Humans Need Not Apply

Rob Lister · Post by **Rob Lister** » Sat Mar 03, 2018 12:58 pm

Doctor X wrote:A computer simply has the ability to run every conceivable move and calculate the probabilities.

A human could do the same . . . if he had the patience . . . and could live a million years.

--J.D.

Not so. The computer too would have to live a million years. Brute force computer chess is a thing of the way-past and it was never really that good when constrained by time, as all tournament games are.

Today's AI players use strategies somewhat similar to those humans use. Minimax being just one.

sparks · Post by **sparks** » Sat Mar 03, 2018 9:54 pm

Computers do not learn. They simply process information based on a well defined set of instructions.

Learning would require an independent goal.

They do not have such. And it cannot be programmed for them by others.

We will get very much better at a simulation of intelligence, but our creations will never ... think for themselves.

Witness · Post by **Witness** » Sun Mar 04, 2018 5:57 am

sparks wrote:Computers do not learn. They simply process information based on a well defined set of instructions.

Learning would require an independent goal.

They do not have such. And it cannot be programmed for them by others.

We will get very much better at a simulation of intelligence, but our creations will never ... think for themselves.

You're in deep waters there, sparky.

First let me state that computers/AI/robots don't have to "think" (whatever that means), they just have – and if I remember well, that was one of the points of the vid in the OP – to be good enough at their tasks. Driving cars better than the average Joe. Mixing or writing music better than the average Joe. Playing chess, or go, better than the average Joe (or even Grand Masters

). Prattling small talk better than the average Joe (and that's the Turing test in a nutshell).

Some here addressed the point, so perhaps you're now convinced that computers can effectively learn. The "set of instructions" is of course "well defined", but you are conflating levels: the fixed part is just like an OS for the AI (or the blood/chemicals allowing our brains to work), inside of that there are parameters (lots of parameters) or even code (for genetic algorithms) which get fine-tuned, for the task at hand, essentially by trial and error. Like good old nature (what a bitch), just faster.

As for the goal, I should hope that we set that. I think a good example would be the annual DARPA contest, where vehicles have to travel as fast as possible (and without crashing, of course) from some point A to some point B through rough terrain. The goal is clear, the zillions of local decisions are up to the machine.

Your last point, about the "simulation of intelligence", seems philosophical to me as I can see no clear border between unintelligent/intelligent. (That's the strength of the Turing test as it dismisses "internalities", consciousness, intelligence even, &c.) To put it slightly unpleasantly: if your lifelong partner simulates love unerringly, won't you accept it as the real thing?

sparks · Post by **sparks** » Sun Mar 04, 2018 4:22 pm

"Good enough" at their tasks says we adequately programmed them, nothing more. They didn't suddenly start seeking goals they invented for themselves. When they do that, I'll concede they're intelligent and can learn after the fashion that we do.

Lot's of peeps can be fooled by a computer that meets the Turing test. That simply means the humans are fucking stupid, not that the machines are smart.

We would love nothing better than to create for ourselves an intellectually equal or superior partner so's we know, at last, that we're not alone. But, alas, the hardware isn't ready for prime time just yet.

sparks · Post by **sparks** » Sun Mar 04, 2018 4:39 pm

Perfect. I'd like 5000 sq ft to my order, and I don't want to spend more than 87 cents US.

Umkay?

BTW: "Soon the robots will built many houses." ?? Only if/when they can properly parse past and future my friend.

sparks · Post by **sparks** » Sun Mar 04, 2018 5:37 pm

I know I don't need to remind you that RUR is a work of fiction. And as such, represents our ability to imagine (change our own programming) not the computers ability to imagine (which they don't have because we haven't been able to figure out the code yet) and which, I dare say, we will never accomplish.

sparks · Post by **sparks** » Sun Mar 04, 2018 6:04 pm

Ask any surviving drug addict.

Witness · Post by **Witness** » Mon Mar 05, 2018 3:45 am

sparks wrote:"Good enough" at their tasks says we adequately programmed them, nothing more. They didn't suddenly start seeking goals they invented for themselves. When they do that, I'll concede they're intelligent and can learn after the fashion that we do.

Lot's of peeps can be fooled by a computer that meets the Turing test. That simply means the humans are fucking stupid, not that the machines are smart.

We would love nothing better than to create for ourselves an intellectually equal or superior partner so's we know, at last, that we're not alone. But, alas, the hardware isn't ready for prime time just yet.

2. More or less agree; still I'm of the "good enough is OK" persuasion.

3. Heartily agree.

Now to point 1. Here's an example of something which hasn't been programmed (by us, anyway), but which has goals – or so it seems to me at least, even if it has no nervous system:

What's your take on that kind of robot?

sparks · Post by **sparks** » Mon Mar 05, 2018 4:44 am

Fascinating. Never underestimate 3.5 billion years of evolution.

Mentat · Post by **Mentat** » Mon Mar 05, 2018 6:10 pm

Burger-flipping robot begins first shift

Flippy, a burger-flipping robot, has begun work at a restaurant in Pasadena, Los Angeles.

It is the first of dozens of locations for the system, which is destined to replace human fast-food workers.

sparks · Post by **sparks** » Mon Mar 05, 2018 11:52 pm

One of Gawd's bad hair days perhaps?

Rob Lister · Post by **Rob Lister** » Tue Mar 06, 2018 12:43 pm

Mentat wrote:Burger-flipping robot begins first shift

Flippy, a burger-flipping robot, has begun work at a restaurant in Pasadena, Los Angeles.

It is the first of dozens of locations for the system, which is destined to replace human fast-food workers.

That's got to be the dumbest implementation of robotics ever. Besides, Hardee's perfected the burger machine 50 years ago.

Witness · Post by **Witness** » Wed Mar 07, 2018 3:21 am

sparks · Post by **sparks** » Wed Mar 07, 2018 4:02 am

Great lookin' legs.

For a fuckin' fembot.

Anaxagoras · Post by **Anaxagoras** » Wed Mar 07, 2018 10:35 am

The latter. Probably should have gone in Rob's bionics thread.

sparks · Post by **sparks** » Thu Mar 08, 2018 1:33 am

Rob's 'ad 'er.

WTF do you think happened to her legs?

'Fess up Listy...

Witness · Post by **Witness** » Fri Mar 09, 2018 9:08 pm

Rob Lister wrote:
Mentat wrote:Burger-flipping robot begins first shift

Flippy, a burger-flipping robot, has begun work at a restaurant in Pasadena, Los Angeles.

It is the first of dozens of locations for the system, which is destined to replace human fast-food workers.
That's got to be the dumbest implementation of robotics ever. Besides, Hardee's perfected the burger machine 50 years ago.

Flippy the burger-flipping robot that started work this week in a California restaurant has been forced to take a break because it was too slow.

The robot was installed at a Cali Burger outlet in Pasadena and replaced human cooks.

But after just one day at work the robot has been taken offline so it can be upgraded to work faster.

Its human helpers are also getting extra training to help the robot keep up with demand.

http://www.bbc.com/news/technology-43343956

Slavery for everybody, cool.

Witness · Post by **Witness** » Fri Mar 16, 2018 5:05 am

Witness · Post by **Witness** » Tue Apr 03, 2018 4:49 pm

Who knows, perhaps it really is

NSFW:

Rob Lister · Post by **Rob Lister** » Thu Apr 05, 2018 2:11 pm

ed · Post by ed » Thu Apr 05, 2018 2:16 pm

Mentat wrote:Burger-flipping robot begins first shift

Flippy, a burger-flipping robot, has begun work at a restaurant in Pasadena, Los Angeles.

It is the first of dozens of locations for the system, which is destined to replace human fast-food workers.

And lets be sure to implement a $15 minimum wage.

Mentat · Post by **Mentat** » Thu Apr 05, 2018 11:29 pm

ed wrote:
Mentat wrote:Burger-flipping robot begins first shift

Flippy, a burger-flipping robot, has begun work at a restaurant in Pasadena, Los Angeles.

It is the first of dozens of locations for the system, which is destined to replace human fast-food workers.
And lets be sure to implement a $15 minimum wage.

I constantly get shit on here for being against minimum wage by you lot. Usually promptly followed by being called a troll the two faced lying asshole who claimed to not call me a troll. So I don't really care to hear about it.

sparks · Post by **sparks** » Fri Apr 06, 2018 1:47 am

Automation eventually must reduce us all to street urchins. Unless, of course, you're the one building the automation.

Then you have a few more years before life in a tent becomes the norm.

Witness · Post by **Witness** » Sat Apr 07, 2018 2:29 am

The Daily Star wrote:Sick sex robot fantasy BANNED from world’s first cyborg brothel

If you really want to click: https://www.dailystar.co.uk/news/latest ... ed-fantasy

Grammatron · Post by **Grammatron** » Wed Apr 11, 2018 10:46 pm

Grammatron · Post by **Grammatron** » Tue May 08, 2018 6:25 pm

[gfycat]GrimyQuickHammerheadshark[/gfycat]

Doctor X · Post by **Doctor X** » Fri May 11, 2018 11:58 am

So it is a x[X--Ed.]erox. . . .

--J.D.

ed · Post by ed » Fri May 11, 2018 12:02 pm

Abdul Alhazred wrote:Food Lion to deploy robot associates to 171 stores
The Produce News

...

These friendly robotic associates have been roaming the aisles of several Giant/Martin’s stores since last year, and now Food Lion is also putting the concept to the test. Marty maneuvers around using advanced artificial intelligence and the same laser imaging technology found in self-driving vehicles. He can perform checks that help the store run more smoothly and safely for both customers and associates.

Marty clean sweeps the store 12 times each day to identify slip and fall hazards on the floor and alert associates. When he finds a severe hazard, he pages associates and positions himself by the spill, notifying customers to be cautious. Marty can also scan the shelves for out-of-stocks and to check that shelf pricing is aligned with the front end register system.

...

I think we should go for a $25 minimum wage.

Grammatron · Post by **Grammatron** » Fri May 11, 2018 11:22 pm

[gfycat]UniformAdmiredHydra[/gfycat]

Rob Lister · Post by **Rob Lister** » Fri May 11, 2018 11:30 pm

I know they're very loud and not likely to sneak up on you but if you arm that thing with a few nasty looking weapons and multiply it by a few dozen there's no enemy in the world that wouldn't retreat.

Is that Boston Dynamics?

Anaxagoras · Post by **Anaxagoras** » Sat May 12, 2018 1:02 am

Wasn't Boston Dynamics bought out by Google?

They were, but then google sold it to Softbank (Japanese mobile phone company)

https://www.engadget.com/2017/06/08/sof ... ts-schaft/

Could be that because Boston Dynamics makes robots with potential military applications, it didn't fit with Google's brand image.

Witness · Post by **Witness** » Sat May 12, 2018 1:16 am

BBC wrote:Towards a lip-reading computer

Scientists at Oxford say they've invented an artificial intelligence system that can lip-read better than humans.

The system, which has been trained on thousands of hours of BBC News programmes, has been developed in collaboration with Google's DeepMind AI division.

"Watch, Attend and Spell", as the system has been called, can now watch silent speech and get about 50% of the words correct. That may not sound too impressive - but when the researchers supplied the same clips to professional lip-readers, they got only 12% of words right.

http://www.bbc.com/news/technology-39298199

Anaxagoras · Post by **Anaxagoras** » Sat May 12, 2018 1:24 am

I question whether there's any such thing as a "professional lip-reader"? Maybe as a side gig. What's the demand for lip-reading services? Spooks I suppose?

Witness · Post by **Witness** » Sat May 12, 2018 1:35 am

https://en.wikipedia.org/wiki/Lip_reading

What extraordinary Google-fu I have!

sparks · Post by **sparks** » Sat May 12, 2018 8:16 pm

Anaxagoras · Post by **Anaxagoras** » Sun May 13, 2018 12:53 pm

Witness wrote:https://en.wikipedia.org/wiki/Lip_reading

What extraordinary Google-fu I have!

Well, I guess my next question is, if "professional lip readers" can only correctly identify 12% of words spoken, what is the practical utility of lip-reading?

Pyrrho · Post by **Pyrrho** » Sun May 13, 2018 10:22 pm

Link: https://twitter.com/lukew/status/995769258347909120

xouper · Post by **xouper** » Sun May 13, 2018 10:49 pm

Anaxagoras wrote:. . . if "professional lip readers" can only correctly identify 12% of words spoken, what is the practical utility of lip-reading?

That reminds me, if you haven't already seen this, here's a comical example of bad lip reading (4 minutes):
https://www.youtube.com/watch?v=gneBUA39mnI

Witness · Post by **Witness** » Thu May 24, 2018 11:23 pm

We need an acronym for AI BS:

The Daily Fail wrote:Would YOU turn your loved one into a robot clone? Swedish scientists are using AI to build androids that are 'fully conscious copies' of dead relatives, report claims

– Scientists are looking for volunteers to offer up their dead relatives for the study
– They would build realistic robot clones based on deceased family and friends
– Using artificial intelligence, the scientists can reconstruct the voices of the dead
– Experts have previously detailed how we may be able to preserve our dead family members in the future, perhaps by uploading their minds to machines

http://www.dailymail.co.uk/sciencetech/ ... tives.html

So we'll have 17 generations angrily criticizing us from the mantelpiece? Hope we'll still be able to pull the plug…

sparks · Post by **sparks** » Fri May 25, 2018 10:03 am

What horseshit this way is flung.