a guy who actually writes AI software


Shhhh.  AI Car projects are paying for the next five to ten years of my life.
 
2021-10-14 8:02:58 AM  

Captain Shaky: that's how it learns.


Seems WAYMO cars are v....e...r....y slow learners...
 
2021-10-14 8:03:40 AM  

BafflerMeal: To be fair, this ain't about AI or self driving cars per se. They don't decide to go this route of free will. This is about mapping software.


Okay, but why are there not regular cars driving to the end of that dead end road every minute? Don't you think there is something the AI must lack (e.g., the "I"), that all the people driving in SF have?
 
2021-10-14 8:07:21 AM  

BafflerMeal: Evil Twin Skippy: wax_on: Well obviously self driving cars are never going to work, we should give up now.

Actually, yes.

/Spoken as a guy who actually writes AI software

Shhhh.  AI Car projects are paying for the next five to ten years of my life.


I worked on the survivability analysis for the Littoral Combat Ship.

So ... I feel ya.

(In my defense our analysis showed the problems they fleet was going to have with them. Which is why the firm is still in business.)
 
2021-10-14 8:17:20 AM  

Flowery Twats: Captain Shaky: that's how it learns.

Seems WAYMO cars are v....e...r....y slow learners...


Much like the starry eyed morons who think that we will EVER see self driving cars in any of our lifetimes.
 
2021-10-14 8:29:19 AM  

BafflerMeal: Evil Twin Skippy: wax_on: Well obviously self driving cars are never going to work, we should give up now.

Actually, yes.

/Spoken as a guy who actually writes AI software

Shhhh.  AI Car projects are paying for the next five to ten years of my life.


Are self driving cars not ready to go solo? Why are they not likely to work? Not trying to be snarky, just curious
 
2021-10-14 8:47:53 AM  

I just lurk here: BafflerMeal: Evil Twin Skippy: wax_on: Well obviously self driving cars are never going to work, we should give up now.

Actually, yes.

/Spoken as a guy who actually writes AI software

Shhhh.  AI Car projects are paying for the next five to ten years of my life.

Are self driving cars not ready to go solo? Why are they not likely to work? Not trying to be snarky, just curious


The big picture problem is that no computer architecture seems capable of solving big picture problems. The can be trained to recognize. They can be trained to restore a lost balance. They can be given complex networks to traverse. But as soon as they run across an exception that they can't handle they are useless at best, and a hazard at worst.

And that's after we solve all of the current implementation issues. Which are many. Some include:

* Lidar can't see far enough ahead to safety drive at highway speed. Or actually any speed past about 20 miles per hour.
* The wavelength of radar is too wide to resolve small vehicles. Or people.
* Visual systems have to be super optimized to work in real time. Part of that optimization limits their tracking to objects moving relative to the car. Things like a stopped car or road barrier blend into the background
* For as many cameras that are on a car with a visual system they would be classified as legally blind for their ability to actually resolve data in real time. To simplify the world to something the computer can process in real time, it goes beyond cartoony into the realm of 1970s video game.
* The only way driverless cars can legally operate under most state insurance codes is under the supervision of a licensed driver. Who is at the controls and legally responsible for the car's action. To allow the car to be fully driverless somebody needs to be on the hook for the inevitable damage that will result from an accident. And no manufacturer in the world is offering to indemnify themselves for the damage done by their products.

Also, don't believe the hype about driverless cars being somehow safer. They can only legally take control during the safest part of driving. And they still manage to have accidents. And while legalistically most are blamed on the other drivers, driverless cars are sure involved in a lot more accidents that normal cars for the miles driven.
 
2021-10-14 9:04:15 AM  
News readin' lady is very cute.
 
2021-10-14 9:19:11 AM  

Enigmamf: wildcardjack: Must be using Apple Maps. That's tried to have me drive down a bike path.

Somehow, I doubt Google's self-driving car division is using Apple Maps.


It must use that stupid Google map version that takes over your iOs screen and can't be manipulated. Even better than apple maps!
 
2021-10-14 9:21:54 AM  

waxbeans: wax_on: Well obviously self driving cars are never going to work, we should give up now.

It would be for the best


There's billions, maybe trillions in hype money riding on this.

We (I mean our humble leasers) probably just need to lower the standards (by a lot). Just look at the carnage on the streets today.
 
2021-10-14 9:41:48 AM  

Evil Twin Skippy: I just lurk here: BafflerMeal: Evil Twin Skippy: wax_on: Well obviously self driving cars are never going to work, we should give up now.

Actually, yes.

/Spoken as a guy who actually writes AI software

Shhhh.  AI Car projects are paying for the next five to ten years of my life.

Are self driving cars not ready to go solo? Why are they not likely to work? Not trying to be snarky, just curious

The big picture problem is that no computer architecture seems capable of solving big picture problems. The can be trained to recognize. They can be trained to restore a lost balance. They can be given complex networks to traverse. But as soon as they run across an exception that they can't handle they are useless at best, and a hazard at worst.

And that's after we solve all of the current implementation issues. Which are many. Some include:

* Lidar can't see far enough ahead to safety drive at highway speed. Or actually any speed past about 20 miles per hour.
* The wavelength of radar is too wide to resolve small vehicles. Or people.
* Visual systems have to be super optimized to work in real time. Part of that optimization limits their tracking to objects moving relative to the car. Things like a stopped car or road barrier blend into the background
* For as many cameras that are on a car with a visual system they would be classified as legally blind for their ability to actually resolve data in real time. To simplify the world to something the computer can process in real time, it goes beyond cartoony into the realm of 1970s video game.
* The only way driverless cars can legally operate under most state insurance codes is under the supervision of a licensed driver. Who is at the controls and legally responsible for the car's action. To allow the car to be fully driverless somebody needs to be on the hook for the inevitable damage that will result from an accident. And no manufacturer in the world is offering to indemnify themselves for the damage done by their products ...


Any yet driverless busses are on the roads and being used right now in the UK UK's first self-driving bus takes to Cambridge roads - Garagewire
 
2021-10-14 9:52:25 AM  

marsoft: Evil Twin Skippy: I just lurk here: BafflerMeal: Evil Twin Skippy: wax_on: Well obviously self driving cars are never going to work, we should give up now.

Actually, yes.

/Spoken as a guy who actually writes AI software

Shhhh.  AI Car projects are paying for the next five to ten years of my life.

Are self driving cars not ready to go solo? Why are they not likely to work? Not trying to be snarky, just curious

The big picture problem is that no computer architecture seems capable of solving big picture problems. The can be trained to recognize. They can be trained to restore a lost balance. They can be given complex networks to traverse. But as soon as they run across an exception that they can't handle they are useless at best, and a hazard at worst.

And that's after we solve all of the current implementation issues. Which are many. Some include:

* Lidar can't see far enough ahead to safety drive at highway speed. Or actually any speed past about 20 miles per hour.
* The wavelength of radar is too wide to resolve small vehicles. Or people.
* Visual systems have to be super optimized to work in real time. Part of that optimization limits their tracking to objects moving relative to the car. Things like a stopped car or road barrier blend into the background
* For as many cameras that are on a car with a visual system they would be classified as legally blind for their ability to actually resolve data in real time. To simplify the world to something the computer can process in real time, it goes beyond cartoony into the realm of 1970s video game.
* The only way driverless cars can legally operate under most state insurance codes is under the supervision of a licensed driver. Who is at the controls and legally responsible for the car's action. To allow the car to be fully driverless somebody needs to be on the hook for the inevitable damage that will result from an accident. And no manufacturer in the world is offering to indemnify themselves for the damage done
by their products ...

Any yet driverless busses are on the roads and being used right now in the UK UK's first self-driving bus takes to Cambridge roads - Garagewire


That was a three month trial, never went above 20mph, followed on a 1km closed loop, and had two safety engineers on it at all times.
 
2021-10-14 9:54:19 AM  

marsoft: Any yet driverless busses are on the roads and being used right now in the UK UK's first self-driving bus takes to Cambridge roads - Garagewire


Did you even read the article you posted? It is driverless bus. Singular. It is is in the future tense. It hadn't actually started operations as of June 1 when the article was written. And "On the roads around the University" is a pretty controlled environment.

Let's see how they are faring in October...

https://www.connectingcambridgeshire.​c​o.uk/smart-places/smart-cambridge/auto​nomous-vehicles/

Ok, there were three shuttles. Not just one. But it never carried any passengers. And they had two safety operators on board at all time. And they talked about how the trial was supposed to run in June. And then... nothing.
 
ZAZ [TotalFark]
2021-10-14 10:18:32 AM  
For added entertainment put a traffic cone in the street.
 
2021-10-14 10:26:15 AM  

Evil Twin Skippy: And while legalistically most are blamed on the other drivers, driverless cars are sure involved in a lot more accidents that normal cars for the miles driven.


Translation for this sentence: They actually obey speed limits, which shocks human drivers in some places where nobody obeys the speed limit, who drive 20 MPH above the speed limit around a blind corner, expecting everybody else to do the same (as all human drivers do in these sorts of areas), encounter an AI car doing the speed limit, and rear end it.  If you rear end somebody, it's almost always your fault-most human drivers drive too fast (as in this scenario) and drive too closely to stop if traffic is slower (as in this scenario).

As it is not legal to program driverless cars to go faster than the speed limit, the only practical cure for this is to raise the speed limit (although it is probably set what it is for a reason).

Note that, again, the driverless car is not at fault here.
 
2021-10-14 10:40:25 AM  

Evil Twin Skippy: wax_on: Well obviously self driving cars are never going to work, we should give up now.

Actually, yes.

/Spoken as a guy who actually writes AI software


Spoken like a guy who writes AI software that doesn't understand the exponential speed growth of computer power and upcoming technological advancements.
You really must suck at your job. You should go into farm work or construction. Seems a better fit for someone with no imagination.
 
2021-10-14 10:50:39 AM  

Evil Twin Skippy: I just lurk here: BafflerMeal: Evil Twin Skippy: wax_on: Well obviously self driving cars are never going to work, we should give up now.

Actually, yes.

/Spoken as a guy who actually writes AI software

Shhhh.  AI Car projects are paying for the next five to ten years of my life.

Are self driving cars not ready to go solo? Why are they not likely to work? Not trying to be snarky, just curious

The big picture problem is that no computer architecture seems capable of solving big picture problems. The can be trained to recognize. They can be trained to restore a lost balance. They can be given complex networks to traverse. But as soon as they run across an exception that they can't handle they are useless at best, and a hazard at worst.

And that's after we solve all of the current implementation issues. Which are many. Some include:

* Lidar can't see far enough ahead to safety drive at highway speed. Or actually any speed past about 20 miles per hour.
* The wavelength of radar is too wide to resolve small vehicles. Or people.
* Visual systems have to be super optimized to work in real time. Part of that optimization limits their tracking to objects moving relative to the car. Things like a stopped car or road barrier blend into the background
* For as many cameras that are on a car with a visual system they would be classified as legally blind for their ability to actually resolve data in real time. To simplify the world to something the computer can process in real time, it goes beyond cartoony into the realm of 1970s video game.
* The only way driverless cars can legally operate under most state insurance codes is under the supervision of a licensed driver. Who is at the controls and legally responsible for the car's action. To allow the car to be fully driverless somebody needs to be on the hook for the inevitable damage that will result from an accident. And no manufacturer in the world is offering to indemnify themselves for the damage done by their products.

Also, don't believe the hype about driverless cars being somehow safer. They can only legally take control during the safest part of driving. And they still manage to have accidents. And while legalistically most are blamed on the other drivers, driverless cars are sure involved in a lot more accidents that normal cars for the miles driven.


Thank you very much for the detailed response! I've always wondered how their near real time vision interpretation worked. I always assumed the car had banks of computers in order to process images so fast. I didn't realize they did it by just streamlining the algorithm so much.

My iota of knowledge about computer vision comes from photogrammetry and the Wikipedia articles I read on SIFTs.

The KITT aficionado in me has a sad now.
 
2021-10-14 11:05:02 AM  

Karne: Evil Twin Skippy: wax_on: Well obviously self driving cars are never going to work, we should give up now.

Actually, yes.

/Spoken as a guy who actually writes AI software

Spoken like a guy who writes AI software that doesn't understand the exponential speed growth of computer power and upcoming technological advancements.
You really must suck at your job. You should go into farm work or construction. Seems a better fit for someone with no imagination.


Boy/Girl/[Diminutive Genderless Term for a Whipper Snapper]

When I started programming, at the age of 9 on an IBM PCjr, I was storing programs on 360 kilobyte floppy disks that had to fit in 128 kilobytes or RAM. This was 1984. And I was lucky. My friends on the Commodore64 had half of that. It had a clock speed of 4.77 megahertz.

We had clock rates in the low Megahertz. There were no hardware floating point units. Because at that time floating point wasn't considered a core business function for a home-pc.

My first computer with hardware accelerated floating point wasn't until college. Drexel made us buy Macs, and I got one of the first models with the PowerPC chip. That computer had a hard drive with 250 megabytes. It had 8 megabytes of ram. It ran at a clock speed of 66 Megahertz.

The Macbook I am writing this post on is from 2015. It has 4 cores, runnng at 2.5 gigahertz. It has 16 gigabytes of RAM. It has an SSD with a capacity if 1 terabyte.

My brand new work laptop has a slightly fancier CPU with 8 cores running at 2.6 Gigahertz. The only reason I "upgraded" was that work has decided to go all Windows.

Point 1) The days of exponential performance growth are over. I run the same programs. They take the same amount of time to run. The 5 year old computer for all of it's cores isn't much of an improvement. And trust me when I say I have contacts in the industry, the focus has been on power consumption for over a decade.

Point 2) Now given what I've seen in 35+ years, if you don't think I can imagine an exponential improvement in hardware, memory, and software tools, I really don't know where to start with you
 
2021-10-14 11:17:00 AM  
And the reason the focus has been on power consumption is that they really have hit a literal wall with chip micronization. If they make the chips any smaller electrons start tunneling across traces. You also have the problem that single electrons behave very differently than a large population of electrons.

Multiple cores and caches and ram on chip don't solve the problem that the speed of the computer is limited by its slowest part. That used to the the hard drive. Most of the improvement in the past 10 years has been because of the widespread use of SSDs. Basically: if your application isn't really storing anything it's not really doing anything. Even the bitcoin farms are running across this problem.

Now if your program can run in RAM, and only in RAM, great. But as soon as you need to talk to a sensor, storage device, or network, your application is limited by the speed of that external factor.

Oh right... if your application is running in RAM, then the multiple cores get to argue about who is accessing the frontside bus. So... to really exploit multiple cores your application needs to fit into level 1 cache sized chunks.
 
2021-10-14 11:20:32 AM  
But that is just the word for a middle-aged fogey who writes large scale expert systems and naval simulations. Feel free to learn all of this yourself. The hard way.

(And if you want the authentic Evil Twin Skippy experience, spend 10 years as a Senior Network Engineer. Oh the shiat I can tell you about network routing and how utterly braindead certain internet protocols are... )
 
2021-10-14 11:41:49 AM  

Evil Twin Skippy: Karne: Evil Twin Skippy: wax_on: Well obviously self driving cars are never going to work, we should give up now.

Actually, yes.

/Spoken as a guy who actually writes AI software

Spoken like a guy who writes AI software that doesn't understand the exponential speed growth of computer power and upcoming technological advancements.
You really must suck at your job. You should go into farm work or construction. Seems a better fit for someone with no imagination.

Boy/Girl/[Diminutive Genderless Term for a Whipper Snapper]

When I started programming, at the age of 9 on an IBM PCjr, I was storing programs on 360 kilobyte floppy disks that had to fit in 128 kilobytes or RAM. This was 1984. And I was lucky. My friends on the Commodore64 had half of that. It had a clock speed of 4.77 megahertz.

We had clock rates in the low Megahertz. There were no hardware floating point units. Because at that time floating point wasn't considered a core business function for a home-pc.

My first computer with hardware accelerated floating point wasn't until college. Drexel made us buy Macs, and I got one of the first models with the PowerPC chip. That computer had a hard drive with 250 megabytes. It had 8 megabytes of ram. It ran at a clock speed of 66 Megahertz.

The Macbook I am writing this post on is from 2015. It has 4 cores, runnng at 2.5 gigahertz. It has 16 gigabytes of RAM. It has an SSD with a capacity if 1 terabyte.

My brand new work laptop has a slightly fancier CPU with 8 cores running at 2.6 Gigahertz. The only reason I "upgraded" was that work has decided to go all Windows.

Point 1) The days of exponential performance growth are over. I run the same programs. They take the same amount of time to run. The 5 year old computer for all of it's cores isn't much of an improvement. And trust me when I say I have contacts in the industry, the focus has been on power consumption for over a decade.

Point 2) Now given what I've seen in 35+ years, if you don't think I can imagine an exponential improvement in hardware, memory, and software tools, I really don't know where to start with you


I guess we will see how well your prediction of driverless cars pans out.
I'm sure there were unimaginative 'smart' people connected to every advancement in history.
As they say - 'Those that say that something can't be done should move out of the way for the people that know it can'.
 
2021-10-14 11:44:58 AM  

Karne: I guess we will see how well your prediction of driverless cars pans out.
I'm sure there were unimaginative 'smart' people connected to every advancement in history.
As they say - 'Those that say that something can't be done should move out of the way for the people that know it can'.


It's not a prediction. It's an expert opinion. Believe me, don't believe me, I really don't care. I just want to make sure you have all the facts.
 
2021-10-14 3:00:34 PM  
I'm fairly close to the depot these things dispatch in and out of, so they're running around my neighborhood and up/down my street all day.

Based on my observations the waymo vehicles are far better at safely navigating the streets in the neighborhood that 90% of the neighborhood residents.

The area I live in the residents make it a point to drive like complete lawless farkheads, so the waymo vehicles are a nice change of pace.
 
2021-10-14 3:20:48 PM  
I see. Waymo is too cheap to actually buy accurate maps.
 
2021-10-14 3:51:35 PM  

moike: I'm fairly close to the depot these things dispatch in and out of, so they're running around my neighborhood and up/down my street all day.

Based on my observations the waymo vehicles are far better at safely navigating the streets in the neighborhood that 90% of the neighborhood residents.

The area I live in the residents make it a point to drive like complete lawless farkheads, so the waymo vehicles are a nice change of pace.


Yup.  If the standard is "better than the average human driver", they already meet it, because the "average human" is a farking dumbass.
 
2021-10-14 7:39:38 PM  

moike: I'm fairly close to the depot these things dispatch in and out of, so they're running around my neighborhood and up/down my street all day.

Based on my observations the waymo vehicles are far better at safely navigating the streets in the neighborhood that 90% of the neighborhood residents.

The area I live in the residents make it a point to drive like complete lawless farkheads, so the waymo vehicles are a nice change of pace.


Also, they're nice cars that make a pleasant, futuristic sound. Once again, better than many of the ones that pass through my neighborhood. I'm looking at you, rednecks with the Harleys and modded pickups.
 
Displayed 50 of 51 comments


Oldest | « | 1 | 2 | » | Newest | Show all


View Voting Results: Smartest and Funniest

This thread is closed to new comments.

Continue Farking




On Twitter


  1. Links are submitted by members of the Fark community.

  2. When community members submit a link, they also write a custom headline for the story.

  3. Other Farkers comment on the links. This is the number of comments. Click here to read them.

  4. Click here to submit a link.