本字幕由TME AI技术生成
Npr
Alright
We have a ridle for you see
If you can guess what we're talking about
Today i was born at a denny's restaurant
My name is letin for envy
I played a lot of video games in my younger years
Today i wean't ted poetry and translation
I'm thirty one years old
And i've been down one out so many times that my unofficial motto is that comcomtichip deout business
Business wldn't it
It
If we said
Said the comfor for moomoment and the sevaluable company on the planet
Yeah
That's
That's right the answer is the computer chip designer in video and in videoous chips are the leading choice for artificial intelligence developers in the last couple of years
Dvidious sales have shot up like a jet plane
It's stostoked price more like a rocket
But there are several chip companies out there like amd and others
So why is in videa the one dominating the ai sce
This is the inicicator from planet
Money
I'm willing want
And i'm doing woods tit t on the show in video
We learn about how this company went from selling niche graphics cards
But gaming to becoming the biggest company in the world
And we ask whether in video has the most to defend its very valuable
Castle
I'm reital martin after hosting morning addition for years
I know that the news can can wear down
So we made a new podcast called wild card
Where a special deck of cards and a whole bunch of fascinating gas help us sort out
What makes life meaningful
It's part game show part exessential deep dive
And is seriously fun jojome on wild card
Wherever we get your podgast only from npr on this week's episode of wildcard
Comedian
Taylor
Thomlinson explains
How you can use fear as a motivating force
I was afraid that i would get years down the road and a man
I really wish i had pursued that
Or i wish i had developed this town
That might have taken me somewhere
I'm right til martin joined us for m pis wildcard podcast the game
Where cards control the conversation
嗯
Npr plus is a new way to support public media and get more from your favorite np r podcasts like fresh air
Sometimes of was actually prefeshed quesquestion
If it makes kes you uncomfortable to talk about of its two personal just tell me
Here's the question for behind the scenes content bonus eppisodes that's more sign up at plus that npr dot org in video is often called a chip maker
That's a bit of an overeplification in of of all incvidia doesn't actually manusung the semicduductor chips that's done by companies like tsmc and samsung in videa
Does the designing like where the circuits or go al in the chip
It also makes the software
So that developers can work with those chips
This combination has its roots ts in sannzi
Jensson
Hugg and two other engineer friends met at a denny's restaurant in sanhouzi
Their idea was you improve video games by building specialized chips for rendering three d graphics
This decision would become incredibly look er
But one person who spent a lot of time on earthing in videos
History is david rosen foll
David is one half of the tech podcast acquired which made this epic series of podcasts on the company's history in video corporate communications
Gatden
Touchson said
Who were your sources
Who told you all this in the company
And we like well i would
Nobody we just to know
We watch a lot of youtube videos with jensen
And we read a lot of favs this led to an in person interview with jhnson huu gg towards end end layear
When he wdid publalized what i should see what i realized that i didn't learn about it until to like isshould to bioaccounts working for jensen is super intense and not everyeverybody failis a shared publicly for example
But david says
Those who stay a loyal davson is is really ly singed in soilicon valley
And i think i mean the man is sixty years old
Nobody from his generation is still running their company in the same way
That he is
When he 's probably more engaged in every detail in in videa
Than any of those other founders ever were at their compies
David says
The way other jennson is involved in the details is one key to white in video wrote the ai wave
So is stute tly he realize moving the invidia ship in the direction
That it is today with a iron machine learning starting in like twenty years ago
It's incredible fossiside
You know it's obviously way before chetch bbt
But it's even before driverous cars or even voice recognition
How did he see this coming well
I think this
This is the key agensity
If he were to say that sentencto him at think so deresponsiand say no
It was not before all these days
All of those things were happening
They were just happening in the deep scientific computer science research community
He was so deeply plugged into all of this
That he knew the principal community uh who now are some of the true leaders in ai research at open ai and in robic
An elsewhere
These days he knew them personally he was so inthe white papers
He was visiting them at universities
All back in the universities at this time
And video was mainly known by fifteen nearelds upgrading their gaming computers
But in the early two thousand of people
Jesson essentially made a huge bet on selling supercomputing power to a wider range of people
But as part of this
The company laong a software development framework called cuda
Cuda acts kind of like a middle man between the software developer and the chip and the system would be crucial for ai after a turning point in twenty twelve
This was the big bang moment that kicked off research and investment at a commercial scale into artificial intelligence
It was at this annual competition for researchers to submit ai systems that could recognize images from a massive database
The database ase called led imagnet
The progress had been gradual entries could correctly identify what a majority of images were
But typically have have an era rate of around rectually or thirty percent that year a team from the universive toronto submits this entry that they call alex nait yet blows away the rest to the field
So the percentage of images it got wrong was around fifteen percent
One Five
This was a acquanomlep
The team used an existing mecessing ai called a neural network
Which was powerful
But had always been limited by how much computing power
It required to course
So instead alex that did something unit the standard approach had been to train these models using central processing units or cpu the generalist brains in your computer
But these calculate their instructions sequently that can think of it like one calculation than another
So instead this team ran these training calculations
Which a graphics processing units
Gpu
That you would play video games with a graphics card handles tens of thousands
Set of grastructions at a time
That i've really widen the pipe of an amount of compute
That i can stuff through this thing and any given point in time and who the exgraphics cuds in video
Of course people at google and people at facebook said holy crap you could use
They realize these image classification systems to build way better social media feed recommenders so for image recognition programmers or training the softwers to notice patterns and say pictures of cats whiskers fur for legs and similarly the big social media companies realize they could train software to notice patterns in what kinds of pictures and posts and movies
People like on the internet that was beknow billions of billion
The bilies of dollars of profits
So there was a good tenyear run
Where like nothing else mattered
And that meant that places like youtube and facebook and instagram were hiring a lot of ai talent
And david says
This was actually part of the motivation for elon mosquk and same openments to found open ai
They were really worried that google and facebook had just become a duoppoly of all the ai development in research talent
Because it was the only huitdoeconconomicvipable use case open as launch of chat gbt in late
Twenty
Twenty two was the next hinge moment for ai and for invidia
It woke investors in everyday people to advancements in ai
That wo't dreining about the future
It's where invidia suddenly started journey becoming one of the most valuable companies on the planet
Because the pollonance of tag inveted invidia has the giant mote
David reckends
It's protected from competition in the foreciable able ture ture
Because of tda develoveloment ment stestem these systems build totop of themselves over time and get more and more complex and powerful
So it would be like somebody going and starting a new phone operating system from scratch and saying ok
What are all the list of things we need to build to make this phone operating system viable to compete are android id
Ios sh that's a tall order
Davd says
Network affects are also important and millions of other developers use cudda
So if you're a college student
That's the language you're going to learn it's a self reinforcing cycle
This mote is something
That has gotten people speculating about government action news outlets at the new york times have reported that the department of justice has cleared the way for possible enti trust action against in video
Jenson though has said he doesn't like the wood mode and prefers to think of the company is working along side the millions of ai coo system
Whichever words you use to describe it in video is standing on top of the world
Right now though even with all its chip sales
Its stock price is very highly priced by conventional measures
One mss steff
And there is a long way to fall all the way into the moat built with bare cuta
Barcota is that's type elit
The epsob was produced by jelly a ritchie with engineering by scene of ve fredo
The spectacch by cuba cats for camkcaken canddon is the shows
Editor and the indicator is production of npr
New from the embedded podcast
What happens when three republican women challenge their own party
Maybe we need to speak out a little bit border
Maybe we need to do something we get people's attentioned
They have a front row seat to democracy
Now you do too listen to supermajority from mpr s embedded in wpln have you ever thought about the political leanings of sports fans
The most democratic leaning sports vabies was aw mba like about a comfortable distance
But that could change
If the wmba continues with its explosive growth
Because as new fans show up some of them are bringing old culture war battles
Listen to its bineminute from npr
If you haven't kept up with the best songs out this year we've got you covered on all songs considered from npr music
We hit pause to look back at the best tracks of twenty twenty four four so far chapel one is just a complete star on this song
And it's been wedged in my brain
Ever since i first heard it listen to all songs considered every tuesday
Wherever you get padstcasts
