How machine intelligence is remaking the American economy | Big Think

How machine intelligence is remaking the American economy
----------------------------------------------------------------------------------
American companies like Amazon and Netflix are already using artificial intelligence, says data scientist Michael Li, and ones that will not adapt will be left behind. What we need is to expand the discussion and possible regulation of this new technology that is transforming our lives.
----------------------------------------------------------------------------------
MICHAEL LI :
Michael Li, PhD, is the founder of The Data Incubator, an education startup training STEM PhDs to be data scientists and quants. He has worked at Foursquare, Google, A16Z, NASA, JPMorgan, and D.E. Shaw. He is a Hertz, NSF, and Marshall Scholar and is an alumnus of Princeton and Cambridge. Michael founded a student-run conference in quant trading that features leading industry figures and raises money for educational non-profits like America Needs You.
----------------------------------------------------------------------------------
TRANSCRIPT:
Michael Li: We hear a lot about data science, AI, machine learning.
These are all things that are in the milieu right now. I think they fundamentally point at the same idea, the same concept, around you might call it machine intelligence—where it’s about how do you use computers and the vast amount of data that’s out there, that kind of big data, and then leverage that to make more intelligent decisions as an organization, as a government, as a nonprofit.
This really comes from a few major secular trends that are happening.
One is the plummeting cost of computation and the plummeting cost of storage.
So now we have the capacity to store that data relatively cheaply and be able to process that data relatively cheaply
And then the other major trend is that everyone’s walking around with smartphones. Everyone’s interacting with the internet for a large portion of the day. And so we’re able to capture huge parts of the human experience and digitize that information and store it in the cloud.
So when we have all these connected devices that are measuring us, we can actually say a lot about human behavior. And that’s actually really, really fascinating.
And from that we’re able to create products, services that are so much more rich and so much more personalized than we’ve been able to do before.
And so if you think about maybe even the simplest example, it might be something like Netflix with a recommendation engine that’s able to serve up content in a very targeted way so that they give you, they show you out of their library (of probably millions of possible videos for you to watch) the five to ten that you’re most likely to want to watch.
And they can do this from what’s called “look-alike analysis” where they would look at what other people, who have watched a similar set of videos as you have, how have they rated those videos. How much they’ve liked those videos.
And then see what other videos those people have liked that you haven’t yet watched. And that’s probably a good candidate for a video that you should watch.
So that kind of look-alike analysis—or if you’re a data scientist you probably call that a recommendation engine—That’s actually a very powerful technique and it’s sort of very fundamental to a business that has tens of millions of videos and they know you’re only going to watch one tonight.
How do you pick out that one good video so that’s not such a huge search problem for a consumer but it’s actually a pleasurable experience for them? And that has implications beyond Netflix.
If you think about a company like Amazon, that’s incredibly important for them. They have billions of items in their store. You need to be able to figure out what to buy and so they can tell you the right item that can maybe get you to buy something that you otherwise wouldn’t have purchased. And that has a direct impact on their bottom line. And it also makes consumers happier, right?
It helps you reduce the amount of time you spend searching for products and services. So I think these kind of data-enabled services where companies can give you what you want when you want it, that’s becoming increasingly powerful within the kind of consumer market and it’s becoming increasingly the standard.
So I think what we’ve seen is that for a lot of legacy enterprises that are not digital first, that haven’t been able to embrace data and data science, there’s an almost a kind of an adversarial relationship between the consumer and that product or service, where you’re saying as a consumer, “Hey, I have this great experience when I’m interacting with G...

[cestlavegan]

Very eloquent and thoughtful. I hope to hear more from this guy

[shaftwood]

Interesting. I very rarely get anything useful from any of the automated suggestions for movies or products from Amazon. I find those suggestions are actually pretty poor and reflect what some company wants me to buy instead of what's really the best product. A little research on my part and I'm much happier buying a higher quality product.

[Alitari]

My concern is more about companies selling us stuff that we don't want but convince us because 'everyone else like you is buying it, you wouldn't want to be the only person who doesn't have it' ... and using the massive amount of information they know about us, what colours we like, what ads work for us, etc. to manipulate us into buying something we don't really want ... instead of watching one movie, a service like Net Flicks figuring out ways to make us binge watch.

[thijsjong]

I get ads fo rbikes after I bought 1 and ads for phones after i bought 1. It seems i decide too quickly for adaptive algorithms to influence me.

[bonecanoe]

Because of automation human labor is needed less and less every year. The sociological implications of this are huge--you will soon have hundreds of millions if not billions of humans who are not "needed" in the modern economy. How will they fit into the world then? This is already a big problem now--drug use and suicide are already climbing fast, and it's going to get a lot worse.

[jesuspg]

Mexicans are not stealing your jobs, machines are

[shinymike]

I need AI to send me a large sum of money.

[christopherdahl]

I'm a chooser not a consumer. You only offer something that resembles something that you have that you want to make me want.

Don't be Datazuckers.

[psd]

brace yourself for the know it all macroeconomics experts talking -out of their ass- based on intuition in the comment section.

[muskduh]

"The human worker will go the way of the horse and carriage"--Wassily Leontief

[TheyCallMeNewb]

I'm at something of a loss. Does this interviewee have a book?

[herbertholmes]

Wouldn't this feed biases in regard to what content we consume?

[monteharrison]

Uptalking? Causes me to lose interest? In anything you have to say?

[apostolapostolov]

Why didn't he mention bitcoin or etherium? Is he afraid?

[Furiends]

"so much more rich and so much more personalized" I call bullshit. Please name an application thats rich and personalized via massive interactions of privacy and constant data collection?

[cybersekkin]

Sadly, We are using it more to put people out of work.

[Johnny_Ultimate]

the tldr is ai is making it easier for me to consume? Swell....

[AzazelEblis]

By demonetizing youtube channels randomly, and putting anti-gay ads on gay youtube channels?

[AudiTTQuattro]

...so far, ai sucks. Maybe some day it will "know" what I'm interested in etc., but not yet. Besides, finding new things can be way more interesting.

[danbee]

Mandate all A.I. to learn the man Carl Sagan. Perhaps the internet should become free for all if it's intended use and notion of it's developer have veered severely from a communicative tool for all to another local news station just looking claim its number one( *yawn* ) even if it means covering crime and tragedy to get you to buy "prescription" drugs cause your so addicted and anxious.