WEBVTT - OpenAI Part 1: The Most Silicon Valley Man Alive

0:00:00.320 --> 0:00:04.320
<v Speaker 1>It's a Wednesday in November twenty twenty three. I'm standing

0:00:04.360 --> 0:00:08.440
<v Speaker 1>in a dark restaurant in San Francisco. There are a

0:00:08.480 --> 0:00:12.160
<v Speaker 1>few dozen people. It's the after party for an AI conference.

0:00:13.320 --> 0:00:17.119
<v Speaker 1>People are standing and chatting in small circles. Waiters are

0:00:17.160 --> 0:00:21.280
<v Speaker 1>bringing trays of arderubs around. I'm biting into a tiny

0:00:21.280 --> 0:00:25.280
<v Speaker 1>Porcini mushroom doughnut when I hear a whisper. Sam Altman

0:00:25.360 --> 0:00:31.280
<v Speaker 1>has walked into the restaurant. I try to spot him

0:00:31.320 --> 0:00:34.040
<v Speaker 1>in the dim lighting. I've talked to Sam here and

0:00:34.080 --> 0:00:36.840
<v Speaker 1>there over the years. He's been around the startup scene

0:00:36.840 --> 0:00:39.240
<v Speaker 1>for more than a decade, and he's friendly to journalists

0:00:39.960 --> 0:00:43.680
<v Speaker 1>and over time, Sam Altman has become basically the most

0:00:43.720 --> 0:00:47.600
<v Speaker 1>Silicon Valley man alive. He's the CEO of open Ai,

0:00:48.080 --> 0:00:50.280
<v Speaker 1>which shot to tech startum a year ago with the

0:00:50.320 --> 0:00:54.160
<v Speaker 1>release of chat GPT. His company is worth eighty six

0:00:54.200 --> 0:00:57.040
<v Speaker 1>billion dollars. It's one of the most valuable startups in

0:00:57.080 --> 0:01:01.040
<v Speaker 1>the world. For the past six month months, Sam had

0:01:01.080 --> 0:01:05.200
<v Speaker 1>been everywhere. It seemed like every major news outlet wanted

0:01:05.240 --> 0:01:07.640
<v Speaker 1>to profile him, and he was saying yes to each one.

0:01:08.400 --> 0:01:12.119
<v Speaker 1>Headlines referred to him as the chat GPT King, the

0:01:12.120 --> 0:01:16.600
<v Speaker 1>Oppenheimer of our age, and an AI overlord. He was

0:01:16.640 --> 0:01:19.319
<v Speaker 1>like an ambassador for the AI future, zipping around the.

0:01:19.280 --> 0:01:21.000
<v Speaker 2>Globe meeting with world leaders.

0:01:21.520 --> 0:01:24.840
<v Speaker 1>He testified in front of Congress and amid all that

0:01:25.400 --> 0:01:28.280
<v Speaker 1>he decided to pop by this AI conference party a

0:01:28.280 --> 0:01:33.360
<v Speaker 1>little surprise visit. So I see him standing in a

0:01:33.360 --> 0:01:36.399
<v Speaker 1>corner of the restaurant. A few people have already clustered

0:01:36.440 --> 0:01:39.840
<v Speaker 1>around him. I decide to say hi. He's wearing a

0:01:39.920 --> 0:01:42.800
<v Speaker 1>suit and a tie, more dressed up than usual, and

0:01:42.840 --> 0:01:46.880
<v Speaker 1>he's shorter than I remember. We shake hands. He eyes

0:01:46.920 --> 0:01:49.040
<v Speaker 1>my conference badge and says, it's good to see you.

0:01:49.760 --> 0:01:52.720
<v Speaker 1>I'm surprised by how personable he is, how friendly he

0:01:52.760 --> 0:01:57.320
<v Speaker 1>comes across. He acts like he remembers me. He already

0:01:57.360 --> 0:01:59.720
<v Speaker 1>knows I'm working on this podcast. So I say to him,

0:02:00.120 --> 0:02:01.840
<v Speaker 1>I'm about to put in a request to your comms

0:02:01.840 --> 0:02:03.280
<v Speaker 1>team to see if we can find a time to

0:02:03.280 --> 0:02:08.160
<v Speaker 1>interview you. He's like, sure, sounds good. Several people are hovering,

0:02:08.280 --> 0:02:10.400
<v Speaker 1>hoping to get his attention for a minute, so I

0:02:10.440 --> 0:02:13.400
<v Speaker 1>step away, and after about ten minutes at the party,

0:02:13.639 --> 0:02:18.760
<v Speaker 1>he leaves too. The next day, Sam continues his tour

0:02:18.840 --> 0:02:22.200
<v Speaker 1>as the Statesman of AI. He speaks at APEC, a

0:02:22.240 --> 0:02:25.320
<v Speaker 1>big conference at President Biden and China's leader Shijin being

0:02:25.400 --> 0:02:29.399
<v Speaker 1>both attended. Here's Sam at that conference talking about AI.

0:02:29.880 --> 0:02:33.000
<v Speaker 3>I think this will be the most transformative and beneficial

0:02:33.040 --> 0:02:36.640
<v Speaker 3>technology humanity has yet invented. I think this is like

0:02:36.880 --> 0:02:39.360
<v Speaker 3>the greatest weep forward of any of the of the

0:02:39.360 --> 0:02:43.280
<v Speaker 3>big technological revolutions we've had so far. So I'm super excited.

0:02:43.760 --> 0:02:46.400
<v Speaker 3>I can't imagine anything more exciting to work on, and

0:02:46.520 --> 0:02:48.720
<v Speaker 3>getting to do that is like the professional honor of

0:02:48.760 --> 0:02:50.400
<v Speaker 3>a lifetime. It's so fun to get.

0:02:50.320 --> 0:02:54.960
<v Speaker 1>To work on that, the most transformative technology humanity has

0:02:55.080 --> 0:03:00.680
<v Speaker 1>yet invented, the professional honor of a lifetime. Sam clearly

0:03:00.720 --> 0:03:04.480
<v Speaker 1>believes the world is about to change dramatically and fast,

0:03:05.320 --> 0:03:08.160
<v Speaker 1>and he's aware that a lot of that change is

0:03:08.160 --> 0:03:13.040
<v Speaker 1>connected to him. And the day after that, less than

0:03:13.040 --> 0:03:15.600
<v Speaker 1>forty eight hours after I saw Sam at that party,

0:03:16.040 --> 0:03:18.360
<v Speaker 1>less than twenty four hours after he spoke at a

0:03:18.440 --> 0:03:22.760
<v Speaker 1>major international gathering of world leaders, he was fired.

0:03:23.440 --> 0:03:28.080
<v Speaker 4>Breaking news. Sam Altman is out as CEO of OpenAI.

0:03:28.280 --> 0:03:29.080
<v Speaker 2>This is a stunner.

0:03:29.520 --> 0:03:31.920
<v Speaker 5>The tech world has been thrown into chaos over the

0:03:31.919 --> 0:03:35.000
<v Speaker 5>weekend when the company that gave us chat GPT fired

0:03:35.080 --> 0:03:39.400
<v Speaker 5>it's CEO. The move came as a complete surprise to everyone,

0:03:39.520 --> 0:03:41.760
<v Speaker 5>including open AI's biggest investor.

0:03:42.040 --> 0:03:46.000
<v Speaker 3>Microsoft board says it pushed Altman out after a review

0:03:46.120 --> 0:03:49.520
<v Speaker 3>found he was quote not consistently candid in his communications

0:03:49.520 --> 0:03:50.080
<v Speaker 3>with the board.

0:03:52.480 --> 0:03:56.440
<v Speaker 1>It's hard to emphasize enough how shocking this was. Sam

0:03:56.480 --> 0:04:00.080
<v Speaker 1>Altman got fired. It was the juiciest tech news of

0:04:00.120 --> 0:04:03.560
<v Speaker 1>the year and completely unexpected. The news came on a

0:04:03.600 --> 0:04:06.720
<v Speaker 1>Friday afternoon, and my colleagues and I all immediately knew

0:04:06.760 --> 0:04:10.440
<v Speaker 1>our weekends were out the window. The next few days

0:04:10.520 --> 0:04:14.280
<v Speaker 1>felt like a hurricane of news. The board implied Sam

0:04:14.320 --> 0:04:18.160
<v Speaker 1>had lied to them. They announced a new CEO. Then

0:04:18.279 --> 0:04:21.920
<v Speaker 1>came a regretful public apology by someone who had fired Sam.

0:04:22.440 --> 0:04:27.160
<v Speaker 1>There were pledges of loyalty, an employee revolt, and at

0:04:27.160 --> 0:04:30.880
<v Speaker 1>the same time something else was happening. It started to

0:04:30.920 --> 0:04:34.200
<v Speaker 1>seem like Sam was actually going to come back. He

0:04:34.320 --> 0:04:38.360
<v Speaker 1>was rallying the support of his employees and Microsoft. It

0:04:38.400 --> 0:04:45.520
<v Speaker 1>was looking like Sam could win. On Tuesday night, Sam

0:04:45.600 --> 0:04:49.239
<v Speaker 1>was reinstated as CEO at the company that just five

0:04:49.320 --> 0:04:53.720
<v Speaker 1>days earlier had basically tried to destroy him. Sam managed

0:04:53.720 --> 0:04:57.560
<v Speaker 1>to get the upper hand once again. It was high drama,

0:04:58.000 --> 0:05:00.920
<v Speaker 1>a jaw dropping turn of events, but for those who

0:05:00.920 --> 0:05:05.200
<v Speaker 1>know Sam, it actually made perfect sense. One of Sam's

0:05:05.240 --> 0:05:08.640
<v Speaker 1>mentors is an investor named Paul Graham. He once told

0:05:08.680 --> 0:05:14.520
<v Speaker 1>a reporter Sam is extremely good at becoming powerful. To me,

0:05:15.000 --> 0:05:19.920
<v Speaker 1>this seems like Sam's defining characteristic. Often in Silicon Valley

0:05:20.000 --> 0:05:23.839
<v Speaker 1>we talk about tech visionaries who are programming geniuses or

0:05:23.960 --> 0:05:27.520
<v Speaker 1>obsessed with the details and design of the product. That's

0:05:27.600 --> 0:05:32.480
<v Speaker 1>not Sam. His strongest, most unique skill is wielding power,

0:05:33.320 --> 0:05:40.880
<v Speaker 1>and that might have consequences for all of us. You're

0:05:40.960 --> 0:05:45.720
<v Speaker 1>listening to Foundering, I'm Ellen Hewitt. In this season of Foundering,

0:05:45.880 --> 0:05:48.960
<v Speaker 1>we're going to chronicle the rise of Sam Alman, the

0:05:49.120 --> 0:05:51.799
<v Speaker 1>all out arms race to build the leading AI company,

0:05:52.360 --> 0:05:54.960
<v Speaker 1>the claim that this new technology could threaten to wipe

0:05:54.960 --> 0:05:57.919
<v Speaker 1>out all of humanity if we're not careful, and then

0:05:58.080 --> 0:06:00.680
<v Speaker 1>the coup that almost took down the guy at the

0:06:00.720 --> 0:06:02.960
<v Speaker 1>top of it all before he managed to.

0:06:02.960 --> 0:06:04.119
<v Speaker 2>Claw his way back.

0:06:06.960 --> 0:06:09.560
<v Speaker 1>For this season, you'll hear reporting done by me and

0:06:09.640 --> 0:06:12.680
<v Speaker 1>my colleagues at Bloomberg who have been covering AI during

0:06:12.720 --> 0:06:15.840
<v Speaker 1>the boom of the last few years. We interviewed some

0:06:15.880 --> 0:06:18.080
<v Speaker 1>of the leading minds in AI to try to cut

0:06:18.080 --> 0:06:21.400
<v Speaker 1>through the hype and understand the debate about whether AI

0:06:21.480 --> 0:06:24.599
<v Speaker 1>will be a tool to improve human existence or to

0:06:24.680 --> 0:06:28.599
<v Speaker 1>extinguish it. But this is also the story of Sam,

0:06:28.920 --> 0:06:31.039
<v Speaker 1>the man at the center of it all, and we

0:06:31.080 --> 0:06:35.480
<v Speaker 1>spoke with Sam's friends, family, and collaborators to demystify him

0:06:35.640 --> 0:06:41.200
<v Speaker 1>and how he rose to power. This first episode is

0:06:41.279 --> 0:06:44.320
<v Speaker 1>all about how Sam got here. He's a man who

0:06:44.360 --> 0:06:46.839
<v Speaker 1>has always understood the importance of being in the right

0:06:46.920 --> 0:06:50.440
<v Speaker 1>room at the right time, with exactly the right few people.

0:06:51.800 --> 0:06:55.840
<v Speaker 1>The full story of Sam's rise is important because understanding

0:06:55.880 --> 0:06:59.440
<v Speaker 1>who he is and what he believes will shed light

0:06:59.480 --> 0:07:03.760
<v Speaker 1>on an earth question should we trust this man to

0:07:03.880 --> 0:07:10.000
<v Speaker 1>oversee this technology? In the summer of twenty twenty three,

0:07:10.160 --> 0:07:13.480
<v Speaker 1>about five months before he was fired, my colleague Emily

0:07:13.560 --> 0:07:17.040
<v Speaker 1>Chang asked him this exact question at a Bloomberg conference.

0:07:18.000 --> 0:07:21.080
<v Speaker 6>You have an incredible amount of power at this moment

0:07:21.120 --> 0:07:21.480
<v Speaker 6>in time.

0:07:21.680 --> 0:07:24.200
<v Speaker 2>Why should we trust you?

0:07:24.200 --> 0:07:27.560
<v Speaker 3>You shouldn't If this really works, it's like quite a

0:07:27.560 --> 0:07:30.560
<v Speaker 3>powerful technology, and you should not trust one company, and

0:07:30.640 --> 0:07:31.960
<v Speaker 3>certainly not one personalize it.

0:07:33.800 --> 0:07:38.080
<v Speaker 1>You shouldn't trust me. If he believes that, then why

0:07:38.120 --> 0:07:40.560
<v Speaker 1>did Sam fight his way right back to the top

0:07:40.600 --> 0:07:44.320
<v Speaker 1>of Open AI, back into that position of control, as

0:07:44.400 --> 0:07:46.960
<v Speaker 1>if to insist he should be the person to lead

0:07:47.040 --> 0:07:50.280
<v Speaker 1>this company as they develop AI as fast as possible.

0:07:51.560 --> 0:07:54.560
<v Speaker 1>To a lot of people, the stakes are terrifying. They're

0:07:54.640 --> 0:07:58.200
<v Speaker 1>higher than anything else really in the valley. Some people

0:07:58.240 --> 0:08:03.000
<v Speaker 1>refer to Artificial General intelligen or AGI as the last invention,

0:08:03.760 --> 0:08:07.160
<v Speaker 1>something so potent that everything that comes after will look

0:08:07.360 --> 0:08:11.640
<v Speaker 1>unrecognizable compared to everything leading up to it. Inside Open AI,

0:08:11.920 --> 0:08:15.000
<v Speaker 1>some employees talk about how they are building God.

0:08:16.560 --> 0:08:17.680
<v Speaker 2>We'll be right back.

0:08:23.680 --> 0:08:25.600
<v Speaker 1>To get a better sense of the kind of person

0:08:25.720 --> 0:08:28.160
<v Speaker 1>Sam is and how he got where he is now.

0:08:28.640 --> 0:08:32.000
<v Speaker 1>I want to take you back to his adolescence. Sam

0:08:32.040 --> 0:08:35.240
<v Speaker 1>had a privileged upbringing in Saint Louis. He's the oldest

0:08:35.240 --> 0:08:38.440
<v Speaker 1>of four siblings. His mom was a dermatologist and his

0:08:38.559 --> 0:08:41.720
<v Speaker 1>dad was a real estate developer. He attended a private

0:08:41.760 --> 0:08:45.040
<v Speaker 1>high school called John Burrows. There's an anecdote about him

0:08:45.040 --> 0:08:48.120
<v Speaker 1>from that period that sticks out. When some students wanted

0:08:48.160 --> 0:08:51.640
<v Speaker 1>to boycott an assembly about sexuality, Sam stood up in

0:08:51.640 --> 0:08:53.880
<v Speaker 1>front of the whole school and announced he was gay.

0:08:54.520 --> 0:08:56.560
<v Speaker 1>It's a pretty gutsy move for a teenager in the

0:08:56.559 --> 0:09:00.800
<v Speaker 1>early two thousands, and unsurprisingly, Sam smart.

0:09:01.280 --> 0:09:06.120
<v Speaker 7>And generally Sam he was. He was an exceptional student,

0:09:06.320 --> 0:09:12.360
<v Speaker 7>He was an exceptional writer, he was an exceptional big personality.

0:09:13.160 --> 0:09:16.200
<v Speaker 1>That's Andy Abbott. He was one of Sam's English teachers,

0:09:16.240 --> 0:09:18.800
<v Speaker 1>and he's now the head of school. And this is

0:09:18.840 --> 0:09:21.600
<v Speaker 1>a pretty nerdy school where it's cool to get good

0:09:21.640 --> 0:09:25.240
<v Speaker 1>grades and be a high achiever. And even in that environment,

0:09:25.679 --> 0:09:26.600
<v Speaker 1>Sam stood out.

0:09:26.880 --> 0:09:34.280
<v Speaker 7>Sam's just a really natural leader, incredibly charismatic, curious guy.

0:09:34.520 --> 0:09:39.440
<v Speaker 7>He's a typical you know. He was the editor of

0:09:39.480 --> 0:09:44.080
<v Speaker 7>the yearbook and he represented the school in the Model

0:09:44.200 --> 0:09:48.080
<v Speaker 7>United Nations. He designed our website, you know, before we

0:09:48.600 --> 0:09:51.440
<v Speaker 7>hired people to do our website. He could just do

0:09:51.520 --> 0:09:52.120
<v Speaker 7>that stuff.

0:09:52.559 --> 0:09:53.959
<v Speaker 2>Sam even played water polow.

0:09:54.240 --> 0:09:58.760
<v Speaker 7>He was pretty good. I'm not a connoisseur, but I'm like,

0:09:58.840 --> 0:09:59.720
<v Speaker 7>he was pretty good.

0:10:00.360 --> 0:10:03.840
<v Speaker 1>He remembers Sam as being really confident, and apparently for

0:10:03.920 --> 0:10:07.400
<v Speaker 1>good reason. Sounds like Sam was just this exceptional kid.

0:10:07.800 --> 0:10:13.640
<v Speaker 7>Well, he's the smartest guy in the room, and he's charismatic.

0:10:14.200 --> 0:10:17.199
<v Speaker 7>I remember thinking, and I'm just this is just an

0:10:17.240 --> 0:10:22.400
<v Speaker 7>embarrassing confession. I hope he doesn't go into technology. He's

0:10:22.440 --> 0:10:26.000
<v Speaker 7>so creative and he's so, he's such a good writer,

0:10:26.160 --> 0:10:30.199
<v Speaker 7>and I hoped he would be an author or something

0:10:30.320 --> 0:10:34.280
<v Speaker 7>like that. And I mean, nobody could have anticipated the

0:10:34.320 --> 0:10:42.960
<v Speaker 7>magnitude of open Ai, but everybody knew that this guy's

0:10:43.000 --> 0:10:45.320
<v Speaker 7>better at most things than most of us are.

0:10:46.200 --> 0:10:49.160
<v Speaker 1>This speaks to a pattern that'll become a crucial factor

0:10:49.200 --> 0:10:54.160
<v Speaker 1>in Sam's career. He's very good at impressing people, especially

0:10:54.240 --> 0:10:58.360
<v Speaker 1>the right people, older people, people with influence, people who

0:10:58.400 --> 0:11:01.440
<v Speaker 1>are at a position to help him. Someone who knows

0:11:01.440 --> 0:11:05.240
<v Speaker 1>Sam says his superpower is figuring out who's in charge

0:11:05.480 --> 0:11:10.199
<v Speaker 1>and charming them. So we have young Sam. Even though

0:11:10.200 --> 0:11:12.719
<v Speaker 1>he was a teenager, he acted like someone older with

0:11:12.760 --> 0:11:16.880
<v Speaker 1>more agency and confidence. Adults found this quality of his admirable,

0:11:17.480 --> 0:11:21.280
<v Speaker 1>and he acted like this toward his three younger siblings too.

0:11:21.520 --> 0:11:24.280
<v Speaker 1>In a big New Yorker profile on Sam, his younger

0:11:24.320 --> 0:11:26.480
<v Speaker 1>brother said that as kids, they used to play a

0:11:26.480 --> 0:11:30.120
<v Speaker 1>board game called Samurai, and Sam always won because he

0:11:30.240 --> 0:11:33.160
<v Speaker 1>declared himself the leader and said, I have to win

0:11:33.320 --> 0:11:37.040
<v Speaker 1>and I'm in charge of everything. When Sam's brother told

0:11:37.040 --> 0:11:40.640
<v Speaker 1>this story, it was a jocular exchange. But Annie, their

0:11:40.720 --> 0:11:44.960
<v Speaker 1>youngest sibling and only sister, sees it differently. These days,

0:11:45.040 --> 0:11:47.800
<v Speaker 1>she's estranged from Sam and the rest of her immediate family,

0:11:48.400 --> 0:11:51.000
<v Speaker 1>but when she was a kid, she remembered that same

0:11:51.120 --> 0:11:54.640
<v Speaker 1>quality of Sam's wanting to be in charge, and to

0:11:54.720 --> 0:11:57.240
<v Speaker 1>her it wasn't funny, it was domineering.

0:11:58.559 --> 0:12:03.760
<v Speaker 8>From my perspective, with nine year age difference, he very

0:12:03.840 --> 0:12:06.600
<v Speaker 8>much wanted to be and acted like the third parent

0:12:07.960 --> 0:12:12.360
<v Speaker 8>and like being the older sibling in charge, in control.

0:12:13.360 --> 0:12:16.240
<v Speaker 1>For instance, even though the family was Jewish, they used

0:12:16.240 --> 0:12:19.600
<v Speaker 1>to get a Christmas tree until Sam put his foot down.

0:12:20.360 --> 0:12:23.520
<v Speaker 8>I don't have memories of Christmas tree because when Sam

0:12:23.600 --> 0:12:29.319
<v Speaker 8>got farmits food at thirteen, he decided that we as

0:12:29.320 --> 0:12:32.360
<v Speaker 8>a family unit were Jews and needed to no longer

0:12:32.400 --> 0:12:33.360
<v Speaker 8>celebrate Christmas.

0:12:33.880 --> 0:12:35.280
<v Speaker 2>There were no more Christmas trees.

0:12:36.080 --> 0:12:39.680
<v Speaker 1>When their dad passed away in twenty eighteen, Annie remembers

0:12:39.720 --> 0:12:42.600
<v Speaker 1>that Sam dictated to each of his younger siblings how

0:12:42.640 --> 0:12:44.600
<v Speaker 1>many minutes they could talk at the funeral.

0:12:45.160 --> 0:12:47.600
<v Speaker 8>To be at your dad's funeral, to be like, oh,

0:12:47.640 --> 0:12:49.679
<v Speaker 8>I'm the oldest sibling, so I get to choose how

0:12:49.679 --> 0:12:52.080
<v Speaker 8>long all the sibling which it is bizarre, and there's

0:12:52.120 --> 0:12:55.960
<v Speaker 8>a level of it that's so hilarious and so benign.

0:12:56.200 --> 0:13:01.240
<v Speaker 8>Surface level, classic older sibling bullshit where it's like, all right,

0:13:01.320 --> 0:13:03.160
<v Speaker 8>older sibling wanting to make.

0:13:03.000 --> 0:13:04.040
<v Speaker 2>Up the rules to the game.

0:13:04.280 --> 0:13:07.040
<v Speaker 8>Like there's a level of it that's very light and funny,

0:13:07.720 --> 0:13:11.120
<v Speaker 8>and there's also a level of it that's very dark

0:13:11.160 --> 0:13:14.719
<v Speaker 8>and deeply unsettling of how does that behavior come up

0:13:14.760 --> 0:13:17.520
<v Speaker 8>in other places if you believe that you get to

0:13:17.520 --> 0:13:20.880
<v Speaker 8>be the authority on something that you are not the

0:13:20.960 --> 0:13:21.640
<v Speaker 8>authority on.

0:13:22.600 --> 0:13:25.640
<v Speaker 1>A spokeswoman for open Ai told us that Sam recalls

0:13:25.640 --> 0:13:31.400
<v Speaker 1>these incidents differently, but she declined to elaborate. And so

0:13:31.720 --> 0:13:34.520
<v Speaker 1>when Sam finished high school, he started down this path

0:13:34.559 --> 0:13:38.160
<v Speaker 1>that's pretty textbook for the tech industry, studying computer science

0:13:38.160 --> 0:13:41.760
<v Speaker 1>at Stanford, founding a startup, and dropping out of Stanford,

0:13:42.480 --> 0:13:46.200
<v Speaker 1>and he made one incredibly important decision he applied to

0:13:46.280 --> 0:13:51.160
<v Speaker 1>why Combinator. Why Combinator is a startup accelerator. It's basically

0:13:51.200 --> 0:13:54.079
<v Speaker 1>a boot camp for startups. You and your co founders

0:13:54.080 --> 0:13:57.560
<v Speaker 1>apply and if you're accepted, you spend three months hacking

0:13:57.600 --> 0:14:00.120
<v Speaker 1>away trying to build a company. At the end of

0:14:00.160 --> 0:14:02.560
<v Speaker 1>that period, you give a demo to investors and try

0:14:02.559 --> 0:14:06.000
<v Speaker 1>to raise venture capital. Sam was actually in the first

0:14:06.080 --> 0:14:09.559
<v Speaker 1>ever group of founders at y Combinator. Everyone calls it YC.

0:14:09.800 --> 0:14:12.560
<v Speaker 1>By the way, it was two thousand and five, so

0:14:12.960 --> 0:14:15.720
<v Speaker 1>YC was totally unknown, just a bunch of young guys

0:14:15.840 --> 0:14:19.120
<v Speaker 1>hanging out in Cambridge, Massachusetts for a summer writing code.

0:14:19.680 --> 0:14:24.160
<v Speaker 1>But WYC would eventually become this enormously powerful network. Now

0:14:24.200 --> 0:14:27.560
<v Speaker 1>it's basically the number one elite program for startups. It's

0:14:27.680 --> 0:14:30.200
<v Speaker 1>really hard to get in, and the alumni network is

0:14:30.320 --> 0:14:34.440
<v Speaker 1>incredibly strong. Sam was nineteen years old when he joined YC,

0:14:35.200 --> 0:14:38.640
<v Speaker 1>and once again he impressed just the right person, Paul Graham,

0:14:38.680 --> 0:14:42.040
<v Speaker 1>the head of YC. The first batch included some other

0:14:42.120 --> 0:14:45.320
<v Speaker 1>really impressive people, like the founders of Reddit and Twitch,

0:14:45.840 --> 0:14:49.000
<v Speaker 1>but someone who knows Sam says that he was immediately

0:14:49.080 --> 0:14:53.920
<v Speaker 1>Paul's favorite. Paul later wrote, within about three minutes of

0:14:53.960 --> 0:14:56.640
<v Speaker 1>meeting him, I remember thinking, ah, so this is what

0:14:56.760 --> 0:14:58.920
<v Speaker 1>Bill Gates must have been like when he was nineteen.

0:15:00.000 --> 0:15:03.480
<v Speaker 1>The startup Sam was building was called Looped. Looped is

0:15:03.520 --> 0:15:06.000
<v Speaker 1>one of the forgotten apps of its time the early

0:15:06.040 --> 0:15:10.040
<v Speaker 1>two thousands, when people were really excited about having GPS

0:15:10.200 --> 0:15:13.040
<v Speaker 1>on their phones for the first time. It used location

0:15:13.200 --> 0:15:15.800
<v Speaker 1>data to connect people to their friends and local businesses,

0:15:16.400 --> 0:15:18.560
<v Speaker 1>kind of like a mix of Yelp, and four Square.

0:15:19.560 --> 0:15:22.120
<v Speaker 1>Here's Sam making the pitch at a developers conference.

0:15:22.520 --> 0:15:25.160
<v Speaker 9>Luped is about connecting with people on the go, which is,

0:15:25.280 --> 0:15:27.360
<v Speaker 9>after all, the main reason you have a phone. We

0:15:27.440 --> 0:15:30.000
<v Speaker 9>show you where people are, what they're doing, and what

0:15:30.120 --> 0:15:31.200
<v Speaker 9>cool places are around you.

0:15:32.080 --> 0:15:34.680
<v Speaker 1>Sam started building the startup in two thousand and five.

0:15:35.240 --> 0:15:38.600
<v Speaker 1>The iPhone didn't exist yet, so Looped was trying to

0:15:38.600 --> 0:15:41.600
<v Speaker 1>do this for flip phones and it was kind of

0:15:41.600 --> 0:15:47.560
<v Speaker 1>hard to get traction. At one point early on, Sam's

0:15:47.600 --> 0:15:50.760
<v Speaker 1>company was in a desperate situation. They really needed to

0:15:50.760 --> 0:15:53.440
<v Speaker 1>get a deal with a mobile carrier. They learned that

0:15:53.520 --> 0:15:56.240
<v Speaker 1>Boost Mobile, which was part of Sprint, was looking to

0:15:56.240 --> 0:15:59.280
<v Speaker 1>add a location feature and needed a partner, but they

0:15:59.280 --> 0:16:03.360
<v Speaker 1>were about to sign with someone else, so Sam flew

0:16:03.400 --> 0:16:07.320
<v Speaker 1>down to Boost's headquarters in Irvine in southern California. When

0:16:07.360 --> 0:16:09.880
<v Speaker 1>he tells the story, he says that he just showed up,

0:16:10.160 --> 0:16:13.200
<v Speaker 1>waited outside the right executive's office, and asked for just

0:16:13.280 --> 0:16:16.640
<v Speaker 1>ten minutes. Here's how that executive remembers it.

0:16:16.640 --> 0:16:20.000
<v Speaker 6>It's ever recall I got a phone call from Sam

0:16:20.520 --> 0:16:23.920
<v Speaker 6>and he was in Irvine, and he said that he

0:16:24.000 --> 0:16:27.760
<v Speaker 6>explained who he was and what Looped was somebody at

0:16:27.800 --> 0:16:31.080
<v Speaker 6>Sprint had told him to get in touch with us.

0:16:31.360 --> 0:16:34.600
<v Speaker 1>That's Lowell Winer. He was at the time the head

0:16:34.640 --> 0:16:37.400
<v Speaker 1>of business development for Boost. And he's going to tell

0:16:37.440 --> 0:16:40.200
<v Speaker 1>a story that has a few asides, but that I

0:16:40.200 --> 0:16:42.520
<v Speaker 1>think captures a lot about what Sam was like.

0:16:42.640 --> 0:16:46.560
<v Speaker 6>Early on, we were a day or two away from

0:16:46.640 --> 0:16:51.040
<v Speaker 6>signing a contract with another startup that was further along

0:16:51.160 --> 0:16:54.360
<v Speaker 6>than Looped. He asked to come by that day, you know,

0:16:54.400 --> 0:16:58.720
<v Speaker 6>which is incredibly unusual, but given the timing that, you know,

0:16:58.720 --> 0:17:00.160
<v Speaker 6>we were at the eleventh hour we were at to

0:17:00.200 --> 0:17:03.760
<v Speaker 6>sign this contract. He had come, you know, referred to

0:17:03.840 --> 0:17:06.840
<v Speaker 6>us by our parent company, it was worth at least

0:17:06.840 --> 0:17:11.000
<v Speaker 6>a meeting. So Sam shows up at the office with

0:17:11.040 --> 0:17:13.800
<v Speaker 6>one of one or two other guys from Looped. We

0:17:13.880 --> 0:17:16.560
<v Speaker 6>go sit in the conference room, you know, we share

0:17:16.600 --> 0:17:18.880
<v Speaker 6>what we were looking to do. Sam started to share

0:17:18.880 --> 0:17:22.480
<v Speaker 6>about Looped. He was I think nineteen at the time,

0:17:23.840 --> 0:17:26.200
<v Speaker 6>you know, I think maybe in cargo short sitting cross

0:17:26.280 --> 0:17:29.200
<v Speaker 6>legged in a in a chair in the conference room

0:17:29.920 --> 0:17:32.000
<v Speaker 6>and just kind of holding court.

0:17:32.280 --> 0:17:34.120
<v Speaker 1>I want to pause here for a second on this

0:17:34.280 --> 0:17:38.200
<v Speaker 1>cargo shorts detail. For a lot of Sam's young life,

0:17:38.280 --> 0:17:41.680
<v Speaker 1>he was a cargo shorts. Devotee wore them all the time.

0:17:42.160 --> 0:17:44.399
<v Speaker 1>People kind of poked fun at him for it, to

0:17:44.440 --> 0:17:46.639
<v Speaker 1>the point where he felt the need to address it

0:17:46.760 --> 0:17:49.000
<v Speaker 1>on a podcast called Masters of Scale.

0:17:49.520 --> 0:17:51.880
<v Speaker 3>Honestly, I don't think they're that ugly, and I find

0:17:51.880 --> 0:17:55.840
<v Speaker 3>them incredibly convenient. Like I I, you can like put

0:17:55.880 --> 0:17:57.320
<v Speaker 3>a lot of stuff like I like to. I'd still

0:17:57.359 --> 0:17:59.960
<v Speaker 3>read paperback books. I like paperback books. I like to

0:18:00.000 --> 0:18:03.680
<v Speaker 3>carry on around with me. I have like an iPhone

0:18:04.119 --> 0:18:06.320
<v Speaker 3>seven plus, which is kind of like works really well

0:18:06.320 --> 0:18:09.800
<v Speaker 3>in cargo pockets. I carry like computer chargers, cables. They're

0:18:09.840 --> 0:18:14.280
<v Speaker 3>just like, you know, efficient. Why people care about that

0:18:14.280 --> 0:18:15.160
<v Speaker 3>so much that I can't.

0:18:15.000 --> 0:18:19.840
<v Speaker 1>Tell you that last comment? That's very Sam to remark

0:18:19.880 --> 0:18:23.000
<v Speaker 1>that the things normal people might talk about don't make

0:18:23.080 --> 0:18:27.040
<v Speaker 1>sense or aren't rational. It's like he has no patience

0:18:27.080 --> 0:18:29.840
<v Speaker 1>for the things most of us might think are funny.

0:18:29.880 --> 0:18:33.880
<v Speaker 1>He has more important things to think about anyway. Here

0:18:33.920 --> 0:18:36.000
<v Speaker 1>he is a nineteen year old in a meeting with

0:18:36.080 --> 0:18:40.280
<v Speaker 1>mobile network executives wearing cargo shorts, sitting cross legged in

0:18:40.320 --> 0:18:43.680
<v Speaker 1>a conference room chair. Even though this encounter was almost

0:18:43.760 --> 0:18:47.200
<v Speaker 1>twenty years ago, Lowell remembers vividly what Sam looked like

0:18:47.240 --> 0:18:49.639
<v Speaker 1>in that moment, because it was such an odd picture.

0:18:50.160 --> 0:18:56.639
<v Speaker 6>He's a he was small in stature. I don't think

0:18:56.680 --> 0:18:59.159
<v Speaker 6>he's a big guy now, but he was. You know,

0:18:59.240 --> 0:19:02.600
<v Speaker 6>he was quite slim at the time. You know, it's

0:19:02.600 --> 0:19:05.760
<v Speaker 6>no easy feat to sit cross legged in a conference chair.

0:19:05.880 --> 0:19:07.719
<v Speaker 6>I mean, he looked like he could have still been

0:19:07.720 --> 0:19:08.359
<v Speaker 6>in high school.

0:19:08.760 --> 0:19:11.680
<v Speaker 1>I've heard other people describe Sam's weird way of sitting.

0:19:12.280 --> 0:19:14.160
<v Speaker 1>He's older now, so he doesn't do it as much,

0:19:14.600 --> 0:19:17.000
<v Speaker 1>but one person who knows him told me he used

0:19:17.040 --> 0:19:19.359
<v Speaker 1>to squat on the seat of a chair like a

0:19:19.400 --> 0:19:25.040
<v Speaker 1>perched bird, with his knees up toward his chin. Despite

0:19:25.040 --> 0:19:28.440
<v Speaker 1>this unconventional way of presenting himself or because of it,

0:19:29.080 --> 0:19:32.720
<v Speaker 1>he almost immediately convinced Lowell that boost Mobile should switch

0:19:32.760 --> 0:19:35.440
<v Speaker 1>plans at the eleventh hour and go with this other partner.

0:19:35.840 --> 0:19:38.320
<v Speaker 6>It was pretty clear within a half hour of this

0:19:38.400 --> 0:19:41.800
<v Speaker 6>meeting starting that Luked and Sam were the right partner

0:19:41.880 --> 0:19:47.240
<v Speaker 6>to do this. There was both excitement and like, shit,

0:19:47.560 --> 0:19:50.719
<v Speaker 6>we have to go sell it internally. But I recall

0:19:50.840 --> 0:19:53.520
<v Speaker 6>stepping outside of the room with the colleague that I

0:19:53.560 --> 0:19:55.960
<v Speaker 6>was in the meeting with and saying to him, we

0:19:56.040 --> 0:19:57.200
<v Speaker 6>need to switch gears.

0:19:57.440 --> 0:20:00.800
<v Speaker 1>And Lowell still remembers this way that Sam looks unassuming,

0:20:01.400 --> 0:20:01.919
<v Speaker 1>but he's not.

0:20:03.520 --> 0:20:09.000
<v Speaker 6>On the one hand, visually he looked incredibly young, but

0:20:09.840 --> 0:20:12.080
<v Speaker 6>if you'd shut your eyes and were just listening to him,

0:20:12.280 --> 0:20:14.359
<v Speaker 6>his command of material and his ability to communicate and

0:20:14.359 --> 0:20:17.959
<v Speaker 6>engage was on par with anyone you know I had

0:20:18.000 --> 0:20:21.680
<v Speaker 6>met with over the course of my tech career. It Uh,

0:20:21.760 --> 0:20:26.879
<v Speaker 6>it was freakish. Yeah, it was screakish, and not his appearance,

0:20:27.080 --> 0:20:30.840
<v Speaker 6>but his his poise and command for.

0:20:30.840 --> 0:20:35.640
<v Speaker 1>That age, Sam's relentlessness paid off. He knew he had

0:20:35.640 --> 0:20:37.879
<v Speaker 1>to get that deal, and he did what he needed

0:20:37.880 --> 0:20:41.000
<v Speaker 1>to do to get it, including flying across the state

0:20:41.160 --> 0:20:45.040
<v Speaker 1>to surprise someone at their office. Sam later said he

0:20:45.119 --> 0:20:48.560
<v Speaker 1>learned an important lesson the way to get things done

0:20:48.840 --> 0:20:54.639
<v Speaker 1>is to just be really persistent. So he inks this

0:20:54.720 --> 0:20:57.720
<v Speaker 1>deal for Looped, it's going to power location sharing for

0:20:57.800 --> 0:21:00.720
<v Speaker 1>boost Mobile. It's the partnership that led to this two

0:21:00.720 --> 0:21:02.120
<v Speaker 1>thousand and six ad campaign.

0:21:02.560 --> 0:21:04.679
<v Speaker 6>Oh you see where I meant here, but I know

0:21:04.760 --> 0:21:08.000
<v Speaker 6>where you at Poop Slooped the GPS.

0:21:08.040 --> 0:21:09.640
<v Speaker 7>Now you know where your friends.

0:21:09.440 --> 0:21:13.160
<v Speaker 1>At Sam's peers at YC were surprised that he pulled

0:21:13.160 --> 0:21:16.399
<v Speaker 1>this off. Lup's business model was pretty wobbly and the

0:21:16.440 --> 0:21:20.520
<v Speaker 1>product wasn't all that impressive, but Sam's particular strength was

0:21:20.560 --> 0:21:21.639
<v Speaker 1>starting to become clearer.

0:21:21.960 --> 0:21:24.080
<v Speaker 4>It was obvious Sam was an incredible deals guy.

0:21:24.440 --> 0:21:27.000
<v Speaker 1>That's a clip from a podcast interview with Emmett Sheer,

0:21:27.320 --> 0:21:30.119
<v Speaker 1>one of the other guys in that very first YC batch.

0:21:30.760 --> 0:21:33.639
<v Speaker 1>You can hear how impressed he was by Sam's particular talent.

0:21:34.200 --> 0:21:38.040
<v Speaker 4>He was somehow convincing the phone companies to give his

0:21:38.520 --> 0:21:42.040
<v Speaker 4>startup that like didn't really have a product like deals.

0:21:42.160 --> 0:21:44.600
<v Speaker 4>I still don't know how he did that, And like

0:21:44.840 --> 0:21:47.080
<v Speaker 4>that was the only obvious thing about Sam at the time,

0:21:47.240 --> 0:21:50.280
<v Speaker 4>was like that he was ambitious, but most of us

0:21:50.320 --> 0:21:53.320
<v Speaker 4>were pretty ambitious, and he was great at great great

0:21:53.320 --> 0:21:53.760
<v Speaker 4>deals guy.

0:21:54.400 --> 0:21:58.120
<v Speaker 1>For the next few years, luped kept growing. Sam presented

0:21:58.160 --> 0:22:01.080
<v Speaker 1>at Apple's Developer Conference into thousand and eight in a

0:22:01.200 --> 0:22:04.840
<v Speaker 1>very unique outfit. He wore two polo shirts layered on

0:22:04.880 --> 0:22:07.280
<v Speaker 1>top of each other, so he's in a hot pink

0:22:07.320 --> 0:22:10.840
<v Speaker 1>polo with a second lime green collar poking out underneath.

0:22:11.040 --> 0:22:16.440
<v Speaker 4>I'd like to invite up Sam Altman, Sam K.

0:22:16.440 --> 0:22:23.240
<v Speaker 9>Scott, we are incredibly psyched about lupt on the iPhone.

0:22:23.400 --> 0:22:26.000
<v Speaker 9>Luped is about connecting with people and what cool places

0:22:26.040 --> 0:22:28.520
<v Speaker 9>are around you. The orange pin up there is where

0:22:28.520 --> 0:22:31.200
<v Speaker 9>I am right now, and the blue pins represent my friends.

0:22:31.920 --> 0:22:33.280
<v Speaker 9>We make serendipity happen.

0:22:33.760 --> 0:22:36.520
<v Speaker 1>And when you listen to Sam pitching Loot, you can

0:22:36.560 --> 0:22:41.120
<v Speaker 1>hear this other part of him, this earnestness and optimism.

0:22:41.480 --> 0:22:43.639
<v Speaker 1>I've listened to and read a lot of interviews with

0:22:43.680 --> 0:22:47.080
<v Speaker 1>Sam and he's always using the words super and excited,

0:22:47.520 --> 0:22:49.280
<v Speaker 1>sometimes super excited.

0:22:49.480 --> 0:22:52.400
<v Speaker 3>It's super cool, super easy to make, super important. It's

0:22:52.440 --> 0:22:53.160
<v Speaker 3>been super great.

0:22:53.240 --> 0:22:54.640
<v Speaker 9>I'm super excited to announce.

0:22:54.800 --> 0:22:57.159
<v Speaker 3>I'm super excited, super excited for that.

0:22:58.960 --> 0:23:02.440
<v Speaker 1>But he was not super excited about what happened next

0:23:02.520 --> 0:23:06.880
<v Speaker 1>at lut. After several years, Looped fizzled out. Sam made

0:23:06.920 --> 0:23:09.160
<v Speaker 1>a deal to sell the company for a modest sum

0:23:09.280 --> 0:23:12.359
<v Speaker 1>in twenty twelve. He walked away with a reported five

0:23:12.400 --> 0:23:16.400
<v Speaker 1>million dollars. Most people would be pretty happy with this outcome,

0:23:16.680 --> 0:23:19.359
<v Speaker 1>but in Silicon Valley terms, Loop twa was kind of

0:23:19.359 --> 0:23:23.320
<v Speaker 1>a failure. But that's okay because by then Sam had

0:23:23.320 --> 0:23:26.560
<v Speaker 1>won over other people who could help him. One was

0:23:26.640 --> 0:23:29.720
<v Speaker 1>Peter Teel, who was a billionaire, an investor, and the

0:23:29.760 --> 0:23:33.240
<v Speaker 1>co founder of PayPal and Palanteer. He's also one of

0:23:33.240 --> 0:23:36.000
<v Speaker 1>the most powerful gay men in Silicon Valley, which lent

0:23:36.080 --> 0:23:39.959
<v Speaker 1>him and Sam a sense of camaraderie. When Sam left looped,

0:23:40.240 --> 0:23:43.199
<v Speaker 1>Peter gave him a bunch of money to invest. They

0:23:43.200 --> 0:23:47.399
<v Speaker 1>were close, and Sam's peers noticed because Peter Teel is

0:23:47.480 --> 0:23:52.920
<v Speaker 1>notoriously pessimistic and even nihilistic, and Sam's public image, by contrast,

0:23:53.359 --> 0:23:56.920
<v Speaker 1>is pretty earnest and optimistic. At the same time, Sam

0:23:57.040 --> 0:23:59.800
<v Speaker 1>was deepening his relationship with Paul Graham, the head of YC,

0:24:00.840 --> 0:24:05.160
<v Speaker 1>and that closeness was giving Sam tangible benefits. When Paul

0:24:05.240 --> 0:24:07.600
<v Speaker 1>had the chance to invest very early on in Stripe

0:24:07.640 --> 0:24:12.040
<v Speaker 1>the Payments startup, he invited Sam to invest too. Sam

0:24:12.119 --> 0:24:15.080
<v Speaker 1>later said that that was by some measures, his most

0:24:15.200 --> 0:24:19.399
<v Speaker 1>profitable angel investment ever, and he got it purely because

0:24:19.440 --> 0:24:22.880
<v Speaker 1>of this personal relationship he had built. Paul is known

0:24:22.880 --> 0:24:26.359
<v Speaker 1>for writing essays about how to build startups, essays filled

0:24:26.400 --> 0:24:31.040
<v Speaker 1>with blunt, quotable entrepreneurial wisdom. Often in those essays he

0:24:31.160 --> 0:24:35.200
<v Speaker 1>praised Sam. He advised young eager founders to emulate Sam,

0:24:35.800 --> 0:24:38.040
<v Speaker 1>and Paul is also responsible for one of the most

0:24:38.040 --> 0:24:42.479
<v Speaker 1>infamous quotes about Sam. You could parachute him into an

0:24:42.480 --> 0:24:45.600
<v Speaker 1>island full of cannibals and come back in five years

0:24:45.960 --> 0:24:52.720
<v Speaker 1>and he'd be the king. At first, this quote sounded

0:24:52.760 --> 0:24:56.040
<v Speaker 1>to me like a compliment, but lately I've wondered if

0:24:56.040 --> 0:25:00.639
<v Speaker 1>maybe it's not anyway. Around twenty twelve, Why had become

0:25:00.800 --> 0:25:04.560
<v Speaker 1>very influential. It had relocated its headquarters from Cambridge to

0:25:04.600 --> 0:25:07.320
<v Speaker 1>Silicon Valley and was now a breeding ground for some

0:25:07.359 --> 0:25:12.040
<v Speaker 1>of the most successful Internet businesses. Airbnb, Dropbox, and Stripe

0:25:12.119 --> 0:25:15.600
<v Speaker 1>all got their start at YC, and from the outside

0:25:15.800 --> 0:25:18.560
<v Speaker 1>it looked as though Sam was beginning to mimic Paul

0:25:18.720 --> 0:25:23.960
<v Speaker 1>as a startup guru. He was also advising young founders

0:25:23.960 --> 0:25:27.480
<v Speaker 1>at YC, and just like Paul, he started writing essays

0:25:27.880 --> 0:25:32.600
<v Speaker 1>filled with mysterious, often perplexing advice for startup founders, such

0:25:32.640 --> 0:25:36.240
<v Speaker 1>as the most successful founders do not set out to

0:25:36.280 --> 0:25:39.359
<v Speaker 1>create companies. They are on a mission to create something

0:25:39.400 --> 0:25:44.520
<v Speaker 1>closer to a religion or here's another A big secret

0:25:44.680 --> 0:25:46.600
<v Speaker 1>is that you can bend the world to your will

0:25:46.680 --> 0:25:50.359
<v Speaker 1>a surprising percentage at the time, most people don't even try.

0:25:51.760 --> 0:25:55.080
<v Speaker 1>In a blog post called how to Be Successful, Sam

0:25:55.119 --> 0:25:58.720
<v Speaker 1>told founders that they should quote have almost too much

0:25:58.760 --> 0:26:02.200
<v Speaker 1>self belief. The most successful people I know believe in

0:26:02.280 --> 0:26:08.160
<v Speaker 1>themselves almost to the point of delusion. That last point

0:26:08.240 --> 0:26:11.159
<v Speaker 1>will sound familiar to people who know Sam. One of

0:26:11.160 --> 0:26:15.080
<v Speaker 1>them told me that Sam has complete self belief. That

0:26:15.160 --> 0:26:17.920
<v Speaker 1>Sam gives off the impression that he believes in himself

0:26:18.000 --> 0:26:21.840
<v Speaker 1>one hundred percent without that nagging feeling most of us have,

0:26:22.080 --> 0:26:25.760
<v Speaker 1>like a little voice of fear or uncertainty. And here's

0:26:25.800 --> 0:26:28.280
<v Speaker 1>Lowell again, the former boost Mobile executive.

0:26:29.600 --> 0:26:35.000
<v Speaker 6>He was extraordinarily self assured, and not in an egotistical way,

0:26:35.680 --> 0:26:42.600
<v Speaker 6>but just very comfortable with himself, with his capacity both

0:26:42.640 --> 0:26:44.119
<v Speaker 6>intellectually and relationally.

0:26:44.920 --> 0:26:48.239
<v Speaker 1>For years, Paul had built up Sam's image, made him

0:26:48.280 --> 0:26:52.119
<v Speaker 1>this kind of startup demigod. Then he decided to anoint

0:26:52.240 --> 0:26:56.200
<v Speaker 1>him fully. Paul stepped down in twenty fourteen and named

0:26:56.240 --> 0:26:57.800
<v Speaker 1>Sam President of YC.

0:26:58.760 --> 0:26:59.640
<v Speaker 2>This was a big deal.

0:27:00.280 --> 0:27:03.199
<v Speaker 1>YC is the center of Silicon Valley, and now Sam

0:27:03.400 --> 0:27:07.080
<v Speaker 1>was its leader. He was twenty eight years old. With

0:27:07.119 --> 0:27:11.320
<v Speaker 1>both Peter Teel and Paul Graham, Sam cultivated these close

0:27:11.359 --> 0:27:14.920
<v Speaker 1>relationships with people and positions of power. Then they gave

0:27:15.000 --> 0:27:20.639
<v Speaker 1>him things like money, connections, influence titles. Essentially, they transferred

0:27:20.640 --> 0:27:24.240
<v Speaker 1>some of their power directly to him, and that in

0:27:24.320 --> 0:27:27.920
<v Speaker 1>turn gave Sam the ability to think big. Even when

0:27:27.960 --> 0:27:30.600
<v Speaker 1>he was working on something pretty silly sounding like looped,

0:27:31.119 --> 0:27:36.040
<v Speaker 1>Sam had sky high ambitions. Paul noticed this. Here he

0:27:36.119 --> 0:27:38.720
<v Speaker 1>is at a conference explaining why he picked Sam to

0:27:38.760 --> 0:27:40.120
<v Speaker 1>be his successor at YC.

0:27:41.119 --> 0:27:43.520
<v Speaker 3>It's turned into this giant thing and I'm no good

0:27:43.520 --> 0:27:44.720
<v Speaker 3>at running giant things.

0:27:45.160 --> 0:27:48.000
<v Speaker 4>Sam, however, is going to be good at running a

0:27:48.040 --> 0:27:48.560
<v Speaker 4>giant thing.

0:27:48.880 --> 0:27:51.240
<v Speaker 1>At this point, there was no evidence Sam could run

0:27:51.280 --> 0:27:54.240
<v Speaker 1>a giant thing or even a medium sized thing. But

0:27:54.320 --> 0:27:57.560
<v Speaker 1>once he took over, Sam did make YC much bigger

0:27:57.600 --> 0:28:01.919
<v Speaker 1>than before sprawling. He gave more money to more startups,

0:28:02.119 --> 0:28:03.560
<v Speaker 1>and he expanded overseas.

0:28:04.359 --> 0:28:08.720
<v Speaker 10>The Sam Altman era at YC was about expansion.

0:28:09.320 --> 0:28:12.320
<v Speaker 1>That's John Coogan. He's a startup founder who went through

0:28:12.440 --> 0:28:16.480
<v Speaker 1>YC twice, including ones for his company Soylent, you know,

0:28:16.840 --> 0:28:19.000
<v Speaker 1>the one that made food powder for tech people.

0:28:20.400 --> 0:28:24.800
<v Speaker 10>Why C went from you know, a summer program for

0:28:24.960 --> 0:28:28.800
<v Speaker 10>startups with just a couple companies to doing all sorts

0:28:28.840 --> 0:28:33.000
<v Speaker 10>of things, venture investing with the Continuity Fund, nonprofit work

0:28:33.119 --> 0:28:38.600
<v Speaker 10>with YC research. There's definitely a there's a feeling of like, Okay,

0:28:39.240 --> 0:28:42.960
<v Speaker 10>how much should YC really be doing? Should this organization

0:28:43.360 --> 0:28:44.360
<v Speaker 10>be doing everything.

0:28:44.960 --> 0:28:47.600
<v Speaker 1>During his time at YC, John noticed a couple things

0:28:47.640 --> 0:28:48.240
<v Speaker 1>about Sam.

0:28:48.760 --> 0:28:54.840
<v Speaker 10>He's very good at switching gears and listening very intently.

0:28:55.120 --> 0:28:58.520
<v Speaker 10>I think that's what's really that's really like a superpower.

0:28:58.520 --> 0:29:01.280
<v Speaker 10>I heard someone describe him as the Michael Jordan of listening.

0:29:01.800 --> 0:29:04.320
<v Speaker 1>A lot of people have described this intensity of Sam's

0:29:04.360 --> 0:29:07.960
<v Speaker 1>to me. When he listens, he also stares. It can

0:29:07.960 --> 0:29:11.880
<v Speaker 1>almost feel unsettling. And now that Sam was running YC,

0:29:12.480 --> 0:29:15.480
<v Speaker 1>he could become kind of a covert operator. He had

0:29:15.520 --> 0:29:17.720
<v Speaker 1>all of his deal making skills from his Loop days,

0:29:18.080 --> 0:29:21.120
<v Speaker 1>and now his network and influence was way bigger. If

0:29:21.120 --> 0:29:23.000
<v Speaker 1>there was ever a problem, he could make a call

0:29:23.080 --> 0:29:24.800
<v Speaker 1>and fix it fast.

0:29:25.000 --> 0:29:26.960
<v Speaker 10>Early, very early in my business career, I had a

0:29:27.080 --> 0:29:32.120
<v Speaker 10>very tough negotiation going on, and I wrote Sam this

0:29:32.240 --> 0:29:36.640
<v Speaker 10>email saying like, Hey, I'm in this negotiation. I just

0:29:36.680 --> 0:29:39.600
<v Speaker 10>want to reality check some stuff with you, and he

0:29:40.040 --> 0:29:42.800
<v Speaker 10>called me immediately. We talked for like five minutes and

0:29:42.800 --> 0:29:45.080
<v Speaker 10>he completely solved my problem, and it was like one

0:29:45.080 --> 0:29:47.240
<v Speaker 10>of the best business deals in my life. It was like,

0:29:47.680 --> 0:29:51.120
<v Speaker 10>actually left a really really like big impact on me.

0:29:51.600 --> 0:29:55.360
<v Speaker 10>I've personally seen Sam resolve a one hundred million dollar

0:29:55.440 --> 0:29:59.240
<v Speaker 10>issue in a thirty minute phone call, fifteen minute phone call.

0:30:00.040 --> 0:30:02.360
<v Speaker 10>It's really remarkable, and I think Sam just thinks about

0:30:02.360 --> 0:30:04.920
<v Speaker 10>it in human terms. It's like, this person wants X,

0:30:04.960 --> 0:30:07.160
<v Speaker 10>this person wants why. How can we bring these two

0:30:07.200 --> 0:30:07.760
<v Speaker 10>people together.

0:30:08.640 --> 0:30:13.920
<v Speaker 1>YC was growing, but Sam's ambitions were even larger. When

0:30:14.000 --> 0:30:17.600
<v Speaker 1>he became president, he started taking on pet projects. One

0:30:17.640 --> 0:30:21.240
<v Speaker 1>of his interests was nuclear fusion. To encourage more people

0:30:21.240 --> 0:30:24.960
<v Speaker 1>to build startups in this area, he expanded yc's scope

0:30:24.960 --> 0:30:28.800
<v Speaker 1>to include hard tech startups, that is, companies building tech

0:30:28.840 --> 0:30:32.160
<v Speaker 1>where there's some doubt that it'll work at all. Until

0:30:32.160 --> 0:30:36.080
<v Speaker 1>that point, YC had been almost exclusively about Internet software.

0:30:37.440 --> 0:30:40.120
<v Speaker 1>He also spun up a research arm at YC and

0:30:40.200 --> 0:30:43.960
<v Speaker 1>assigned researchers to projects he wanted to explore. One example

0:30:44.080 --> 0:30:47.760
<v Speaker 1>was universal basic income. The idea is to give every

0:30:47.800 --> 0:30:51.560
<v Speaker 1>person a regular income, regardless of whether they work. So

0:30:51.760 --> 0:30:55.600
<v Speaker 1>under Sam's command, YC created a study to give money

0:30:55.840 --> 0:30:59.640
<v Speaker 1>no strings attached to families in Oakland. This move is

0:30:59.680 --> 0:31:02.680
<v Speaker 1>one that Sam turns too often. He'll have an idea

0:31:02.720 --> 0:31:05.320
<v Speaker 1>for something he thinks should exist, and then he pulls

0:31:05.360 --> 0:31:08.040
<v Speaker 1>in people and money to encourage someone to do it.

0:31:08.640 --> 0:31:10.960
<v Speaker 1>Bending the world to his will. As he wrote in

0:31:11.000 --> 0:31:14.480
<v Speaker 1>his blog, he's kind of like the maestro of an orchestra,

0:31:15.000 --> 0:31:20.400
<v Speaker 1>not playing the instruments himself, but conducting his symphony. And

0:31:20.480 --> 0:31:23.280
<v Speaker 1>in twenty fifteen he'll have the chance to do this again,

0:31:23.760 --> 0:31:27.360
<v Speaker 1>this pattern of finding an important topic, then arranging to

0:31:27.360 --> 0:31:29.600
<v Speaker 1>get the right people and the right money on board

0:31:29.640 --> 0:31:35.000
<v Speaker 1>to build it. This time the focus is AI. It'll

0:31:35.000 --> 0:31:38.480
<v Speaker 1>all take place at a dinner. At dinner, that'll change everything.

0:31:39.360 --> 0:31:40.160
<v Speaker 2>We'll be right back.

0:31:46.240 --> 0:31:49.440
<v Speaker 1>Okay, here's a bit of Silicon Valley lore. It's often

0:31:49.440 --> 0:31:53.120
<v Speaker 1>told that in twenty fifteen Elon Musk was worried, or

0:31:53.120 --> 0:31:55.560
<v Speaker 1>at least Elon was going around telling everyone he was

0:31:55.640 --> 0:32:00.680
<v Speaker 1>deeply distraught over the state of artificial intelligence. Back then,

0:32:00.880 --> 0:32:04.240
<v Speaker 1>the biggest powerhouse in AI was Google. It had so

0:32:04.440 --> 0:32:07.200
<v Speaker 1>much money, and it had hired all the best researchers.

0:32:07.600 --> 0:32:10.600
<v Speaker 1>They had Google Brain, and then they acquired deep Mind.

0:32:11.120 --> 0:32:12.920
<v Speaker 1>It's a British lab that was working on some of

0:32:12.960 --> 0:32:16.000
<v Speaker 1>the most exciting things in AI at the time, AI

0:32:16.000 --> 0:32:19.760
<v Speaker 1>that could be more fluid and self taught. Google's early

0:32:19.840 --> 0:32:23.440
<v Speaker 1>lead particularly disturbed Elon because he was worried about the

0:32:23.480 --> 0:32:27.760
<v Speaker 1>possibility that AI could start to grow too powerful, especially

0:32:27.800 --> 0:32:31.479
<v Speaker 1>if an AI entity began to improve itself, and he

0:32:31.520 --> 0:32:33.720
<v Speaker 1>was really worried about the guy in charge of it all.

0:32:34.120 --> 0:32:36.840
<v Speaker 11>So I used to be close friends with our page,

0:32:36.840 --> 0:32:38.080
<v Speaker 11>and I would say at his house and we'd have

0:32:38.160 --> 0:32:41.440
<v Speaker 11>these conversations long into long into the evening about AI,

0:32:41.600 --> 0:32:44.800
<v Speaker 11>and I would I would be constantly urging him to

0:32:45.000 --> 0:32:50.200
<v Speaker 11>be careful about the danger of AI, and and he

0:32:50.360 --> 0:32:53.200
<v Speaker 11>just he was really not concerned about the danger of

0:32:53.240 --> 0:32:55.600
<v Speaker 11>AI and was quite cavalier about it.

0:32:56.160 --> 0:33:00.240
<v Speaker 1>That's Elon on CNBC. I'm always surprised by how catual

0:33:00.320 --> 0:33:03.040
<v Speaker 1>he sounds when he's telling the story of basically ending

0:33:03.040 --> 0:33:06.200
<v Speaker 1>a friendship. But he loves this anecdote because it makes

0:33:06.280 --> 0:33:08.000
<v Speaker 1>him sound smart and forward thinking.

0:33:08.600 --> 0:33:11.960
<v Speaker 11>And at the time, Google, especially after their acquisition of

0:33:11.960 --> 0:33:14.400
<v Speaker 11>deep Mind, had three quarters of the world's AI talent.

0:33:14.440 --> 0:33:16.760
<v Speaker 11>They had obviously a lot of computers and a lot

0:33:16.760 --> 0:33:20.520
<v Speaker 11>of money. But the person who controls that does not,

0:33:20.920 --> 0:33:23.320
<v Speaker 11>or at least that did not seem to be concerned

0:33:23.360 --> 0:33:26.959
<v Speaker 11>about AI safety. That sounded like a real problem. So

0:33:28.080 --> 0:33:31.480
<v Speaker 11>and then the final straw was Larry calling me a

0:33:31.520 --> 0:33:37.000
<v Speaker 11>specist for being pro human consciousness instead of machine consciousness.

0:33:37.040 --> 0:33:41.360
<v Speaker 11>And I'm like, well, yes, I guess I am.

0:33:41.920 --> 0:33:44.400
<v Speaker 1>At the time, a lot of experts thought the idea

0:33:44.480 --> 0:33:47.440
<v Speaker 1>of developing AI so powerful and that it threatened the

0:33:47.480 --> 0:33:51.880
<v Speaker 1>human species was laughable because AI was still having trouble

0:33:51.960 --> 0:33:55.080
<v Speaker 1>distinguishing between a picture of a chihuahua and a picture

0:33:55.160 --> 0:33:58.200
<v Speaker 1>of a blueberry muffin. But to Elon it was a

0:33:58.240 --> 0:34:01.600
<v Speaker 1>real threat. Here he is speaking with Walter Isaacson at

0:34:01.600 --> 0:34:04.560
<v Speaker 1>a conference. He sounds serious, even alarmed.

0:34:05.200 --> 0:34:08.880
<v Speaker 11>I don't think most people understand just how quickly machine

0:34:08.920 --> 0:34:12.960
<v Speaker 11>intelligence is advancing. It's much faster than almost anyone realizes.

0:34:13.080 --> 0:34:17.680
<v Speaker 11>Even within Silicon Valley, people really have no idea.

0:34:17.600 --> 0:34:18.640
<v Speaker 7>Why is that dangerous?

0:34:19.520 --> 0:34:24.279
<v Speaker 11>If there's some digital superintelligence, particularly if it's engaged in

0:34:24.360 --> 0:34:31.160
<v Speaker 11>recursive self improvement, and its optimization or utility function is

0:34:31.200 --> 0:34:34.840
<v Speaker 11>something that's detrimental to humanity, then it will have a

0:34:34.920 --> 0:34:36.040
<v Speaker 11>very bad effect.

0:34:37.120 --> 0:34:37.239
<v Speaker 6>You know.

0:34:37.280 --> 0:34:40.240
<v Speaker 11>It could be just something like getting rid of spam

0:34:40.400 --> 0:34:42.640
<v Speaker 11>email or something, and it's like concludes, well, the best

0:34:42.640 --> 0:34:44.680
<v Speaker 11>way to get over spam is to get rid of humans,

0:34:45.640 --> 0:34:49.600
<v Speaker 11>you know, But why would we lose source of all spam.

0:34:49.920 --> 0:34:52.759
<v Speaker 6>I know we've all watched Howl in two thousand and one.

0:34:53.400 --> 0:34:56.200
<v Speaker 1>The audience actually laughed at him because what he was

0:34:56.200 --> 0:35:01.319
<v Speaker 1>saying sounded so outlandish. At the time, Elon was increasingly

0:35:01.360 --> 0:35:04.800
<v Speaker 1>feeling like he had to do something to dilute Google's power,

0:35:05.280 --> 0:35:07.879
<v Speaker 1>but he might have also been driven by something else.

0:35:08.520 --> 0:35:11.360
<v Speaker 1>Here's Ashley Vance, my colleague at Bloomberg, who wrote a

0:35:11.360 --> 0:35:13.840
<v Speaker 1>biography of Elon twenty thirteen.

0:35:13.920 --> 0:35:18.480
<v Speaker 12>Elon was He was not the guy we know today.

0:35:18.520 --> 0:35:22.640
<v Speaker 12>I mean, he was doing okay, but Tesla was kind

0:35:22.640 --> 0:35:25.879
<v Speaker 12>of just barely starting a hitted stride. SpaceX was doing

0:35:25.920 --> 0:35:29.759
<v Speaker 12>pretty well. Elon was worth probably a few billion dollars.

0:35:30.280 --> 0:35:33.400
<v Speaker 12>Definitely not the richest person in the world.

0:35:34.000 --> 0:35:36.799
<v Speaker 1>Elon's friends at the time were among the richest people

0:35:36.840 --> 0:35:39.879
<v Speaker 1>in the world, Larry Page and Saragey Brinn, the co

0:35:39.960 --> 0:35:40.920
<v Speaker 1>founders of Google.

0:35:41.600 --> 0:35:43.959
<v Speaker 12>I was interviewing Elon a lot then, I was working

0:35:44.000 --> 0:35:47.279
<v Speaker 12>on a book about him, and it was clear to

0:35:47.320 --> 0:35:49.759
<v Speaker 12>me that he was looking over at his friends who

0:35:49.840 --> 0:35:52.520
<v Speaker 12>were doing really well and everything they did, and they

0:35:52.560 --> 0:35:57.560
<v Speaker 12>had these software empires and this growing AI empire. And

0:35:59.280 --> 0:36:04.440
<v Speaker 12>my sense of where Elon's thoughts about AI started to

0:36:05.000 --> 0:36:08.680
<v Speaker 12>originate was part jealousy. I thought he looked at Google

0:36:08.719 --> 0:36:12.040
<v Speaker 12>and all the success it was having and the success

0:36:12.080 --> 0:36:15.000
<v Speaker 12>as friends were having and he had nothing like that.

0:36:15.080 --> 0:36:17.319
<v Speaker 12>But he would never admit this out loud that he

0:36:17.360 --> 0:36:18.240
<v Speaker 12>was jealous.

0:36:18.440 --> 0:36:21.680
<v Speaker 1>Okay, So AI was top of mind for Elon, and

0:36:21.719 --> 0:36:25.800
<v Speaker 1>in twenty fifteen he attended this dinner at the Rosewood Hotel.

0:36:26.440 --> 0:36:29.479
<v Speaker 1>It's this swanky place on sand Hill Road in Menlo Park,

0:36:29.840 --> 0:36:32.000
<v Speaker 1>right near some of the biggest venture capital firms in

0:36:32.000 --> 0:36:36.480
<v Speaker 1>the valley. There were about ten people at this dinner,

0:36:36.880 --> 0:36:40.160
<v Speaker 1>but for our purposes, here are the four most important ones.

0:36:40.600 --> 0:36:44.280
<v Speaker 1>There's Elon, There's Sam. Then there are two other people,

0:36:44.840 --> 0:36:49.160
<v Speaker 1>Ilia Sitskiver and Greg Brockman. Ilia worked at Google. It

0:36:49.200 --> 0:36:52.720
<v Speaker 1>was a really well respected researcher in AI, and Greg

0:36:52.800 --> 0:36:55.360
<v Speaker 1>had been one of the most important people at Stripe,

0:36:55.719 --> 0:36:57.719
<v Speaker 1>taking it from a team of five people to a

0:36:57.760 --> 0:37:01.520
<v Speaker 1>company worth billions of dollars. At dinner, they talked very

0:37:01.520 --> 0:37:04.360
<v Speaker 1>seriously about AI and the threat that it could be

0:37:04.440 --> 0:37:07.959
<v Speaker 1>misused or that it could lead to catastrophe. They talked

0:37:07.960 --> 0:37:10.200
<v Speaker 1>about what it might take to build something that could

0:37:10.200 --> 0:37:15.040
<v Speaker 1>compete with Google. They had the pieces there. Ilia's AI skills,

0:37:15.239 --> 0:37:20.120
<v Speaker 1>Greg's operational experience, Elon's money, and Sam was there to

0:37:20.280 --> 0:37:24.040
<v Speaker 1>orchestrate it all. At that dinner, Elon pledged to put

0:37:24.080 --> 0:37:27.160
<v Speaker 1>a billion dollars towards this project. He came up with

0:37:27.160 --> 0:37:30.000
<v Speaker 1>the name too. You can hear the pride in his

0:37:30.080 --> 0:37:32.360
<v Speaker 1>voice when he describes the idea on CNBC.

0:37:33.320 --> 0:37:36.440
<v Speaker 11>Open ai refers to open source, So the intent was,

0:37:36.640 --> 0:37:39.080
<v Speaker 11>what's the okay? So what was the opposite? What's the

0:37:39.120 --> 0:37:39.640
<v Speaker 11>opposite of.

0:37:41.320 --> 0:37:41.640
<v Speaker 2>Google?

0:37:41.640 --> 0:37:45.520
<v Speaker 11>Would be an open source nonprofit because Google is closed

0:37:45.520 --> 0:37:50.600
<v Speaker 11>sourced for profit, and that profit motivation can be potentially dangerous.

0:37:51.600 --> 0:37:54.719
<v Speaker 1>So that's the birth of open ai. They had an

0:37:54.800 --> 0:37:58.239
<v Speaker 1>initial vision. Open ai would be a research lab and

0:37:58.320 --> 0:38:01.280
<v Speaker 1>it would share its work openly with the public instead

0:38:01.320 --> 0:38:03.200
<v Speaker 1>of trying to keep it private for its own gain.

0:38:03.920 --> 0:38:07.000
<v Speaker 1>And it would be a nonprofit not focused on enriching

0:38:07.000 --> 0:38:10.000
<v Speaker 1>the company, but instead focused on building a safe AI

0:38:10.120 --> 0:38:14.040
<v Speaker 1>that would benefit humanity. It sounds good in theory, but

0:38:14.120 --> 0:38:18.880
<v Speaker 1>those nonprofit open source ideals were about to get complicated fast.

0:38:19.400 --> 0:38:22.480
<v Speaker 1>Open Eye's co founders soon found themselves locked in a

0:38:22.520 --> 0:38:27.160
<v Speaker 1>power struggle and then urgently racing to raise billions of dollars,

0:38:27.920 --> 0:38:30.840
<v Speaker 1>and Sam would come out on top again in a

0:38:30.880 --> 0:38:35.520
<v Speaker 1>way that cemented his power even further. That's next time

0:38:35.640 --> 0:38:45.160
<v Speaker 1>on foundering. Foundering is hosted by me Ellen Hewitt. Sean

0:38:45.239 --> 0:38:49.600
<v Speaker 1>Wen is our executive producer. Molly Nugent is our associate producer.

0:38:50.080 --> 0:38:54.480
<v Speaker 1>Blake Maples is our audio engineer, Mark Million and vandermain

0:38:54.719 --> 0:38:58.560
<v Speaker 1>seth Fiegerman, Tom Giles and Molly Schutz are our story editors.

0:38:59.200 --> 0:39:02.480
<v Speaker 1>We had production hell up from Jessica Nix and Antonio mufferetch.

0:39:03.360 --> 0:39:06.080
<v Speaker 1>Thanks for listening. If you like our show, leave a

0:39:06.120 --> 0:39:09.080
<v Speaker 1>review and most importantly, tell your friends.

0:39:09.640 --> 0:39:10.359
<v Speaker 2>See you next time.