AI is eating our world | Fabian Westerheide | TEDxHeidelberg
-
0:11 - 0:15Well, I got my first computer
at the age of eight, -
0:15 - 0:17and since then, I'm hooked
in the digital world. -
0:17 - 0:19I was involved in over 30 companies
-
0:19 - 0:23as an entrepreneur
and investor in the last years, -
0:23 - 0:24mainly in the software space.
-
0:24 - 0:27And through influence by science fiction -
-
0:27 - 0:30which I love if it's Gibson
or Hertling or Charles Stross -
0:30 - 0:31and a series like Stargate -
-
0:31 - 0:38I discovered that the natural evolution
of software is artificial intelligence. -
0:38 - 0:41Someone said it could be
the last invention we do as humankind. -
0:41 - 0:45And that's what I would like
to speak with you today about, -
0:45 - 0:47that AI is eating our world
-
0:47 - 0:50and that this is something
we should be happy about. -
0:51 - 0:54Artificial intelligence
is not a new concept. -
0:54 - 0:56Already, the Romans,
the Greeks, the Chinese, -
0:56 - 0:59they had a concept of human-like machines.
-
0:59 - 1:01So it is through human history
-
1:01 - 1:04that we're thinking about machines
who could be more like us. -
1:04 - 1:06Today, we understand machines
-
1:06 - 1:10who look more like us, who think like us,
who see the world like us. -
1:10 - 1:16And also, it got recently
very popular in Hollywood again. -
1:17 - 1:21If it's Gates, if it's Elon Musk,
if it's Hawking, -
1:21 - 1:24they all speak currently
about artificial intelligence. -
1:24 - 1:26And the reason is,
-
1:26 - 1:30three factors came together
in the last three years, -
1:30 - 1:32which are interesting to [note].
-
1:32 - 1:34The first is the increase
in processing power, -
1:34 - 1:35the second is data,
-
1:35 - 1:37and the third is progress
in machine learning. -
1:38 - 1:41We all know that computers
used to be very, very big -
1:41 - 1:43and very slow and very stupid.
-
1:43 - 1:46And they used to fill rooms like this.
-
1:46 - 1:47Rooms like this were computers,
-
1:47 - 1:49and they were just smart calculators,
-
1:49 - 1:51just a calculator.
-
1:51 - 1:56But today, we all have a supercomputer
in our pocket, in our smartphone. -
1:56 - 1:59Thanks to Moore's law,
which is valid for the last 50 years, -
1:59 - 2:00every 18 months,
-
2:00 - 2:05a computer is half the size,
half the price, and double the speed. -
2:05 - 2:08So computers are getting faster
from time to time, from year to year, -
2:08 - 2:09and this won't stop.
-
2:09 - 2:11This is like your brain would get better.
-
2:11 - 2:15Imagine getting smarter every 18 months,
what you could do with this. -
2:15 - 2:16So computers get better.
-
2:16 - 2:19This is the first thing:
they get smarter and faster. -
2:19 - 2:21The second is we have more data.
-
2:21 - 2:25It's like for us humans,
more data means you have more to process. -
2:25 - 2:27And thanks to the internet 25 years ago,
-
2:27 - 2:30we produce, every day,
more data than the day before. -
2:30 - 2:33Whatever you do on your smartphone,
whatever you do offline, -
2:33 - 2:35everything's tracked, everything's stored.
-
2:35 - 2:38Whatever we do, everything,
especially digital, we do, -
2:38 - 2:41this is somewhere safe
and stored on a server. -
2:42 - 2:46And that is why companies
like Google, like Facebook, -
2:46 - 2:50they use these huge, complex
processing powers in combination with data -
2:50 - 2:52to make sense out of the world they have.
-
2:52 - 2:57And they use machine learning
to use all these computers and data -
2:57 - 3:00to make even their business models more,
-
3:00 - 3:03you know, to understand better
what customers do, to make more money. -
3:04 - 3:05Just imagine what this means
-
3:05 - 3:09if computers are getting stronger,
faster and you have more data to do it. -
3:09 - 3:11So machine learning is the idea
-
3:11 - 3:14that you do not need
to program everything, yes? -
3:14 - 3:16If you don't use machine learning,
-
3:16 - 3:18to improve a computer or computer program,
-
3:18 - 3:19you have to write it in the code.
-
3:19 - 3:22You run it, you test it,
and you have to write the code again. -
3:22 - 3:25With machine learning,
the machine is able to adapt itself. -
3:25 - 3:28And with the machine learning something,
the last years had a breakthrough, -
3:28 - 3:31which is really interesting,
and it's called deep learning. -
3:31 - 3:34And deep learning
are neural nets and layers. -
3:34 - 3:36They asked me to explain this
a little bit more. -
3:36 - 3:40So this means it's mimicking
how the brain works. -
3:40 - 3:44And it's - in easy words,
you have two points, -
3:44 - 3:46and when the machine
is doing something right, -
3:46 - 3:48the path between two points
is getting stronger, -
3:48 - 3:49if the machine is doing wrong,
-
3:49 - 3:52the path between two points
is getting weaker. -
3:52 - 3:54And you can imagine
it's like you drive a car. -
3:54 - 3:56When you're 18, you need to learn rules.
-
3:56 - 3:59You go to school, learn these rules,
you have drive-hours - -
3:59 - 4:01In Germany, you have to take
30 hours or more - -
4:01 - 4:03you learn this until you get
more experienced, -
4:03 - 4:05that you don't think anymore;
-
4:05 - 4:07you have something like experience;
it's like instinct. -
4:07 - 4:09This is the same with machines:
-
4:09 - 4:11the more you run the machine,
the better it gets. -
4:11 - 4:15And these machines are in the step
of being cognitive machines. -
4:15 - 4:22So the machines today using deep learning,
they can gather data, yes, -
4:22 - 4:26then they can run simulations
with your mathematical complex models, -
4:26 - 4:28and then they propose actions.
-
4:28 - 4:31And today, you don't tell them anymore
what they have to do; -
4:31 - 4:33they run their own actions,
they execute them, -
4:33 - 4:37they measure the results,
and then they optimize. -
4:37 - 4:39So I repeat this:
-
4:39 - 4:41machines optimize themselves.
-
4:41 - 4:44And we don't know anymore
what's happening inside. -
4:44 - 4:47Machines start trading stocks,
machines drive cars, -
4:47 - 4:50machines develop their own encryptions,
-
4:50 - 4:53and we don't know anymore how they do it.
-
4:53 - 4:56Within their certain frames,
they decide on themselves, -
4:56 - 4:58they improve on themselves.
-
4:58 - 5:00And they do this simultaneously.
-
5:00 - 5:02There are companies out there
who run a machine, -
5:02 - 5:04and they copy them eight times,
-
5:04 - 5:09and within two days, they have 16, 20,
and more different outcomes and versions. -
5:09 - 5:13While a human can do this only afterwards,
machines do this parallel. -
5:13 - 5:15So what I want to say
is we've come to the point, -
5:15 - 5:19I call this the age
of "narrow artificial intelligence." -
5:19 - 5:21It's narrow artificial intelligence
-
5:21 - 5:25because these machines
are very intelligent, -
5:25 - 5:29very good in their specific trained field.
-
5:29 - 5:33Training machines got cheaper
and faster and easier than ever. -
5:33 - 5:36The more processing power you have
and the more data you have, -
5:36 - 5:38the easier there is to train a machine.
-
5:38 - 5:40And these machines today,
we used to train them, -
5:40 - 5:45we already have today machines
who we have taught to train themselves. -
5:45 - 5:47So you don't have to teach them
that they have to learn, -
5:47 - 5:49they know how to learn themselves.
-
5:49 - 5:51So you give them a task,
and then they solve it. -
5:51 - 5:54And they solve it in ways
we can't imagine ourselves. -
5:54 - 5:59They solve it different and faster
and maybe more creative than we do. -
5:59 - 6:02And narrow artificial intelligence
are there already. -
6:02 - 6:04We just don't always see them.
-
6:04 - 6:06Google is using them to figure out
-
6:06 - 6:10what you will type in
before you type it in, yes? -
6:10 - 6:12Netflix is using
narrow artificial intelligence -
6:12 - 6:16to decide what you should watch
before you thought about this. -
6:17 - 6:20We all have artificial
intelligences at home. -
6:20 - 6:22Either it's there in your smartphone
-
6:22 - 6:25or it is at Cortana from Microsoft
or it's from Amazon Echo. -
6:25 - 6:27Well, it's not smart for you,
-
6:27 - 6:31but I know kids who speak with Siri
every day for a couple of hours. -
6:31 - 6:33For them, this is real communication;
-
6:33 - 6:34they don't think about this,
-
6:34 - 6:36"Oh, this is just a machine,
and it's stupid." -
6:36 - 6:39We all have seen this - self-driving cars.
-
6:39 - 6:43This is a form of narrow
artificial intelligence too. -
6:43 - 6:44And there's not only one solution:
-
6:44 - 6:45Tesla has one;
-
6:45 - 6:48BMW and Commodore AI
are working solutions. -
6:48 - 6:51There's even more
than one version out of this. -
6:51 - 6:54Ten years ago, we thought
that this is future, but this is real. -
6:54 - 6:57And it will come -
we will see it in the cities, -
6:57 - 6:59and this will be there and will increase.
-
6:59 - 7:03There are self-driving trucks
based on artificial intelligences. -
7:03 - 7:05This is real, too, it's proven.
-
7:05 - 7:08It's just policies and legislation
take way longer -
7:08 - 7:10until we see it on the road.
-
7:10 - 7:15You can use artificial intelligence
almost in every object; it's like a layer. -
7:15 - 7:16You see it or don't see it.
-
7:16 - 7:17It is everywhere.
-
7:17 - 7:21Wherever you have a computer, you can put
artificial intelligence in there, -
7:21 - 7:23and they can do something for us.
-
7:24 - 7:27That means machines work.
-
7:27 - 7:30Machines work for us;
that's why we invent them. -
7:30 - 7:32We invent them so they work for us.
-
7:32 - 7:37They work for us
24 hours a day, 365 days a year. -
7:38 - 7:42They don't complain, they don't want
a raise, they don't go on a strike, -
7:42 - 7:44and they get cheaper
the more you use them, -
7:44 - 7:47they get more efficient
the more we use them. -
7:48 - 7:53We use machines to do
the physical work we don't like to do. -
7:53 - 7:58We use machines to do
the boring work we don't like to do. -
7:59 - 8:02And we use machines
to do the thinking work -
8:02 - 8:04we honestly don't like to do anymore.
-
8:05 - 8:07(Laughter)
-
8:08 - 8:10And what is the benefit for us?
-
8:10 - 8:14Well, the benefit for us
is we have more "lebenszeit," -
8:14 - 8:17we have more time to live.
-
8:17 - 8:19We only have, in an average life,
-
8:19 - 8:224,500 weeks.
-
8:22 - 8:25So time is - and we learned
this today - is very scarce; -
8:25 - 8:29the most scarce thing we have,
the only limited resource in our life. -
8:29 - 8:33So when the machines do the work for us,
we have more time to live. -
8:33 - 8:36We have more time
for the important things in life, -
8:36 - 8:37which is family, which is friends,
-
8:37 - 8:40which is traveling, your wife,
your husband, your kids. -
8:40 - 8:43It's reading a book;
it's travel the world. -
8:43 - 8:46It's doing something for the community
or start a company. -
8:46 - 8:48When the machines do the work for us,
-
8:48 - 8:50we have more time to be human.
-
8:51 - 8:54However, and this is the first
people ask me, -
8:54 - 8:56"What about jobs?"
-
8:56 - 8:58And honestly, I won't tell you,
-
8:58 - 9:01well, this is, you know,
going to end very easy. -
9:01 - 9:04Yes, machines are there,
-
9:04 - 9:08and they're continuing to do
the jobs for us because we want this. -
9:08 - 9:10That's why we invent them.
-
9:10 - 9:14Because they are supposed to do
the things we don't like to do. -
9:14 - 9:19So there's the estimation that only in,
let's say, in Germany the next 20 years, -
9:19 - 9:225 to 10 million jobs
will be replaced by machines. -
9:22 - 9:23For Europe, it's 50 million.
-
9:23 - 9:25We speak about
hundreds of millions of jobs -
9:25 - 9:29which could be replaced
by machines, robots and computers -
9:29 - 9:31over the next decades.
-
9:31 - 9:35Just think about sales agents,
think about customer support, -
9:35 - 9:38taxi drivers, cab drivers and so on.
-
9:38 - 9:40This is just a small list I compiled.
-
9:40 - 9:42I asked a couple of people
what jobs could be replaced, -
9:42 - 9:44and there are a lot of jobs there.
-
9:44 - 9:49The good thing is it won't replace you
personally right away. -
9:49 - 9:52It's more that you will use
artificial intelligence -
9:52 - 9:55to do more out of your job.
-
9:55 - 9:57You'll use artificial intelligence
and computers -
9:57 - 9:59to leverage what you do today.
-
9:59 - 10:03But in the long run,
companies just hire less people -
10:03 - 10:06because AI plus humans is enough
-
10:06 - 10:09than, let's say, 20 or 30 humans
did this before. -
10:09 - 10:13So AI is just there in supporting you;
it won't replace you right away. -
10:13 - 10:14Just in the long term,
-
10:14 - 10:18we'll need less people
to have the same productivity and output. -
10:18 - 10:19And it's everywhere there.
-
10:19 - 10:23When you sit in front of the computer,
yes, AI will replace you. -
10:23 - 10:27If you're working with people,
AI is not there yet. -
10:27 - 10:29You're on the safe side.
-
10:30 - 10:33And this is something
we should not be afraid of, -
10:33 - 10:35but we need to discuss.
-
10:35 - 10:37This is something we need to be aware of,
-
10:37 - 10:38that this will come,
-
10:38 - 10:40and that's why I'm here today.
-
10:40 - 10:45We need to speak about a system
where we don't have to work anymore. -
10:45 - 10:50Yes, we need to speak about the system
which I call the "jobless future." -
10:50 - 10:53I have this picture of a future
where we don't have to work anymore; -
10:53 - 10:55we only work because we want to.
-
10:55 - 10:57We only do the things
which are our passion. -
10:57 - 10:59We only do the things
which we have a drive for. -
10:59 - 11:02We get up in the morning
because we really want to do this. -
11:02 - 11:07It does not have to be for the money;
it can be for a deeper and higher meaning. -
11:07 - 11:10So we need to speak about the system,
how we can do this. -
11:10 - 11:12I don't have the solution.
-
11:12 - 11:15There are things we can discuss,
like unconditional basic income -
11:15 - 11:16and so forth.
-
11:16 - 11:20But we need to figure
this out as a society, -
11:20 - 11:24how we do this transition
from, you know, a present, -
11:24 - 11:25where we have to work to survive,
-
11:25 - 11:28to a future where this is optional,
-
11:28 - 11:32and where job is not necessary
to have a good living and to be human. -
11:34 - 11:37For the end, let's look a little bit back,
-
11:37 - 11:39for the last 150 years.
-
11:39 - 11:41For the last five generations,
-
11:41 - 11:43life got really, really good
for us humans. -
11:43 - 11:47I can say that we are living
in the greatest time of human history, -
11:47 - 11:49even if we don't see it.
-
11:49 - 11:52Our average life expectancy doubled.
-
11:52 - 11:55Yes, everyone older than 30
here in this room? Yes? -
11:55 - 11:57You can be happy to live today
-
11:57 - 12:02because only 100 years ago
you would be dead, most likely. -
12:02 - 12:04We have way, way more money in our pocket.
-
12:04 - 12:08Our average income and wealth
increased to 600 to 1,000% -
12:08 - 12:09just in the Western countries.
-
12:09 - 12:11And at the same time, we work less.
-
12:11 - 12:15We work half the time
we used to work 100 years ago. -
12:15 - 12:17Average working time was 80 hours.
-
12:17 - 12:20Today, we work 35 to 40 hours on average.
-
12:20 - 12:23So we live longer, we are healthier,
-
12:23 - 12:26we have more money in our pocket,
can buy more things, -
12:26 - 12:27we work less for this,
-
12:27 - 12:28and this is the trend we have.
-
12:28 - 12:31This is the progress
we had in the last 150 years -
12:31 - 12:34because we embrace technologies.
-
12:34 - 12:36Yes, with every new thing
which will come - -
12:36 - 12:38which can be AI, AGI,
-
12:38 - 12:41super intelligence, nanorobotics.
virtual reality and so forth - -
12:41 - 12:42we have challenges.
-
12:42 - 12:46It has challenges for you as a person,
for our system we're living in, -
12:46 - 12:47for companies and so forth,
-
12:47 - 12:48but it's also opportunities.
-
12:48 - 12:52And they would bring us
a better life, a more happy life. -
12:52 - 12:54This is the future I'm fighting for.
-
12:54 - 12:57This is the future
I think we should all fight for. -
12:57 - 13:01And this is what I'm living for,
that we'll all achieve it. -
13:01 - 13:02Thank you.
-
13:02 - 13:05(Applause)
- Title:
- AI is eating our world | Fabian Westerheide | TEDxHeidelberg
- Description:
-
Fabian Westerheide dives into the challenges and opportunities of a world in which intelligent machines and humans coexist together. Artificial intelligence is and will be integrated more into vital parts of human life, and not only human jobs will be replaced by machines. Artificial intelligence will also challenge human thinking, art, and culture. What comes after is the concept of Superintelligence and Singularity.
Fabian has deep roots in the German and European digital ecosystem. After working at numerous tech companies and selling his own startup, he has turned his longtime passion artificial intelligence into an investment activity of his own. As CEO of the family-owned venture capital firm Asgard, Fabian now invests in promising startups in the artificial intelligence and internet-of-things sector. He is the founder and organizer of the annual “Rise of AI” conference and serves as coordinator for artificial intelligence at the German Association of Startups.
This talk was given at a TEDx event using the TED conference format but independently organized by a local community. Learn more at http://ted.com/tedx
- Video Language:
- English
- Team:
- closed TED
- Project:
- TEDxTalks
- Duration:
- 13:23
Peter van de Ven approved English subtitles for AI is eating our world | Fabian Westerheide | TEDxHeidelberg | ||
Peter van de Ven edited English subtitles for AI is eating our world | Fabian Westerheide | TEDxHeidelberg | ||
Peter van de Ven edited English subtitles for AI is eating our world | Fabian Westerheide | TEDxHeidelberg | ||
Retired user accepted English subtitles for AI is eating our world | Fabian Westerheide | TEDxHeidelberg | ||
Retired user edited English subtitles for AI is eating our world | Fabian Westerheide | TEDxHeidelberg | ||
Retired user edited English subtitles for AI is eating our world | Fabian Westerheide | TEDxHeidelberg | ||
Zsuzsa Viola edited English subtitles for AI is eating our world | Fabian Westerheide | TEDxHeidelberg | ||
Zsuzsa Viola edited English subtitles for AI is eating our world | Fabian Westerheide | TEDxHeidelberg |