Artificial intelligence still has some way to go

will ferrell IS john the baptist

what was the human safety driver doing???

seems like the whole point of them being there in the first place is to slam on the manual breaks in case something goes wrong. were they not paying attention to the road?

― Hazy Maze Cave (Adam Bruneau), Monday, March 19, 2018 2:13 PM (three months ago) Bookmark Flag Post Permalink

https://arstechnica.com/cars/2018/06/police-uber-driver-was-streaming-hulu-just-before-fatal-self-driving-car-crash/

"This crash would not have occurred if Vasquez would have been monitoring the vehicle and roadway conditions and was not distracted,'' the report concludes.

Police obtained records from Hulu suggesting that Vasquez was watching "The Voice," a singing talent competition that airs on NBC, just before the crash. Hulu's records showed she began watching the program at 9:16pm. Streaming of the show ended at 9:59pm, which "coincides with the approximate time of the collision," according to the police report.

the whole concept of a safety driver is bullshit, sitting there completely passively "monitoring the vehicle and roadway conditions" while being 100% disengaged from the actual driving, you'd have to have superhuman powers of concentration.

yeah it's easy to clown these people but hard agree

i mean you're either driving or you're not??! even an automatic on cruise control needs full, continuous engagement

you're either driving or you're not

almost wrote these exact words!

I enjoyed this https://arxiv.org/abs/1806.10474v1 (it's currently the #1 hyped article on arxiv-sanity for the last day)

i saw sander speak ... 4 years ago when he was a (spotify?) intern who had just won a galaxy zoo kaggle competition. i am very skeptical about kaggle, but it's hard to argue with his career as an example of what it can do.

Yeah, I was skeptical about Kaggle too. However, I ran a competition a few months back (https://www.kaggle.com/c/data-science-bowl-2018) but it went okay and most of the winners were legit.

Well this guy's a fucking idiot.
https://www.theguardian.com/technology/2018/jul/07/artificial-intelligence-can-tell-your-sexuality-politics-surveillance-paul-lewis

I smell a fraud and a hack

The conclusion is hilarious. What a rube! Fucking Stanford, man

Had the same thoughts when I read that earlier. I know there is some small space given to critics therein, but the piece still seems distinctly lacking.

Phrenology definitely makes a lot more sense when it's a computer doing the analysis.
I'm working hard right now on building a neural network that can analyse people's romantic compatibility based on the balance of their humors.

As well as sexuality, he believes this technology could be used to detect emotions, IQ and even a predisposition to commit certain crimes. Kosinski has also used algorithms to distinguish between the faces of Republicans and Democrats, in an unpublished experiment he says was successful – although he admits the results can change “depending on whether I include beards or not”.

oh good it works only as long as nobody has facial hair

in the future if this kind of tech goes into mass use i may consider converting to Juggalo

I don’t believe in guilt, because I don’t believe in free will,” Kosinski tells me, explaining that a person’s thoughts and behaviour “are fully biological, because they originate in the biological computer that you have in your head”. On another occasion he tells me, “If you basically accept that we’re just computers, then computers are not guilty of crime

this is like first day in philosophy class galaxy brain stuff. he also says elsewhere not publishing his results would have been "morally wrong" so he's obviously great at thinking things through.

“Progress always makes people uncomfortable,” Kosinski adds. “Always has. Probably, when the first monkeys stopped hanging from the trees and started walking on the savannah, the monkeys in the trees were like, ‘This is outrageous! It makes us uncomfortable.’ It’s the same with any new technology.”

LOL the idea that it was some kind of BC outrage culture and not, you know, being on the ground vs. the trees now puts you in direct danger of being killed by a number of carnivorous beasts with that were previously out of reach.

The idea that pictures people choose to post on dating sites aren’t gonna be full of non-facial cues as to their sexual orientation is just bogglingly terrible science.

This isn’t even “AI has a ways to go” it’s “coders should not be allowed anywhere near the social sciences without a nanny”

scary shit

https://www.nytimes.com/2018/07/08/business/china-surveillance-technology.html

I don't really follow (am probably not capable of really following) recent NLP research very closely, but the stuff described here (as well as this paper, which he doesn't mention but seems related to me) seem like a really big deal! Is there some reason to be pessimistic here that I'm missing?

they named it Skynet, huh

xpost

Yeah I was reading that earlier, the idea has been around for ages but it just didn't work very well until recently. It's definitely progress but it alone doesn't change what you can do with language interfaces, it's a foundation to build on and we'll find out over time whether it is indeed as powerful as the hype suggests.

yeah my ongoing impression is that NLP and image recognition and deep training stuff is TOTALLY AMAZING until it fucks up and then everyone groans at how predictably stupid it is

never forget "What is Toronto?????"

otm we have come a long way from "How to wreck a nice beach"

https://www.theregister.co.uk/2018/07/18/pinscreen_fraud_claims/

ha

Machine learning systems designed to keep learning on the job present opportunities for mischief, as this funny example of poisoning the training data from hackernews shows: https://t.co/r9GHp6vYKF pic.twitter.com/FMijdQcJTD

— Reuben Binns (@RDBinns) July 29, 2018

According to new reporting by WalesOnline, police in South Wales scanned the faces of more than 44,000 people at the “Biggest Weekend” event in Swansea earlier this year—and of those, there were only 10 false positives.
That’s a significant improvement over a similar trial run at a 2017 soccer championship in Cardiff, in which 92 percent were incorrect matches.

The South Wales Police have attributed the improved matching to a “new algorithm” from its contractor, NEC.

"With each deployment of the technology we have gained confidence in the technology, and this has enabled the developers at NEC to integrate our findings into their technology updates,” WalesOnline quoted the SWP as saying.

During the Biggest Weekend event, the facial recognition system flagged a person who had an outstanding arrest warrant, and officers took the person into custody.

In the United States, facial recognition is in use—it was recently used to identify the Capital Gazette shooter in Maryland—by some law enforcement and also at some airports.

Just last month, the American Civil Liberties Union used Amazon’s Rekognition tool to show that it falsely identified 28 members of Congress as people who had been arrested.

https://arstechnica.com/tech-policy/2018/08/uk-cops-used-facial-recognition-at-show-found-someone-with-outstanding-warrant/

if you don't have anything to hide, you don't have anything to worry about

how is this shit legal, jfc

don't answer that

machine learning is going to start WW3

quote me if we survive

the good news is that the young men and women of the senate and house who have the ultimate oversight oversight of such things in the US are really big into technology, what with your email and AOL memberships. they are definitely keeping an eye on all of this and understand what it all means

SCOOP: With secret access to NYPD CCTV @IBM created software which tags people based on their skin tone + hair/clothing color. IBM gave NYPD access, then pitched them on a new AI product which identifies people on camera as "Black," "White," and "Asian":https://t.co/4HuhciGhJL pic.twitter.com/tRKEK4Xwtg

— George Joseph (@georgejoseph94) September 6, 2018

Yikes. Might need a "shitty artificial intelligence still has the potential to ruin people's lives" thread

feels like there's no one at the wheel. these stories come out as scoops and leaks and exposes months or years after the fact, and then regulators pretty much have nothing to say about them. i know there's a shitload of stuff happening right now in the political spectrum, but this deserves a "national debate" as much as anything else does right now. the status quo/default is going to be for law enforcement to adopt these kind of technologies across the board, as much as possible. if nothing is done to prevent that now, it will be much more difficult to roll it back in the future, because the standard line would be "why are you take away a technology that has proven to help law enforcement prevent crime?", or whatever

Totally agree. The frustration for me in following surveillance tech news is that often the complaint boils down to "it doesn't work" rather than "this should be illegal in the first place." So like facial recognition is criticized for misidentifying people as if a 100% accurate system would be unproblematic.

reading this now and its pretty scary

https://en.wikipedia.org/wiki/Life_3.0

It would be better I think if AI was limited to narrow AI and general AI was never developed but technology cannot be held back. It will be developed one day. I think any technological civilization will eventually develop any technology that can be developed.

My @nest doorbell automatically locks the front door when it sees a face it doesn't recognize. Today it didn't recognize me, so I went into the app to investigate and... pic.twitter.com/qcgE4Ii1pn

— B.J. May (@bjmay) September 17, 2018

Both funny, and prompting the question why would you have such a thing on your house?

Karl otm

Anyone see that bbc future technologies series from the last several years? Suggested the next war may be transhumans vs non modded peoples, like x men level stuff

the next war is probably going to look depressingly like Saudis bombing the shit out of Yemeni civilians far more than it will look like transhumans vs non modded peoples, and it will probably start less than a year from now. and another one will start the year after that.

The next war will have nothing to do with AI, and everything to do with natural idiocy

Yeah agree

combining this

https://www.youtube.com/watch?v=7Dg49wv2c_g

with this

https://www.youtube.com/watch?v=GuZGK7QolaE

equals a new circle of hell

see you all there

the use applications in the top video are so purposefully naive - if you team loses the big game, you can switch the jersey so that you're wearing the winning team's jersey!!!!

ooooooor, you could make the target a major political figure, make the major political figure say exactly what you want them to say in lifelike fashion, and send it to a bunch of people who still use aol accounts as their primary email addresses to confuse them!

NVIDIA's new vid2vid is the first open-source code that lets you fake anybody's face convincingly from one source video. prior "face2face" stuff was either cartoonish or proprietary. interesting times ahead... https://t.co/JsPVVa3xwa pic.twitter.com/AFhpeObd8N

— Gene Kogan (@genekogan) August 21, 2018

(btw i realize this stuff isn't AI, but it seems relevant)

(also, maybe facial recognition is not a very good long-term security strategy for unlocking phones)

I was recently wondering whether it was possible to truly articulate your research for the next ten years (a common question in grant applications) with just one example. I came up with this:

I’ve been thinking about the following problem: pic.twitter.com/P3hztqShZ3

— Allen Goodman (@0x00B1) October 3, 2018