B1 中級 3 タグ追加 保存
take a close look at this.
See anything interesting?
This is a picture of complete darkness.
Trust me.
No matter how hard you look, you won't see a thing.
For some, this represents the conditions of a nightmare or a horror film.
So if I told you this was a picture of a modern day workplace, would you find that disturbing?
This is a real live electron ICS factory in Chengdu, China.
This is not after hours.
This is in the middle of the day at the height of productivity.
And there's total darkness at this factory.
Workers work 24 hours a day.
They don't take lunch or bathroom breaks the happy to work without lighting.
And I get paid and no one complains.
Why is that?
Well, these workers, they're not human.
This is what it looks like without the need for basic amenities.
No lights, no air conditioning, no running water, not even toilets.
The things we, as modern humans take for granted.
This is a dock or lights out factory in the dock.
Factory machines automate the end to end production line from processing the raw materials to storing the end product.
And it's sometimes weeks before a human enters the factory floor to inspect or maintain the equipment.
In 1955 this notion was just a futuristic fable.
By 2049 most of Chinese manufacturing will look like this.
I want to show you another image.
Well, sure, to the human eye.
It looks the same.
This time.
It's a different kind of factory.
This one is in Yoshino, Japan.
At this factory, dark factory robots build other dark factory robots.
They can build up to 50 per day and go unsupervised for up to 30 days at a time in the dock since 2003 and they've built over 400,000 to date.
That's right.
In some kind of self replicating human free level of inception, machine intelligence is building other intelligent machines capable of operating without the need with human supervision.
Robots like these are increasingly replacing us humans in factories and other blue collar jobs.
And why simple that just more efficient.
Okay, so you might say short order mating a production like to grind, drill polish and assemble Well, that makes sense.
But I have a white color job.
I'm safe, right?
Well, sorry to break its here earlier this year, we were approached by a top tier law firm and they said, Hey, could you build us a junior lawyer using machine intelligence?
When a law firm hires a junior, they Barry's them with years of casework and a decision tree, they shatter the more senior lawyers they try to learn as much as possible.
Take is much of the work load and then become billable in his little time as possible.
And then they leave all of the time and equity invested, gone and the process starts again.
So when they asked me to improve this process, I said, Sure, How could it be a machine intelligence?
Junior has its advantages.
You can train artificial intelligence in much the same way as you would a human, but in much less time, eh?
I can learn from its mistakes, handle the unexpected with much greater context.
Based on its previous learnings.
It doesn't have a family, doesn't take a paycheck, doesn't need friends, won't take a sick day.
And most importantly, where I leave you to go where kill swear.
One Machine Intelligence junior can replace an army of human genius and within months might even make senior partner the junior where in the process of building will be able to digest decades of case.
Notes and output are ready to go contract in seconds.
That means every few seconds around the clock, this junior will be making money for the firm, and we expected to be fully operational in six months.
It's a reality which is as exciting as it is terrifying.
Welcome to my world.
My team are on the front line of this shift, optimizing jobs through data driven experiments.
We take human behavior, turn into data and teach it to machines.
Sounds like a tall order, but with the right people in the room not really once programmed.
I isn't just cheap, it's pretty much free.
You can train a I by ingesting information, making sense of the structure or lack of structure.
And it's fast.
And I recently deployed by J.
Morgan, can do 360,000 hours of human labor in just seconds.
Because humans are patent creatures and because our jobs have some kind of structure to them, machines should be able to learn how to do them, And that's why we're about to see an implosion of what I like to call organic labor or labour performed by a human.
You can imagine your job as a salad with each ingredient, something you need to do each day.
What will start to see is our tasks, the ingredients in our jobs starting to be eaten away slowly we might lose our daily spreadsheet and then I need to generate reporting without reports we won't need to attend.
Meetings will be way more productive.
There'll be less of us required to achieve the same outcome slowly but surely would be less responsible for more tasks until the majority have been swallowed by the machines.
So it got me thinking which jobs are at the highest risk of automation.
We gathered over 11,000 of the world's most widely held jobs and created the short term Automation susceptibility index.
What we found is that no matter whether it was blue, white or any other color, regardless of skill or intelligence level jobs, tasks can be deconstructed into three things.
Structure, variants and predictability.
We found that up to 80% off low skilled jobs already so automated will be for gotten as human options over the next few years.
These are jobs that make you say what people still do, that jobs like cement mixer, candy machine operator, a video store attendant the next most susceptible of jobs like water meter inspector, stock taker and order fill er up to 60% off.
These low variability jobs will succumb to automation over the next decade, up to 25% off.
Low structure, high variability jobs will shrink Judah having bits and pieces off the day to day being automated.
These are primarily senior blue collar positions or standard white collar jobs such as pharmacist, accountants and health inspectors.
Yet only 3% of highly specialized jobs will follow the same fate.
We're talking neuroscientists, chemical engineers, acrobats, actors and animal scientists.
Machines will have a hard time with these jobs as they require a certain intellectual context, multi disciplinary skills and often fundamentally human attributes such as interpersonal communication and random physical manipulation.
But you'll notice, I said, machines will have a hard time with these jobs, but nothing really is impossible.
It comes down to a fundamental question.
Where would we rather pay to watch this guy before my comedy routine or Seinfeld?
Let's take a no overwhelmingly human job.
For example, sales were recently completed a proof of concept for a 14 500 company, which proved sales could be driven by machine intelligence.
We took a bunch of open source information and key revenue driving criteria, and the A.
I was literally about to tell us how to sell to new leads.
In a world where sales people spend up to 70% off the time on research and administration, even a small improvement means real money through the door.
In this reality, the sales person becomes a delivery vehicle for an A I crafted sales message in a very human way.
So what does this leave us?
Our financial freedom, Our sense of self and a sense of purpose is becoming increasingly threatened as the jobs of today become more automated and scarcer.
As a result, this is the price we pay to make things like legal advice, accounting and counseling cheaper and more accessible to the masses.
But what we must remember is that while we can't take the human out of the job, we can never replace the job of being human.
At the end of the day, we are social creatures and we crave human interaction.
Maybe that means it will cost more to interact with the human instead of a machine in the near future.
But as I see it, tomorrow's world will be faster and smarter, a world optimized to leverage the best version of ourselves.
This is exciting, right?
Machines will liberate human purpose.
So how can we prepare to take a look at your job prospects and the job prospects of your Children?
And ask yourself, Could a machine do this?
Prepare your family to be okay with change.
As uncomfortable as it may be, will likely be switching Korea's Farmall frequently in the near future.
Love it or hate it.
Let's take feelings out of this.
Machines can't be stopped, but they will let up our world in more ways than we could ever imagine.
This is the future of wack.
Let's embrace pushing limits, not buttons.
Buttons are so 2017.


What happens when we take humans out of work? | Tomer Garzberg | TED Institute

林宜悉 2020 年 3 月 20 日 に公開
  1. 1. クリック一つで単語を検索


  2. 2. リピート機能


  3. 3. ショートカット


  4. 4. 字幕の表示/非表示


  5. 5. 動画をブログ等でシェア


  6. 6. 全画面再生


  1. クイズ付き動画


  1. クリックしてメモを表示

  1. UrbanDictionary 俚語字典整合查詢。一般字典查詢不到你滿意的解譯,不妨使用「俚語字典」,或許會讓你有滿意的答案喔