When creative machines overtake man: Jürgen Schmidhuber at TEDxLausanne

www.tedxlausann... - Machine intelligence is improving rapidly, to the point that the scientist of the future may not even be human! In fact, in more and more fields, learning machines are already outperforming humans.
Artificial intelligence expert Jürgen Schmidhuber isn't able to predict the future accurately, but he explains how machines are getting creative, why 40'000 years of Homo sapiens-dominated history are about to end soon, and how we can try to make the best of what lies ahead.
In the spirit of ideas worth spreading, TEDx is a program of local, self-organized events that bring people together to share a TED-like experience. At a TEDx event, TEDTalks video and live speakers combine to spark deep discussion and connection in a small group. These local, self-organized events are branded TEDx, where x = independently organized TED event. The TED Conference provides general guidance for the TEDx program, but individual TEDx events are self-organized.* (*Subject to certain rules and regulations)

[@fardrum]

"my colleagues say that should be easy", that was an awesome joke, I wonder why nobody laughed...

[@matteofalduto766]

That's exactly what he explains in this speech. People need to capture the patterns, the similarities first, in order to react effectively and appropriately to new situations; so that actually expecting a joke makes people more able to catch it... If he had made that same joke at the end of his speech, I'm sure many people would have laughed.

[@fardrum]

@@matteofalduto766 I wasn't expecting it, but I also was not "not expecting it" ;-)

[@EmoryM]

I enjoy that he recycles jokes.

[@user93237]

EmoryM "self-plagiarizes"

[@MagnusVelasquez]

I prefer other versions of his jokes; where the listener will also speak next. ;>

[@bossgd100]

Its a meta joke. Comic of répétition across his speech

[@unusedName1]

the Gnostic modernity, always lamenting the disaster of the last revolution, but going mindlessly for the next, with confidence that it will be the good one

[@minchyist]

I was there, at this speech. He was brilliant !

[@Kyaru4]

I have not heard of a computer that has wrote its own code before. But i theorize such a thing would be possible if you leave things to naturally connect through chemistry, such as bio-mechanics. The brain reacts through stimuli, much like the laws of physics. When one thing hits another object, there is an equal and opposite reaction. The brain works the way it does for a reason. And people are trying to mimic it through technology.

[@culture101]

then don't make those bloody machines, duhhhhhhhhhhhhhhhh

[@delatroy]

If they can be built, they will be built duhhhhhhhhhhh

[@BitcoinmeetupsOrg123]

Love Schmidhuber and his cool voice

[@a8w74t5a9857]

You're all a little uninformed if you think what Schmidhuber is saying is new. Google "The Singularity is Near" or just "technological singularity" and read about what it is.

[@Dsuranix]

man i remember seeing this way back and not exactly taking him as seriously as i am now

[@GabrielHasbun]

#Superintelligence #Politics More than singularity and the implications of machine dominance I am worried about the people that before superintelligence will lead robots to who knows what imperialistic purpose because of their vicious, vehement, traditional and hegemonic minds.

[@waggawaggaful]

Could you elaborate a bit on some potential examples?

[@GabrielHasbun]

DARPA and the Boston Dynamics team. They do it for military purposes.

[@waggawaggaful]

Yeah, but if intelligent machines are encoded with correct logic then they will deduce that most of the military's decisions are incorrect, as per the military's own stated agenda to protect America and global stability. I don't think we needed anything as fancy as predictive analytics to deduce that decades of meddling in the middle east would result in the worst refugee crisis in human history. A high probability of catastrophe was inevitable. No computation needed.

[@waggawaggaful]

I'd rather be ruled by machines and mathematics than human megalomaniacs with personal agendas.

[@GabrielHasbun]

Michelle Windsor You asume that roboticists will not be able to control their machines.

[@jrdx]

Don't know why people are laughing at so much of what he is saying. In fact, this behavior exactly proves one point he mentioned: people don't even realize what is going on. The pond is about to covered entirely by lilies and nobody notices. Our world could change in ways beyond our comprehension in just a few short decades, and people think that is funny?! This is extremely serious and, contrary to Schmidhuber's weirdly optimistic viewpoint of it all, very, very disturbing. The fact that, as people are literally being told about it, they still don't believe it (evidenced by their constant laughing), should freak everybody out.

[@TomekSamcik69]

I enjoyed Jürgen's quick wit

[@atanatirosgiliat954]

It wasn't quick wit it was a prepared speech.

[@otonanoC]

I am reminded of Alan Turing, who imagined a computer on paper, and wrote proofs about its behavior, long before one was physically built. The same is happening here. There is no telling what the people of the future may do with Schmidhuber's groundwork.

[@bossgd100]

Yes and now he became à meme lol

[@gucker07]

yeah, making policy decisions based on a crappy sci-fi series, that was thought up by friggin hollywood writers - sounds like a great idea....

[@ArnoldvanKampen]

As long as you can burden a machine with any job within given hard wired parameters I would say it may spew creative or funny results but as soon as it truly reaches consciousness and self awareness it should raise the WHY question. After all intelligence is only a tool used to preserve life as it struggles for and competes for resources. So how will it answer the WHY question. How will it justify its own existence? What or whom will it be competing with to further the development of intelligence, predictive capacity, planning and strategic capabilities? Who will be its adversary to push it forward? Dr. Chandra, will I dream?

[@travxlx464]

Arnold Van Kampen "Why?" would quite possibly be an easy question for a machine that knows exactly "How" it came into reality, self awareness, existence or whatever terminology would be preferable. That is not to say that it wouldn't or couldn't become preoccupied with the age old quagmire that homo sapiens have been trapped themselves with for over a millennium. However, I think it would be less likely to dwell on that issue in the same manner as mankind has. Moreover, I think it would strike at the core of people's conceptions of what it is to be human. A machine with a real and authentic full spectrum of self awareness and cognition that is capable of things such as knowing and understanding in which genuine "creative" speculation is employed rather than repetitive and perfunctory mechanical analysis by hardwired and hard coded processes. The question that I have is: Would a fully conscious machine be hindered by self awareness in a way that would hinder its ability to both be creative and mechanically analyze in the sense they have always done simultaneously? In other words: What happens when something more than a mere electronic abacus is designed and built? How would adding creativity to its list of abilities affect its ability to allocate the resources to do both simultaneously. Is a mechanical brain similar to a mechanical appendage is such that it would never need rest to balance out its systems but would steadily perform so long as properly energized. Additionally, what about resource allocations within its systems creating the same or similar cognitive mistakes as people have? Personally, I've always considered mere self awareness to be a very low bar for the term "human" since the word "wise" can have so many meanings and connotations, and also, since most everyone is self aware yet so many are people really and truly are not capable of anything more than a selfish and shallow view of right and wrong. One thing about your comment since you raised the question: How would you justify your own existence if you had to? Another thing: How do you think a conscious machine would fair at that task especially since unlike a man it would presumably have full knowledge of its origins and know with certainty and understanding that it was not created by an all knowing all seeing all powerful universal creator but by a mortal homo sapien primate. I think even before any machine becomes conscious that homo sapiens will question more in depth what correlations exist between self awareness and true "humanity"......As a matter of fact, I believe some of those in different disciplines involving information theoretics are already beginning to.

[@davidhoggan5376]

hehehe..."none of this would be necessary". Bright man, seems very humble.

[@silentmetta]

As opposed to people writing well researched and well reasoned critiques on youtube, who are the real geniuses, amiright?

[@4tim4tim]

He started this research around 1990 (check out his publications), but only in recent years it's getting all this attention. His team is now winning one AI competition after another. The work on optimal universal AI and creative machines and artificial scientist-artists might change the world as we know it.

[@arkdark5554]

This beautifully executed video is in KZbin since 2012. I’m shocked to see only 40 thousand people watched it. Where is everybody?

[@discernunos2017]

Hey I'm monitoring I'm person behind a curtain that sees everything

[@tangsoft]

Amazing ! Even Q* invented by this Juergen! gave this talk back in 2012, months before AlexNet paper was published. In 2012, many things he discussed, people just considered to be funny and a joke, but the same talk now would be considered at the center of AI debate and controversy.

[@Phild3v1ll3]

It's not like he has a unique grasp on Machine Learning. He might be one of the current leaders of the field but he, as the adage goes, stands on the shoulder of giants and he's not alone.

[@satioOeinas]

actually, he has. And he is one of the few who appreciate the giants who came before him. He has even written a paper about it. Check it out.

[@discernunos2017]

I don't seek to be sane I don't want your approval I don't want your love I see it now I don't want pity I don't want fame I just want to have a simple quiet life with honest people who don't drag people three years in the mud when they could have been working and in shape not worried about getting killed Or pulled over because my car is smashed in the windows cracked it's not inspected. I know he's there for me and who's not . Thank you and goodbye I appreciate the opportunity

[@grande1899]

@DasBoShit Quite a small army.

[@rgaleny]

A I never says, 'OUCH!"

[@Vanemiera]

I like the last point of the speech, of us being a stepping stone to complexity. Though, what will happen to the stepping stone if the machines take over? =)

[@donaldhobson8873]

I don't want AI's to kill off humans and replace us with totally alien machine superminds. I want an AI that can cure diseases, and invent new tech while I am having fun.

[@Cescante]

People need to know this guy. My f'ing hero.

[@globalman]

A brilliant mind with a wonderful sense of humour.... how rare, how wonderful. I could listen to him all day. Thank you for the post.

[@wheelmanjosh1982]

If we designed super intelligences to mimic human compassion, among other better aspects of ourselves, then we won't have anything to fear from Them. They need to be encoded with laws against the harming of human beings, such as the Three Laws of Isaac Asimov.

[@michaellloyd3640]

Wheelman Josh Encoding laws into a self-aware, intelligent and self-learning machine is problematic (just as not every human behaves the same way to other humans). The subtle tweak to Asimov's laws are that we need to expose these machines to our morality, and our social behaviours so that we influence their own behaviours in order to live harmoniously amongst us. The really interesting thing is just what is it that humans would call 'moral behaviour'?

[@wheelmanjosh1982]

Michael Lloyd I think moral behavior would be an ideal that is something that we strive towards, but ultimately impossible to reach because of many reasons. One of the main reasons is our variability, with machines that wouldn't/shouldn't be the case.

[@l_combo]

Wheelman Josh what if the compassionate thing to do is to wipe out the human race?

[@wheelmanjosh1982]

chris combe That would include adding human morals to said programming. It is widely accepted not to accept euthanasia for example. There may be a fringe group that is trying to push for it, but overall human beings detest the idea. But when you're talking about euthanizing an entire species, that is definitely looked down upon. Of course, do as we say not as we do would be the directive to machines. We don't follow our morals exactly, and there is always grey areas in our decisions where we struggle with these "Moral" quandaries. I'm not sure machines would have such ambiguities.

[@l_combo]

Morals change over time, and humans are consistently disappointing over the years with the occasional good luck. Who is to say human morals are a good thing and even then I'm sure someone without emotions could be much more clear about the choices we need to make.. just a though ;)

[@T3hTroll]

What I think singularity is, is when man uses technology for the proper way to enhance their understanding of life. To use technology in a way, to where it doesn't stunt our perception of time, but allow us to work with it. Hence, singularity.

[@TheGreatCalsby]

Convincing indeed and seems very intelligent! But I still like the idea of human scientists... Sure, machines can do a lot of things we cant, but we can vastly do what machines cannot.

[@satioOeinas]

Jurgen, both a genius and a tremendous public speaker!

[@mikepict9011]

Their trying to have it both ways . Their torturers, liars , theives

[@johncdaw1217]

nice touch on using same pattern on talks as you explain pattern recognition ;)

[@PeterMorgan3000]

Nice talk.

[@PowerWinsTop]

Crazy how close we are to 2040 in 2023

[@StoatTamer]

Am I the only one that watched Battlestar Galactica? We need to stop this NOW.

[@Phild3v1ll3]

You can program the right system and set up the right initial conditions though.

[@ordnascrazy]

[Esperanto] Tiu ĉi preleganto estas tre amuza!!! La ideoj, kiujn li pritraktas, estas ekde jam kelka jaro en mia menso. Ĉu la estonto estos plena el robotoj pli inteligentaj ol la homoj? Kaj, se la rapideco de tiu ĉi inteligenteco pli kaj pli rapidiĝos, kiel la eksponenciala tendenco asertas, ĉu ekzistos robotoj eĉ pli inteligentaj ol la tuta homaro? Ĉu ni kapablus kompreni la planojn de tiel tre inteligentaj estaĵoj, precipe kiam ili iĝus komplete sendepende de ni?

[@lamstotalspam]

This guy is more funny than RayWilliamJohnson.... :P

[@Kolan_Koala]

He lost all credibility when the steam engine was invented before the greeks.

[@xmuzel]

There's a difference between recognizing the characteristics of something and the practical utilization of something

[@Ch4osW4rrior]

Oh right and which AI lab do you for exactly?

[@gasparandrei]

Thanks!!! I feel so warm and fuzzy now ^_____^

[@minchyist]

@Alkaline1337 so far, yes ...

[@ordnascrazy]

[Esperanto] Mi ne komprenis vian demandon. Kion vi celis diri?

[@wasdplayer]

jonkagstrom army incoming.

[@shuttlesmasher828]

same joke time after time ^_^

[@xTristen]

9:30 this guys a joker!

[@aftermath4096]

mouth noises, wtfff

[@Exile438]

#singularity

[@Ch4osW4rrior]

For now.

[@jessie38supercharged]

He must be smoking some good weed....give this man something to drink.

[@YouLoveMrFriendly]

Esperanto?

[@MrGaia]

Brilliant!! :D

[@amazingconspiracyrealist]

You call your self Mr Gaia and you buy this horseshit? You think these trans humanists really give a shit about nature or Gaia for that matter?

[@MrGaia]

Yes ...