Mark Zuckerberg
Founder and CEO of Meta
Where to hear and read about Mark Zuckerberg1 episode or test
Mark Zuckerberg founded Facebook at Harvard in 2004 and has run it, now Meta, ever since, growing it into a family of apps used by over 3 billion people a day. The thing most worth learning from him is how he paired permanent founder control with a willingness to bet the company again and again: refusing buyouts, giving up desktop revenue for mobile, and cannibalizing ads to answer TikTok.
what we're really buying is time
Meta
Mark Zuckerberg: the story and the lessons
1How Mark Zuckerberg grew up building software
Strategy games taught Zuckerberg to pursue several paths to victory at once. As a boy he played Civilization, where you can win by conquest, technology, diplomacy or culture.
The hosts liken running a top-10 market-cap company to playing that game, and later describe Meta as keeping many paths open at once.
He learned serious programming as a child. At 10 he begged for a C++ for Dummies book and got a weekly tutor, then built ZuckNet, a chat tool linking his house and his dad's dental office.
C++ requires manual memory management, far harder than the simple languages most teens learn.
AOL Instant Messenger showed him free, viral, hackable software. AIM was free to non-subscribers and had 36 million users by 2001, and teens extended it with bots, graphics and ASCII art profiles.
Early projects with Adam D'Angelo gave him a safety net and a core idea. At Exeter they built Synapse, a Winamp plugin that predicted your next song, and Microsoft and Winamp offered to buy it or hire them, reportedly for around $1 million.
D'Angelo's Buddy Zoo, built from AIM buddy lists, showed that some apps get better as more people join. D'Angelo posted "Mark gets way too lucky" on Friendster, and still thinks about it because the luck was never random.
Harvard side projects taught him what people click on. Course Match let students see which friends were in a class, and Harvard took to it at once. Facemash, a hot-or-not site built in one evening from photos taken from Kirkland House servers, overloaded the servers within hours.
The ad board gave him probation, not expulsion. He met Priscilla Chan at his goodbye party the night before the verdict.
2How Zuckerberg launched thefacebook and spread it school by school
Many people had the idea, so execution decided who won. Stanford, Columbia and Yale had online facebooks, Harvard IT kept promising one, and the Winklevoss twins asked him to code theirs. He built his own instead.
He told the Crimson he makes small things and later puts them together. The February 4, 2004 launch merged Course Match and Buddy Zoo into one product. The Winklevoss suit later settled for $65 million, $45 million of it in stock.
Free open-source tools let him launch with no money or permission. The LAMP stack meant a $10 domain and about $100 a year of hosting, where such a site once needed Oracle or Microsoft enterprise software.
A technical founder no longer needed a business co-founder. Zuckerberg and Eduardo Saverin each put in $1,000 for a two-thirds and one-third split.
Zuckerberg built on real names and user-entered content from day one. Sign-up required a harvard.edu email, and after Facemash he had users enter all their own content, from photo to phone number.
That trained users to contribute and made trust the foundation. Myspace and Friendster were open to anyone and grew faster at first.
He seeded one dense network so every newcomer found it alive. 650 people joined in the first 24 hours and over half of Harvard undergraduates were active within two weeks. 70% of everyone who ever signed up was active on a given day, and that held for years.
Friendster's marketing sent newcomers to an empty site. Zuckerberg's framing: "almost nobody can use the Facebook but for the people who can it's an unbelievably great experience."
He attacked the toughest rival first. Three weeks after Harvard, Facebook opened at Columbia because a competitor was already there, then spread to nearby schools.
Each school ran as a separate network on separate servers, which kept trust high and server load small. Schools opened only when waitlist demand was strong, and Facebook reached about 100 schools by summer.
His 2005 roadmap made every feature feed one engagement loop. The plan listed a redesign, photos, a personalized newspaper, events, local business and an "I'm bored" games feature, each a separate company elsewhere. Photos was written by one engineer in about two weeks.
Events led to tagged party photos, which flowed into the feed and drew people to the next event. By November 2005 Facebook served 230 million page views a day with under 10 million users, passing Google.
Engagement made Facebook profitable within about a year and a half. By October 2005 it made $1 million a month simply by plugging in ad networks.
Of its 5 million users in September 2005, 85% were weekly active.
3How Zuckerberg kept control of Facebook
Sean Parker steered him to investors who accepted founder control. Parker, ousted by Plaxo's board, warned that VCs would take the board and fire him. He helped reset the company so Zuckerberg controlled the board.
Peter Thiel led a $500,000 round at a $5 million pre-money valuation. Sequoia would not have accepted founder control. Reid Hoffman and Mark Pincus put in $37,500 each.
Investors talked him out of a distraction. Wirehog, a file-sharing tool inside Facebook, was dropped after Reid Hoffman told them to focus, since music was "Kryptonite."
Accel's Series A left Zuckerberg in charge. In spring 2005 Accel invested $12.7 million at a $98 million post-money valuation for under 20%, above Don Graham's $60 million offer, with $3 million as a secondary to Zuckerberg, Moskovitz and Parker.
It was the first post-dotcom institutional deal in which the founder kept control. Jim Breyer's personal $1.1 million at 4.5 cents a share became about a 13,000x return.
Zuckerberg shook hands on selling to Yahoo, then walked away. In summer 2006, with growth slowing outside colleges, he accepted Yahoo's $1 billion all-stock offer. When Yahoo's stock fell 20%, the deal dropped to about $800 million and he declined.
His management team mostly wanted to sell. He was optimizing for impact, had no idea as good as Facebook, and had the better product. The hosts call it "history turning on a knife point."
After Yahoo, he chased cash flow so he would never need to sell. Outsourcing ad sales to Microsoft took revenue from $9 million in 2005 to $48 million in 2006 and $153 million in 2007.
In 2007 he turned down a Microsoft buyout worth about $24 billion and took a $240 million investment at a $15 billion valuation instead, closed overnight as a Google meeting loomed.
Control let him make bets a hired CEO could not. Founder control let Facebook ride out the post-IPO crash, the mobile pivot and the TikTok response without being forced to change course.
The hosts credit his speed of learning, low ego about being right and relentlessness, but warn of survivorship bias: control does not by itself produce good decisions.
4News Feed, Beacon and how Zuckerberg found the ads business
Zuckerberg shipped News Feed to everyone and trusted the data over the revolt. It launched September 5, 2006 with no opt-in, drew 30,000 angry emails on day one, and 10% of users joined a group against it.
Engagement soared, so he posted "Calm down. Breathe. We hear you." and added controls. Within two weeks users had adapted. Ruchi Sanghvi, Andrew Bosworth and Chris Cox built it.
News Feed made open registration land quietly. Sign-ups went from 5,000 to 10,000 a day over the summer of 2006 to 70,000 a day by late fall after anyone could join.
Users already knew their posts were broadcast, so few objected to strangers joining.
The developer platform paid off mainly as an ad channel. At F8 in May 2007, Facebook hit its goal of 5,000 developers in two days.
Facebook sat on a browser on an operating system, so it had to give developers data and distribution. When it cut organic distribution, large developers bought ads instead.
Beacon failed and showed users would not always adapt. Beacon published users' purchases to friends, including a diamond ring that spoiled an engagement. It went opt-in, activity fell to roughly zero, and it was killed about two years later.
Zuckerberg had seen ads as gross and wanted them native to the social fabric. The hosts' verdict: "let's get way too clever for our own good first and then work backward to the most basic ad unit."
Zuckerberg hired Sheryl Sandberg once he saw he did not understand ads. Sandberg, who built Google's self-serve ads, took ads and operations while he kept product, engineering and platform, and the company committed to targeted advertising.
Revenue went from $280 million in 2008 to almost $800 million in 2009, $2 billion in 2010 and $3.7 billion in 2011.
He backed a growth team that grew Facebook through the product. When growth stalled in 2008, Chamath Palihapitiya, Alex Schultz, Naomi Gleit and Javier Olivan built People You May Know and crowdsourced translations.
New users who reached about 10 friends in 14 days stayed. Users went from 100 million in August 2008 to 350 million in 2009.
He personally renamed the "awesome button" to "like". Tests showed likes increased comments, and the signal fed ad targeting. Embedding the button across the web let Facebook track people from site to site.
5How Zuckerberg bet Facebook on mobile ads
Mobile broke every pillar of the desktop business. Apps cannot run inside other apps, Facebook Credits could not add a 30% cut on top of Apple's 30%, and all ads sat in a desktop sidebar that phones did not have.
Platform had been seen as 50% or more of Facebook's value. Steve Jobs declined Zuckerberg's 2008 pitch to put the Facebook platform inside iOS.
He bought Instagram for $1 billion during the IPO quiet period. Instagram had 27 million users and no revenue, and the prior record acquisition was $70 million. He decided it over a weekend.
The 2012 IPO went badly and the stock halved. Facebook raised $16 billion at $38, then fell 53.5% to $17.68 by September 4, with mobile revenue admitted as unproven.
Zuckerberg and Sandberg gave up desktop revenue to rebuild around feed ads. They stopped caring about right-side desktop ads, moved engineering to mobile and accepted missed quarters.
Sandberg told him: "Mark, nobody can fire you, and only you can fire me, so if you're in, I'm in." Full-screen feed ads turned out more valuable than any desktop ad.
He let Bosworth argue for far more ads, not fewer. Zuckerberg wanted fewer mobile ads to protect the feed. Bosworth argued relevance was the problem and needed a huge ad marketplace, around 10,000 times more ads.
Zuckerberg pulled Bosworth off profile work to ship feed ads. By Q4 2013 mobile was 53% of revenue and ad revenue grew 76%, while desktop shrank.
Facebook dropped HTML5 and rebuilt native apps. The HTML5 wrapper kept releases fast but engagement low, so Facebook rewrote its clients over about a year.
After Q4 2013 market cap climbed from about $175 billion toward $500 billion almost in a straight line.
6How Zuckerberg buys, copies and outlasts rivals
Zuckerberg buys companies to buy time. In a 2012 email about Instagram he argued there are a finite number of social mechanics, and buying a company let Facebook build its mechanics into the core within 12 to 24 months.
Either the acquired product or Facebook's version would win. He applied the same logic to WhatsApp and to Snapchat's stories when Snap would not sell.
He builds and buys in parallel. He built Messenger while buying WhatsApp, and Photos while buying Instagram.
Facebook bought, beat or copied Myspace, Twitter, Instagram, WhatsApp, Snapchat and TikTok, and few of its hits were invented; News Feed and photo tagging are the exceptions.
Facebook beat Google+ with a companywide lockdown. It shipped video calls against Hangouts and a Circles competitor while Zuckerberg invoked Cato's "Carthage must be destroyed."
One monolithic codebase and a rule that "you get hired into the company not onto a team" let engineers be reassigned fast.
Zuckerberg saw sharing move from the town square to the living room. Snapchat stories and WhatsApp showed people sharing in small private groups, and he made that the core product strategy in 2019.
Instagram's stories improved on Snap's by ranking a row of heads by who you are likely to tap. Facebook also used Onavo, a VPN it bought in 2013, to see rival app traffic.
He answered TikTok immediately, even at the cost of revenue. Apple's tracking change cost about $10 billion in 2022, yet Meta pushed Reels, which cannibalized ads. The stock fell 26% in a day, a $232 billion loss, and 72% by Halloween, then later rose about 5x.
A professional CEO would likely have waited 6 to 12 months. TikTok won attention without a social graph, so Meta copied ByteDance's central structure for AI, infrastructure, growth and integrity.
7Zuckerberg's platform and AI bets
He set up FAIR in 2013 as product-focused AI research. Zuckerberg and Mike Schroepfer recruited Yann LeCun by letting him stay in New York, keep teaching at NYU and publish openly.
The lab targeted feed ranking and ads, "not Microsoft research," and paid off in about two years. Without it, Reels would not have arrived fast enough against TikTok.
Meta gives away Llama to keep suppliers from gaining leverage. Like publishing its data center specs through Open Compute, free models push down the margins of closed AI providers.
Zuckerberg wants to avoid being locked into a competitor's closed ecosystem. Cheaper models are pure margin, since advertisers pay the same either way.
Reality Labs is his bet on owning the next platform. In 2014 he paid $19 billion for WhatsApp and, 34 days later, $2 billion for Oculus. Reality Labs has lost about $60 billion since 2019.
He says the core products could be twice as profitable on Meta's own platform. Even iPhone-like growth would not break even until around 2035.
8Zuckerberg's mistakes and unresolved risks
The lesson that users always adapt did not hold. It worked for News Feed but failed for Beacon and for later privacy changes.
Permissive early privacy rules caught up with Facebook. A 2011 FTC consent decree required 20 years of audits. A quiz taken by 320,000 people exposed data on 87 million users, which became the Cambridge Analytica story.
The hosts judge Cambridge Analytica's real election impact as thin, but the FTC settlement cost $5 billion, and a July 2018 warning erased $119 billion in a day.
Meta tries everything and accepts many failures. Failures include Facebook Live, Watch, Portal, Workplace, Libra, Places and a solar drone.
His 2015 forecast named AR and AI as future focuses, and he was right on both. Teen mental health remains the largest unresolved risk.
Meta
Zuckerberg built Facebook by seeding a dense, trusted network, keeping control of the company, and copying or buying every new social mechanic that threatened it. He kept making large bets, such as turning down Yahoo, rebuilding for mobile, and cannibalizing revenue to answer TikTok, while owning enough of the company that no one could stop him.
Strategy games taught Zuckerberg to pursue several paths to victory at once. As a boy he played Civilization, where you can win by conquest, technology, diplomacy or culture.
The hosts liken running a top-10 market-cap company to playing that game, and later describe Meta as keeping many paths open at once.
He learned serious programming as a child. At 10 he begged for a C++ for Dummies book and got a weekly tutor, then built ZuckNet, a chat tool linking his house and his dad's dental office.
C++ requires manual memory management, far harder than the simple languages most teens learn.
AOL Instant Messenger showed him free, viral, hackable software. AIM was free to non-subscribers and had 36 million users by 2001, and teens extended it with bots, graphics and ASCII art profiles.
Early projects with Adam D'Angelo gave him a safety net and a core idea. At Exeter they built Synapse, a Winamp plugin that predicted your next song, and Microsoft and Winamp offered to buy it or hire them, reportedly for around $1 million.
D'Angelo's Buddy Zoo, built from AIM buddy lists, showed that some apps get better as more people join. D'Angelo posted "Mark gets way too lucky" on Friendster, and still thinks about it because the luck was never random.
Harvard side projects taught him what people click on. Course Match let students see which friends were in a class, and Harvard took to it at once. Facemash, a hot-or-not site built in one evening from photos taken from Kirkland House servers, overloaded the servers within hours.
The ad board gave him probation, not expulsion. He met Priscilla Chan at his goodbye party the night before the verdict.
Many people had the idea, so execution decided who won. Stanford, Columbia and Yale had online facebooks, Harvard IT kept promising one, and the Winklevoss twins asked him to code theirs. He built his own instead.
He told the Crimson he makes small things and later puts them together. The February 4, 2004 launch merged Course Match and Buddy Zoo into one product. The Winklevoss suit later settled for $65 million, $45 million of it in stock.
Free open-source tools let him launch with no money or permission. The LAMP stack meant a $10 domain and about $100 a year of hosting, where such a site once needed Oracle or Microsoft enterprise software.
A technical founder no longer needed a business co-founder. Zuckerberg and Eduardo Saverin each put in $1,000 for a two-thirds and one-third split.
Zuckerberg built on real names and user-entered content from day one. Sign-up required a harvard.edu email, and after Facemash he had users enter all their own content, from photo to phone number.
That trained users to contribute and made trust the foundation. Myspace and Friendster were open to anyone and grew faster at first.
He seeded one dense network so every newcomer found it alive. 650 people joined in the first 24 hours and over half of Harvard undergraduates were active within two weeks. 70% of everyone who ever signed up was active on a given day, and that held for years.
Friendster's marketing sent newcomers to an empty site. Zuckerberg's framing: "almost nobody can use the Facebook but for the people who can it's an unbelievably great experience."
He attacked the toughest rival first. Three weeks after Harvard, Facebook opened at Columbia because a competitor was already there, then spread to nearby schools.
Each school ran as a separate network on separate servers, which kept trust high and server load small. Schools opened only when waitlist demand was strong, and Facebook reached about 100 schools by summer.
His 2005 roadmap made every feature feed one engagement loop. The plan listed a redesign, photos, a personalized newspaper, events, local business and an "I'm bored" games feature, each a separate company elsewhere. Photos was written by one engineer in about two weeks.
Events led to tagged party photos, which flowed into the feed and drew people to the next event. By November 2005 Facebook served 230 million page views a day with under 10 million users, passing Google.
Engagement made Facebook profitable within about a year and a half. By October 2005 it made $1 million a month simply by plugging in ad networks.
Of its 5 million users in September 2005, 85% were weekly active.
Sean Parker steered him to investors who accepted founder control. Parker, ousted by Plaxo's board, warned that VCs would take the board and fire him. He helped reset the company so Zuckerberg controlled the board.
Peter Thiel led a $500,000 round at a $5 million pre-money valuation. Sequoia would not have accepted founder control. Reid Hoffman and Mark Pincus put in $37,500 each.
Investors talked him out of a distraction. Wirehog, a file-sharing tool inside Facebook, was dropped after Reid Hoffman told them to focus, since music was "Kryptonite."
Accel's Series A left Zuckerberg in charge. In spring 2005 Accel invested $12.7 million at a $98 million post-money valuation for under 20%, above Don Graham's $60 million offer, with $3 million as a secondary to Zuckerberg, Moskovitz and Parker.
It was the first post-dotcom institutional deal in which the founder kept control. Jim Breyer's personal $1.1 million at 4.5 cents a share became about a 13,000x return.
Zuckerberg shook hands on selling to Yahoo, then walked away. In summer 2006, with growth slowing outside colleges, he accepted Yahoo's $1 billion all-stock offer. When Yahoo's stock fell 20%, the deal dropped to about $800 million and he declined.
His management team mostly wanted to sell. He was optimizing for impact, had no idea as good as Facebook, and had the better product. The hosts call it "history turning on a knife point."
After Yahoo, he chased cash flow so he would never need to sell. Outsourcing ad sales to Microsoft took revenue from $9 million in 2005 to $48 million in 2006 and $153 million in 2007.
In 2007 he turned down a Microsoft buyout worth about $24 billion and took a $240 million investment at a $15 billion valuation instead, closed overnight as a Google meeting loomed.
Control let him make bets a hired CEO could not. Founder control let Facebook ride out the post-IPO crash, the mobile pivot and the TikTok response without being forced to change course.
The hosts credit his speed of learning, low ego about being right and relentlessness, but warn of survivorship bias: control does not by itself produce good decisions.
Zuckerberg shipped News Feed to everyone and trusted the data over the revolt. It launched September 5, 2006 with no opt-in, drew 30,000 angry emails on day one, and 10% of users joined a group against it.
Engagement soared, so he posted "Calm down. Breathe. We hear you." and added controls. Within two weeks users had adapted. Ruchi Sanghvi, Andrew Bosworth and Chris Cox built it.
News Feed made open registration land quietly. Sign-ups went from 5,000 to 10,000 a day over the summer of 2006 to 70,000 a day by late fall after anyone could join.
Users already knew their posts were broadcast, so few objected to strangers joining.
The developer platform paid off mainly as an ad channel. At F8 in May 2007, Facebook hit its goal of 5,000 developers in two days.
Facebook sat on a browser on an operating system, so it had to give developers data and distribution. When it cut organic distribution, large developers bought ads instead.
Beacon failed and showed users would not always adapt. Beacon published users' purchases to friends, including a diamond ring that spoiled an engagement. It went opt-in, activity fell to roughly zero, and it was killed about two years later.
Zuckerberg had seen ads as gross and wanted them native to the social fabric. The hosts' verdict: "let's get way too clever for our own good first and then work backward to the most basic ad unit."
Zuckerberg hired Sheryl Sandberg once he saw he did not understand ads. Sandberg, who built Google's self-serve ads, took ads and operations while he kept product, engineering and platform, and the company committed to targeted advertising.
Revenue went from $280 million in 2008 to almost $800 million in 2009, $2 billion in 2010 and $3.7 billion in 2011.
He backed a growth team that grew Facebook through the product. When growth stalled in 2008, Chamath Palihapitiya, Alex Schultz, Naomi Gleit and Javier Olivan built People You May Know and crowdsourced translations.
New users who reached about 10 friends in 14 days stayed. Users went from 100 million in August 2008 to 350 million in 2009.
He personally renamed the "awesome button" to "like". Tests showed likes increased comments, and the signal fed ad targeting. Embedding the button across the web let Facebook track people from site to site.
Mobile broke every pillar of the desktop business. Apps cannot run inside other apps, Facebook Credits could not add a 30% cut on top of Apple's 30%, and all ads sat in a desktop sidebar that phones did not have.
Platform had been seen as 50% or more of Facebook's value. Steve Jobs declined Zuckerberg's 2008 pitch to put the Facebook platform inside iOS.
He bought Instagram for $1 billion during the IPO quiet period. Instagram had 27 million users and no revenue, and the prior record acquisition was $70 million. He decided it over a weekend.
The 2012 IPO went badly and the stock halved. Facebook raised $16 billion at $38, then fell 53.5% to $17.68 by September 4, with mobile revenue admitted as unproven.
Zuckerberg and Sandberg gave up desktop revenue to rebuild around feed ads. They stopped caring about right-side desktop ads, moved engineering to mobile and accepted missed quarters.
Sandberg told him: "Mark, nobody can fire you, and only you can fire me, so if you're in, I'm in." Full-screen feed ads turned out more valuable than any desktop ad.
He let Bosworth argue for far more ads, not fewer. Zuckerberg wanted fewer mobile ads to protect the feed. Bosworth argued relevance was the problem and needed a huge ad marketplace, around 10,000 times more ads.
Zuckerberg pulled Bosworth off profile work to ship feed ads. By Q4 2013 mobile was 53% of revenue and ad revenue grew 76%, while desktop shrank.
Facebook dropped HTML5 and rebuilt native apps. The HTML5 wrapper kept releases fast but engagement low, so Facebook rewrote its clients over about a year.
After Q4 2013 market cap climbed from about $175 billion toward $500 billion almost in a straight line.
Zuckerberg buys companies to buy time. In a 2012 email about Instagram he argued there are a finite number of social mechanics, and buying a company let Facebook build its mechanics into the core within 12 to 24 months.
Either the acquired product or Facebook's version would win. He applied the same logic to WhatsApp and to Snapchat's stories when Snap would not sell.
He builds and buys in parallel. He built Messenger while buying WhatsApp, and Photos while buying Instagram.
Facebook bought, beat or copied Myspace, Twitter, Instagram, WhatsApp, Snapchat and TikTok, and few of its hits were invented; News Feed and photo tagging are the exceptions.
Facebook beat Google+ with a companywide lockdown. It shipped video calls against Hangouts and a Circles competitor while Zuckerberg invoked Cato's "Carthage must be destroyed."
One monolithic codebase and a rule that "you get hired into the company not onto a team" let engineers be reassigned fast.
Zuckerberg saw sharing move from the town square to the living room. Snapchat stories and WhatsApp showed people sharing in small private groups, and he made that the core product strategy in 2019.
Instagram's stories improved on Snap's by ranking a row of heads by who you are likely to tap. Facebook also used Onavo, a VPN it bought in 2013, to see rival app traffic.
He answered TikTok immediately, even at the cost of revenue. Apple's tracking change cost about $10 billion in 2022, yet Meta pushed Reels, which cannibalized ads. The stock fell 26% in a day, a $232 billion loss, and 72% by Halloween, then later rose about 5x.
A professional CEO would likely have waited 6 to 12 months. TikTok won attention without a social graph, so Meta copied ByteDance's central structure for AI, infrastructure, growth and integrity.
He set up FAIR in 2013 as product-focused AI research. Zuckerberg and Mike Schroepfer recruited Yann LeCun by letting him stay in New York, keep teaching at NYU and publish openly.
The lab targeted feed ranking and ads, "not Microsoft research," and paid off in about two years. Without it, Reels would not have arrived fast enough against TikTok.
Meta gives away Llama to keep suppliers from gaining leverage. Like publishing its data center specs through Open Compute, free models push down the margins of closed AI providers.
Zuckerberg wants to avoid being locked into a competitor's closed ecosystem. Cheaper models are pure margin, since advertisers pay the same either way.
Reality Labs is his bet on owning the next platform. In 2014 he paid $19 billion for WhatsApp and, 34 days later, $2 billion for Oculus. Reality Labs has lost about $60 billion since 2019.
He says the core products could be twice as profitable on Meta's own platform. Even iPhone-like growth would not break even until around 2035.
The lesson that users always adapt did not hold. It worked for News Feed but failed for Beacon and for later privacy changes.
Permissive early privacy rules caught up with Facebook. A 2011 FTC consent decree required 20 years of audits. A quiz taken by 320,000 people exposed data on 87 million users, which became the Cambridge Analytica story.
The hosts judge Cambridge Analytica's real election impact as thin, but the FTC settlement cost $5 billion, and a July 2018 warning erased $119 billion in a day.
Meta tries everything and accepts many failures. Failures include Facebook Live, Watch, Portal, Workplace, Libra, Places and a solar drone.
His 2015 forecast named AR and AI as future focuses, and he was right on both. Teen mental health remains the largest unresolved risk.
Meta
