By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.
Bloomberg OriginalsMeet The New Mark Zuckerberg | The Circuit
Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.
CHAPTERS
Chapter summaries
The “Zuckissance” and leaning into memes
Emily Chang opens by asking Mark Zuckerberg about the idea of a “Zuckissance,” pointing to his evolving public image and meme culture around him. Zuckerberg responds that as he’s gotten older, he cares less about the noise and tries to be himself, while joking that the memes should at least be good.
- •Emily Chang introduces the term “Zuckissance” and explains it as a “Zuck renaissance.”
- •Emily cites his new look and the way his actions become memes.
- •Zuckerberg says he’s more comfortable leaning into it and being himself.
- •Zuckerberg jokes that people should “make them good memes.”
Why Meta is betting on open-source AI
Chang frames Zuckerberg’s long tenure in the spotlight and Meta’s new push to open-source AI versus closed competitors. At Meta HQ, she connects this to Zuckerberg’s early interest in open source, and he explains how open code and a “hacker mentality” made building early Facebook possible.
- •Emily Chang sets context: years of scrutiny and a new push to open-source AI.
- •Chang notes a 2003 article showing Zuckerberg discussed open source early on.
- •Zuckerberg says open source enabled Facebook’s early development with limited capital.
- •He describes a “hacker mentality” of adapting existing code to build quickly.
Open source as Meta’s strategy: from Open Compute to AI standards
Chang calls Zuckerberg an “unlikely champion” of open source, and he argues Meta has practiced it for years. He points to Meta’s work on infrastructure standards and says a similar dynamic will happen with AI.
- •Chang challenges the idea that Zuckerberg is a natural open-source advocate.
- •Zuckerberg says Meta has been a major open-source proponent for a long time.
- •He references Meta’s work standardizing data-center and server infrastructure via the Open Compute Project.
- •He predicts open, shared approaches will also shape AI industry standards.
Regaining platform control: AI and AR/VR as “next platforms”
Chang asks how much open-sourcing AI is about controlling the next technology wave. Zuckerberg says building on phone platforms controlled by competitors has been “soul-crushing,” and he wants Meta to have more control over the next generation of platforms—especially AI and augmented/virtual reality—while arguing that sharing can strengthen an ecosystem.
- •Chang asks whether open-source AI is a strategic play to own the next wave.
- •Zuckerberg describes frustration with dependency on competitor-controlled platforms.
- •He names AI and augmented/virtual reality as critical future platforms for Meta.
- •He argues open sharing can make the system stronger by growing the ecosystem.
Meta AI product pitch: real-time image generation and usage goals
Chang presses on why someone should use Meta AI instead of ChatGPT. Zuckerberg highlights features he claims are better, including generating images as you type, and says his goal is for Meta AI to become the world’s most used AI assistant, citing “hundreds of millions” of users and saying they’ll reach the goal before year-end.
- •Chang asks why users should choose Meta AI over ChatGPT.
- •Zuckerberg describes a feature that generates images in real time as you type.
- •He says he uses it with his daughters and finds it “really cool.”
- •He sets a goal: Meta AI becoming the most used AI assistant in the world by year-end.
- •He claims they’re on track, with hundreds of millions already using it.
Llama 3.1 and the “many models” worldview
Chang brings up Llama 3.1 and its range of models, including a very large model. Zuckerberg says the key impact of a large, open model is that others can use it as a “teacher” to train smaller specialized models, and he rejects the idea of one universal AI, arguing instead for millions or billions of models tailored to different needs.
- •Chang introduces Llama 3.1 and its big-and-small model family.
- •Zuckerberg says open access lets people use the large model to train smaller ones.
- •He says closed systems make building custom models difficult for organizations.
- •He argues there won’t be “one AI to rule them all,” but many models for many uses.
AI personas, creators, and role-playing tough conversations
Chang explores broader implications: AI-generated influencers, avatars, and even an AI-generated social network. Zuckerberg says “we’ll have all of it,” emphasizes creators on Meta’s platforms, and highlights role-playing difficult social interactions as a top Meta AI use case, describing it as a judgment-free way to practice conversations.
- •Chang asks about a future of AI influencers, captions, and avatars talking to avatars.
- •Zuckerberg says these outcomes will all exist in the ecosystem.
- •He discusses enabling creators to train AI versions of themselves for community interaction.
- •Chang and Zuckerberg discuss AI as a tool for role-playing hard social interactions.
- •Zuckerberg says a key benefit is the lack of judgment or social repercussions.
Metaverse progress check: why Meta renamed and what’s faster/slower
Chang challenges Meta’s continued spending on the metaverse and asks whether progress matches expectations. Zuckerberg says the rebrand helped avoid being pigeonholed as only a social media company, calls the metaverse long-term, and notes some efforts are slower while glasses are going better than expected—especially as a form factor for AI.
- •Chang asks about the Meta rebrand and ongoing metaverse investment.
- •Zuckerberg says the metaverse was always intended as a long-term project.
- •He says some progress is slower than expected, while glasses are going better.
- •He calls glasses stylish and a strong form factor for AI.
- •He says he once expected AR before AI, but AI arrived sooner than he thought.
Is AI investment a bubble? The timing vs. the stakes
Chang asks when AI spending starts paying off and whether it’s a bubble. Zuckerberg compares it to past tech bubbles where value emerged later than expected, says companies might be overbuilding, but argues the investments are rational because falling behind could mean missing the most important technology of the next decade-plus.
- •Chang questions ROI timing for chips, data centers, and infrastructure spend.
- •Zuckerberg says bubbles can still produce long-term value; timing is the hard part.
- •He says there’s a meaningful chance companies are overbuilding now.
- •He argues the downside of being behind is too large given AI’s importance.
- •He frames AI as central for the next 10 to 15 years.
AGI and what “intelligence” means: no single definition
Chang asks how Zuckerberg defines AGI and whether Meta gets there first. Zuckerberg says Meta is already working on Llama 4 and wants to close the gap with competitors, but he doesn’t commit on being first and argues intelligence has multiple dimensions rather than one definition, mentioning interest in capabilities like reading faces and emotions.
- •Chang asks for Zuckerberg’s definition of AGI and whether Meta wins the race.
- •Zuckerberg says they’re already starting work on Llama 4 to close gaps with others.
- •He says predicting breakthroughs is hard and not linear.
- •He argues there’s no single definition because intelligence is multi-dimensional.
- •He mentions interest in modalities like reading faces and emotions.
Nature, parenting, and learning: values, creativity, and going deep
The setting shifts to a quieter, outdoors moment, and Zuckerberg describes spending summer time with family and getting kids into sports. Asked what kids should study, he emphasizes critical thinking and values, shares a story about his daughter writing a novel and using Meta AI to generate images, and connects this to a hiring philosophy of valuing people who can go deep and learn.
- •Chang remarks on the peaceful setting; Zuckerberg says they love being in nature.
- •Zuckerberg describes family time, summer routines, and teaching sports.
- •He says kids should learn critical thinking and values when young.
- •He shares his daughter’s novel-writing and using Meta AI to generate images.
- •He describes a hiring philosophy: depth in one area demonstrates the ability to learn and excel.
US–China AI competition and open innovation vs. lockdown
Chang asks about China’s position in AI/AGI and how the US should approach competition. Zuckerberg says he doesn’t know a lot personally, argues against “locking it all down,” and says the US thrives on open and decentralized innovation; he also says leading companies should work with the US government to maintain an advantage and describes aiming for a continual, months-long lead rather than years.
- •Chang raises Zuckerberg’s long-standing interest in China and asks about China’s AI position.
- •Zuckerberg says he doesn’t personally know a ton about it.
- •He argues the US shouldn’t respond by locking everything down.
- •He says US strength comes from open, decentralized innovation.
- •He suggests companies working with the US government to keep an advantage in defense and leading tech.
- •He frames a realistic target as a continual months-long lead, not a multi-year gap.
Trust, scrutiny, elections, and mental health: Meta’s role and responsibility
Chang asks about Sam Altman’s leadership, trust in Meta given past controversies, the 2024 US election, and teen mental health concerns. Zuckerberg credits Altman while noting the irony of “OpenAI” leading closed models, says open source enables scrutiny and faster quality improvement, says Meta will recommend less political content and he won’t endorse candidates, and argues mental health is a serious issue while disputing that science has proven social media is the cause—emphasizing parent-aligned controls.
- •Chang asks about Sam Altman; Zuckerberg gives him credit and says he’s handling scrutiny gracefully.
- •Zuckerberg calls it ironic that “OpenAI” leads in closed models, though not necessarily bad.
- •Asked why people should trust Meta with AI, Zuckerberg says scrutiny is healthy and open source enables public inspection.
- •On the 2024 election, Zuckerberg says people want less political content; Meta will recommend less of it and he won’t endorse candidates.
- •He comments on a widely seen moment involving Donald Trump and says Meta’s role is enabling expression of views.
- •On teen mental health, he agrees it’s a serious issue and stresses building tools aligned with parents’ needs.
- •He says he doesn’t think science supports a proven connection as commonly portrayed, while acknowledging broader societal issues.
Wakesurfing with Mark: family banter, coaching, and the long arc of tech
The episode closes with a lighter sequence around wakesurfing, featuring instruction, jokes, and family commentary. Zuckerberg explains technique with a simple coaching motto, then reflects on how long he wants to keep building—describing work in 10–15 year chapters—and shares optimism that AI will expand creativity by lowering the technical barriers to building, ending with a playful exchange about sunscreen.
- •Chang and others banter about surfing, family dynamics, and dad jokes (including “not my best work”).
- •Zuckerberg gives wakesurfing instructions and repeats a family sports motto: “your body knows what to do.”
- •Chang attempts to get up on the board; the group jokes about whether the sport is fun in the hard moments.
- •Asked about tenure, Zuckerberg says he thinks in 10–15 year building chapters, with AI and new computing platforms taking time to materialize.
- •He says he doesn’t see an “end point” to tech’s arc and is optimistic AI will increase creativity by helping people build without needing deep coding skills.
- •Chang closes with a humorous sunscreen gift; Zuckerberg quips it matches how much he needs.
