Coming soon! Tractors!
Coming soon! Tractors!
Other urls found in this thread:
pastebin.com
youtube.com
twitter.com
I'm glad they're thinking about us farmers
I look forward to reading about google cars causing an accident because they mixed up traffic signs and cars.
Robotic firemen?
that and chimneys wtf
cute
I wish I didn't have to be a slave to google
Testing.
click click click -faaaaaade- click click -faaaaade- click [verify] PLEASE TRY AGAIN
>that the? no way
click click click -faaaaade- click click click -faaaaade- click click -faaaaaaaaaade- click -faaaaaaaaaaaade- click [verify]
(*connection error*)
>rages
doesn't work on my machine, thanks for the botnet tho
Will there soon be one for spotting traps?
>Will there soon be one for spotting traps?
Probably. It would undoubtedly increase traffic and show that google captcha is embracing diversity. The downside is, users would be deliberately missing at least one result in hopes that the follow-up "retry" captcha would be more traps.
epic
At this point, how could anyone possibly deny that Captcha exists to train the AI??
>pic very fukken related
it was always training the ai, even when it was just text. are you retarded
user was being sarcastic (I'm pretty sure), and I agree that in no way was the captcha ever used to train any AI.
>zoomers know what a tractor is
it's a genius age gate.
>firescapes and chimneys
I think Amazon is partnering with Captcha to identify future alternatives to front door drone delivery.
>be Google AI
>require tens of millions of people saying yes/no to billions of images to learn what a car looks like
>still don't know what a car looks like
>in no way was the captcha ever used to train any AI.
Either your understanding of what an AI is is wrong or you're very naive
Yeah, no. But it is. R.E.T.A.r.D
Ah ok. Well, I was always thinking, from the very beginning, that claiming the captcha was for the altruistic purpose of training an AI, was utter corporate bullshit. I always thought it was to build fingerprinting capabilities (order of choosing squares, how often mistakes are made, timing between choosing squares, how much can you blur the lines between a bus and a van and still get the user to click on a van even when they know it isn't a bus HOW MANY LIGHTS DO YOU SEE)
Let's face it ... if the captcha can tell you when you've made a mistake, then it already knows what the answer is, and doesn't need you to log anything for the research.
I'm pretty sure it's used for research in the same way the text ones were used to refine text-recognition software, and the quality of the images varies from bad to terrible for the same reason: With each downward adjustment in image quality, the performance of the software at recognizing what's in the image is compared to the statistical consensus of human perceivers of it. Exactly how algorithms use that data I haven't the slightest idea, but the data is certainly used to inform them.
If you were riding in this, would you want to collide with a tractor?
If I was riding that, I would want a bullet to collide with my head.
Google car-chan getting wrecked by massive black tractor
That's a prototype car the actual cars would look normal
I feel like that would lose to a pedestrian in a collision
No irony there if he tips the scale at Arch Linux user weight. Also imagine the sluggish acceleration if he's riding in it, the body-lean, the inevitable Robin Reliant rollover.
>implying those things can go where tractors are commonly seen
lel
They know some of the pics are right and some are wrong, if you click one that's considered wrong it will make you solve another. They can also be judging what pic is right/wrong by the amount of clicks it get, if you click a new picture it'll probably act like it's wrong and tell you to solve another, if you click pictures that gets clicks frequently, you will get rewarded, and so on.
Every car is a potential tractor
>bomb the stairs so people cannot escape from a flaming building
youtube.com
How we all feel