The future is here (1)-Technological change

The future is here.

Companies that are driven by technology and use algorithms to make decisions are growing wildly, leaving veteran Internet giants in panic.

Replace people with algorithms, this is happening.

Before we read the news, Alpha Dog Go defeated humans, and AI played Dota to defeat quasi-professional players.

On the one hand, we are worried about being replaced by artificial intelligence. On the one hand, we feel that the day of being replaced is far away: artificial intelligence plays Go, but it is not enough to be productive in many fields.

The premise of “very distant” is: artificial intelligence wants to imitate people, and then surpass people, which is difficult. Because of human emotions, thinking is something that machines can hardly simulate.

Humans are arrogant here.

Because as long as one kind of productivity is more efficient than the old one, then the old one will be replaced.

Algorithms surpass humans, that is, they surpass humans without imitating humans at all.

“Flow” technological change

Regarding the issue of “flow”, there have been three obvious stages in the last 20 to 30 years.

(1) First stop, flooding traditional media

The 1990s were the golden age of newspapers and television media. This is the earliest time when people come into contact with a single medium.

During that time, the media and advertising companies were the most beautiful places. A lot of talents are there.


(2) At the second stop, someone started to think about the scene and experience

Here, the leader will be the product manager. They start to think about “scenes” and “interactions”-I think what users should do in this situation will have a better experience.

They are responsible for thinking: I think users should have a better experience in this situation.

Even, they said: Users don’t know what they want, I’ll define what users want.

Define what the user wants. And it also defines success as genius.

To make products, we must rely on “users think this is better”, but not “I think, users will think this is better.”

So, how do you perceive “user perception”? Three ways:

① Talk to the user.

② As a user, experience it yourself.

③ User behavior data.

Traditional Internet companies make use of ① and ②, and they call this “user research”.

They use “true smart people, all work hard.” To praise excellent product managers. This means that the smarter and more experienced you are, the more you have to walk into users and get to know them.

Otherwise, it is easy to fall into empirical judgment-“I think, users will think.”

And this judgment is often not credible.

This is because the user’s behavior data is not collected enough, not good enough, and the algorithm design is not beautiful enough. They believe that their judgment of eyes and hearts is better than that of data algorithms.

However, the data will get better because it gets better faster than people grow faster.

When it becomes more credible than human experience judgment, we will come to the third stop.

(3) The third station is algorithm-driven.

Design from the aspects of data sampling, data utilization, and algorithm construction.

This is a change:

① Product thinking scenarios and interactions; engineers to implement; users give feedback; return to the first step.

② Engineers design algorithms; engineers request data; product thinking how to get data; product design scenarios and interactions; engineers to implement; data to give feedback; return to a previous step.

Later, we will continue to give examples.

Technological change is happening now

When the algorithm’s ability is getting stronger and stronger, until one day it breaks a threshold, which is the average effect of human work.

Then the algorithm will kill most people’s work in an instant.

And the remaining handful of people will be killed by the algorithm a little bit.

Because the evolution of the algorithm is faster than human learning.

The old productive forces have been replaced by ruin. This kind of thing has happened too many times in human history, especially in recent years.

People started spending online.

Things were gentle at the beginning: some companies began to work harder to recruit algorithm engineers.

Just like now, what is happening.


We have talked about a lot of content in this manuscript, but the core sentence is just one: the algorithm replaces people, and it happens in the present.

If it is replaced, it is an inescapable fate. So as an individual, how should one deal with this matter. -Embrace technology.

To add a case, Netflix.

Maybe you haven’t heard of it, it’s simply the one who made “House of Cards”.

But this company is not simple. In June of this year, its market value was more than 180 billion US dollars, and its price-earnings ratio exceeded 300 times.

Even the same industry, the traditional spoiler Disney, has a market value of $ 168 billion.

This is an explanation of the price-earnings ratio. The price-earnings ratio of general technology companies is tens.

But netflix can reach three hundred.

The meaning behind it is that one side is high risk, and the other side is: it is overvalued by capital, and capital likes it.

So the question is, what business is such an entertainment giant doing?

Netflix was the earliest retailer. Later, it transformed into what it is now.

It shows people, but also makes plays, such as the house of cards.

The difference is that all its users are paid users. It’s just that you have a one-month free trial.

It does a very good recommendation algorithm. Based on what you have seen, based on your preferences, based on your identity information … to recommend content to you.

Now it’s time to talk about what it does with algorithms.

Use algorithms to intervene in content selection and production.
Director David Finch once took the adapted playbook of “House of Cards” and found many TV stations in the United States, but none of them dared to pay for it. No one can say whether an old play 20 years ago still has a market.

Netflix conducted a “TV drama consumer habits database” analysis, they found that the audience who likes to watch the 1990 BBC version of “House of Cards” is also a fan of the ghost director of “Social Network” and “Seven Deadly Sins” David Finch, They are also loyal fans of Oscar-winning actor Kevin Spacey. With powerful big data analysis support, Netflix can fully predict the audience and market response, integrate fans of the original “House of Cards” and Kevin Spicer and David Finch fans, and invest in the new version. House of Cards, a hit.

Use algorithms to make interest recommendations.

Netflix has been holding large competitions to recruit talents to improve its data mining processing capabilities. At the end of 2005, Netflix set up a million dollar prize collection algorithm and architecture that can increase the performance of its recommendation system by 10%. In the end, a team of engineers, statisticians, and research experts won a million prizes and successfully increased the recommendation efficiency of the Netflix movie recommendation engine by 10%.

Corresponds to the utilization algorithm. Many times, algorithms are also using people.

People need benefits, and algorithms need data—a large amount of data that is easy to calculate.

Netflix also made a lot of efforts to feed the data to the algorithm.

Infrastructure
In 2010, Netflix completed two data migrations, the first was to migrate the Netflix data center to Amazon AWS, and the other was to migrate the Oracle database to SimpleDB. By 2011, it was migrated from SimpleDB to Cassandra. Using the routing configuration provided by Cassandra, the cluster can be deployed on multiple continents.

Modification of product form

They eliminated the five-star scoring mechanism and changed it to good and bad, a single judgment.

They cancel the user comment function.

This is done to make the data cleaner and reduce interference factors.

In the eyes of traditional product managers, this is a bold decision.

But netflix did it because the starting point was different. Netflix needs to organize the data and feed it to the algorithm to allow the algorithm to produce benefits.

They trust algorithms and outperform human judgment.

Then they succeeded.