18 Apr 2025 13 min read CEO

What Broke Social Media? Ev Williams on Lessons for the AI Era

What happens when the founder of Blogger, Twitter, and Medium takes a hard look back at the internet he helped create? In this rare, reflective episode, Ev Williams joins Eric Ries for a no-holds-barred conversation about building products that endure, the seductive trap of short-term metrics, and why real innovation starts with intuition—not A/B tests.

Ev shares behind-the-scenes stories of Twitter’s growth, Google’s acquisition of Blogger, and the early dreams that fueled the internet’s rise—and the moments that broke them. He explores how culture, governance, and trust shape a company’s soul, and what today’s AI founders must learn from the “social media experiment.”

Whether you’re building a startup, designing a product, or just trying to understand the digital world around you, this is a must-listen episode in long-term thinking from one of tech’s most quietly influential voices.

PS: Also check out the Hatching Twitter book for its full founding and growth story.

🧱 Founding Stories & Product Creation

1. Blogger (1999–2003)

Ev’s first major hit was Blogger, one of the earliest hosted blogging platforms. The product started scrappily, in a pre-VC era of internet startups:

“Google bought Blogger in 2003. I'd been working on it for four years. The relationships actually weren't great with my co-founder and employees. I had to lay everybody off. Then I was by myself for quite a while starting to rebuild the team.”

Despite its shaky beginnings, Blogger succeeded in democratizing online publishing for millions:

“Millions of people have put their thoughts out there thanks to that.”

Acquisition by Google

The deal was all stock—and while that turned out to be lucrative, Ev’s decision was grounded more in his desire to protect the product:

“It was what’s going to be the best thing for this product, this company that I still believe in a lot.”

Blogger was the first external company ever acquired by Google, a fact Ev underlines with both pride and surprise:

“We were actually the first company to sell to Google.”

He credits Tim O’Reilly’s relationship with Larry and Sergey for catalyzing the acquisition—a lesson that recurs across all his ventures: relationships shape outcomes.

2. Twitter (2006–2011, Board until ~2021)

Ev co-founded Twitter inside a company called Odeo, originally a podcasting platform. As Twitter gained traction, it spun out and eventually became a cultural phenomenon.

“Twitter is… maybe no longer Twitter, but there's been a few things that just feel nice to build something and have it last.”

Leadership & Transition

Ev was CEO during the critical growth phase, but despite leading a 10x increase in valuation, users, and employees in 2 years, he was pushed out:

“In the two years I was CEO, our valuation, user base, and employees increased 10x each… and I was still fired.”

He later stayed on the board for a decade. Dick Costolo, who became CEO, was someone Ev had known for 10 years and had recruited personally—another example of the importance of long-term trust-based networks.

Ev reflects on Twitter’s trajectory and how pressure for growth often overrides long-term health:

“The pressure to just make the numbers go up had very negative impact.”

He also doubts whether social media’s harms were even preventable in a capitalist context:

“Social media was inevitable… We were exploiting human nature.”

3. Medium (2012–present)

Founded in 2012, Medium was Ev’s return to publishing—this time with more polish and vision. It became known for its beautiful reading and writing UX:

“The main thing Medium got right in the very early days is this beautiful writing experience… craftsmanship at the highest level.”

But as the company grew, they made decisions that undercut trust with users—like the much-criticized login modal blocking access to articles:

“No one wanted it… but of course we tested it and the signups went through the roof… Then we hired this VP of marketing… who said the sentiment on Twitter was terrible.”

Despite user outcry, the team kept reverting back to this dark pattern—a perfect example of metrics over experience.

Founder’s Reflection on Medium:

“We basically very rarely invested in [the writing experience] over time… because we had raised a lot of money and at the end of the day the system is internal, it’s psychological, it’s unconscious… it is destroying an incredible amount of value because it’s incompatible with craftsmanship.”

4. Obvious Ventures (2014–present)

Ev also co-founded Obvious Ventures, a purpose-driven VC firm that backs “world positive” companies.

“Even my venture firm just turned 10 last year. James who I formed Obvious with, I knew for 10–15 years.”

Again, Ev points to decade-long relationships with co-founders like James Joaquin and partners like Vishal Vasishth as the foundation of sustainability and impact.

He positions Obvious Ventures as a counterforce to extractive capitalism—still idealistic, but with eyes wide open:

“I choose optimism… not because it always wins intellectually, but because I choose to be.”

🧠 Key Principles and Lessons

This section of the podcast is rich with introspection. Ev Williams lays out a career-spanning set of beliefs about what truly drives lasting value, how creativity is often smothered by metrics, and why relationships and integrity matter more than we admit.

1. Relationships Are the True Moat

Ev repeatedly returns to the idea that long-term personal relationships are the most underrated asset in building companies:

“So many of the successes I’ve had I attribute to these very long-term relationships.”

Examples:

Blogger was acquired by Google largely due to a connection between investor Tim O’Reilly and the Google founders.
Twitter's leadership continuity came from knowing Dick Costolo 10 years before he became CEO.
Medium's next CEO, Tony Stubblebine, had worked with Ev at Odeo back in 2005.
Obvious Ventures was co-founded with James Joaquin, a longtime friend, and team members were chosen from people Ev had known for over a decade.

🧩 Lesson: Your network isn’t just career insurance—it’s the foundation of resilient, long-term products.

2. Longevity > Hype

Ev is proud that his projects—Blogger, Twitter, Medium—have endured:

“It feels really lucky to create something that has that longevity… in the digital world, everything kind of disintegrates.”

This endurance, he argues, is less about brilliant strategy and more about resisting short-termism and cultivating continuity.

He criticizes today’s “epidemic of short-term thinking” and views resisting it as a competitive advantage.

3. Creativity Beats Optimization

A core theme: creative insight—not optimization—creates the most value.

“All the value I and my teams have ever created is from someone saying, ‘Hey, wouldn’t it be cool if…’”

Ev references the book Why Greatness Cannot Be Planned, calling it one of his all-time favorites. It argues that rigid goals kill serendipity and originality:

“In consumer products, especially, the value creation is a creative process. Then it gets treated like an engineering or scientific process.”

This mismatch leads to mediocrity disguised as efficiency.

4. Craftsmanship Is Incompatible With ‘Make Number Go Up’

Ev critiques the tyranny of metrics, particularly A/B testing culture and OKRs:

“It’s destroying an incredible amount of value because it’s incompatible with craftsmanship.”

He illustrates this with the story of Medium’s intrusive login modal, which boosted sign-ups but violated user trust and brand identity:

“No one wanted it… but the signups went through the roof… Then we saw on Twitter that everyone hated it.”

Also mentioned: the story of Groupon’s “two emails a day” test, which boosted revenue short term but harmed long-term user sentiment.

🧩 Lesson: Metrics alone can’t tell you if you're building something worth loving. Sometimes the spreadsheet lies.

5. Subjective Vision Is Not a Flaw—It’s a Feature

Ev likens creative product leadership to film direction:

“If you're in a movie set, there could be hundreds of people working on it… but the director has say over very minute aspects.”

In startups, this subjective judgment often gets replaced by consensus, metrics, or politics. That, he says, erodes magic.

Ev admits founders often cloak ego in mission language:

“There’s an ego selfishness that can get bound up in entrepreneurship… The rhetoric of changing the world becomes cover for ‘I want to do the thing I want to do.’”

The desire to “change the world” may be genuine—but without reflection, it can drift into self-justification.

🔁 Summary Takeaways

Principle	Core Idea	Ev’s Quote or Example
Relationships > Strategy	Long-term partnerships sustain products	“So many of the successes… are because of these long-term relationships.”
Intuition > Metrics	Vision-led products outperform test-driven ones	“Do you want that? What do you feel?”
Trust > Growth Hacks	Login popups and email spam erode brand equity	“No one wanted it… but the signups went through the roof.”
Creativity is Core	Originality requires exploration, not goal-chasing	Why Greatness Cannot Be Planned
Culture Empowers	Torchbearers need protection, not just slogans	“They don’t have the tools to defend their arguments.”

🔧 Product Building Philosophy

Ev Williams doesn’t just talk about building products—he deconstructs the psychological, cultural, and philosophical contradictions baked into modern startup life. He emphasizes that while Silicon Valley celebrates innovation, it often kills originality with its own tools: metrics, OKRs, growth-at-all-costs thinking.

1. The Billion-Dollar Formula: Desire + Ease

Ev shares a dead-simple (but powerful) formula for product thinking:

“Identify a human desire, preferably one that’s been around for a really long time… then make it easier to achieve that desire.”

This philosophy underpins the success of Blogger, Twitter, and even Mozi (his new startup). He notes that most enduring tech products aren’t new inventions, but new efficiencies:

“Most of the really big internet companies are just things that made it faster or easier to get something you already wanted.”

🧩 Implication: Don’t try to invent a desire. Find one that already exists—and reduce the friction.

2. Creative Instinct > Data Paralysis

Ev rails against the idea that you must wait for A/B test results to act:

“You know the product is bad and you know those things are going to make it better. The reason you're paralyzed is you don't have enough data to give you permission.”

He tells founders: If you know something will make the product better—just do it. Especially in early-stage products, data often lags behind intuition.

Eric adds:

“Even I can tell you that’s stupid… the whole idea behind the pivot was: data is for optimization—not permission.”

3. The Myth of the Objective

Ev references Why Greatness Cannot Be Planned again here, saying:

“The subtitle is ‘The Myth of the Objective.’ It argues that for anything truly original, setting goals is counterproductive.”

Instead, he frames product creation as a creative process, not a linear or rational one:

“In consumer products especially, the value creation is a creative process… then it gets treated like an engineering process.”

He likens it to filmmaking: the director doesn’t need a spreadsheet to know a scene isn’t working—they feel it. Likewise, product leaders must trust their aesthetic, emotional, and user-centric judgment.

4. “Make the Numbers Go Up” Is a Trap

Ev describes a recurring arc:

“All the value we created came from ‘Hey wouldn’t it be cool if…’ Then, once you get employees and funding, the job becomes ‘make the numbers go up.’”

This mindset shift—from creative play to performance optimization—leads to decisions that erode product quality and user trust. He revisits Medium’s pop-up login wall as a textbook example:

“No one wanted it. But we tested it, and the sign-ups went through the roof. Then we saw sentiment was terrible.”

What’s worse, the feature kept creeping back even after they knew it was bad—because the spreadsheet said it worked.

5. Craftsmanship Requires Protection

Ev believes great product quality can’t be maintained unless leadership actively protects it:

“You’re chopping down trees to make the numbers go up in your forest.”

He argues that if you don’t explicitly prioritize craftsmanship, internal pressures (like growth OKRs, investor updates, team morale) will drive decision-making instead.

This echoes Eric’s idea of a “culture bank”: you make deposits through hard decisions that prioritize trust and product quality, and withdrawals (like deceptive pop-ups or spammy emails) can bankrupt you.

6. Empower the Creators (Not the MBAs)

Ev describes a common startup dynamic where MBAs with spreadsheets override torchbearers—the ones who truly understand what’s right for the user:

“Within any organization, there are torchbearers—the true believers. But they don’t have the tools to defend their arguments.”

He believes founders and leaders must protect these people—often customer support reps, designers, or ICs—who carry the product’s soul and fight against commoditization.

7. You Can Only Go So Far in a Straight Line

Ev shares a metaphor from AI research: optimizing a robot to exit a maze will make it run into dead ends. Why? Because the goal is too rigid.

“If you always try to get closer to the exit, you run into dead ends. Sometimes you need to go sideways to go forward.”

Applied to product strategy, it means: chasing growth blindly leads to stagnation. Sometimes the best move is non-obvious, inefficient, or counter-intuitive—but essential to reach higher ground.

📌 Summary: Ev’s Product Philosophy in 7 Rules

Principle	Summary
Desire + Ease	Focus on old, deep human wants. Make them easier to fulfill.
Creative First	Build the product you believe in—not what the metrics tell you to.
No Mythical Roadmap	Greatness is discovered, not goal-tracked.
Fight Metric Madness	Optimization kills originality if left unchecked.
Protect the Craft	Good product experiences aren’t accidental—they’re defended.
Empower Torchbearers	The best product ideas come from the edges, not the execs.
Wander with Purpose	Sometimes you need to take a step back—or sideways—to leap forward.

Ev and Eric dive into a core question that’s top-of-mind for today’s founders and investors:

“Can we build AI differently from how we built social media?”

The entire conversation acts as both a postmortem of social media and a warning shot for AI builders.

1. The Shared Feeling: “Something Went Wrong”

Ev acknowledges a widespread belief among new AI entrepreneurs: they don’t want to repeat the mistakes of social media.

“I’ve heard it too… It was just a given [at the AI conference] that ‘something went wrong’ with social media, and the question is how do we avoid that?”

But he immediately poses a deeper question:

“Was that preventable?”

This leads to a subtle but powerful thread: maybe the negative outcomes of social media weren’t just mistakes—they were inevitable outcomes of system design, business models, and human nature.

Ev takes a strong stance:

“I think social media was inevitable… It’s an extension of media… What we were exploiting was human nature.”

He compares the rise of Facebook and Twitter to processed food:

“If 7-Elevens had said, ‘Maybe selling all this sugar to the masses isn’t good,’ would that have stopped the obesity epidemic? No. Because the masses wanted sugar. Someone else would’ve done it.”

The analogy is clear: even if Twitter or Facebook hadn’t pursued certain addictive patterns, competitors would have. The incentives made it nearly impossible to act otherwise.

3. Free Speech & “Healthy Dialogue” Myths

Ev critiques Elon Musk’s rhetoric when taking over Twitter:

“He said he wanted to create a place with healthy dialogue and knowledge sharing. And I thought, ‘What do you think we were trying to do?’”

He pushes back on the naive belief that product design alone can force public discourse into harmony.

“You take that to the extreme—Twitter wouldn’t be Twitter. Instagram wouldn’t be Instagram.”

This becomes a reflection on the limits of design under capitalism: even well-intentioned founders are outgunned by growth incentives.

4. Where AI Is Similar—and Where It Might Be Different

Eric and Ev agree: the stakes for AI may be even higher than they were for social media. But Ev is unsure whether different outcomes are possible unless we rethink governance and incentives.

Ev recounts advising several AI startups (like Anthropic), and being impressed by their early desire to embed governance structures into their companies.

But he cautions:

“Unless the whole system changes—the economic system, the incentives, the investor expectations—it’ll be very hard to make good choices consistently.”

5. Governance Matters (but It’s Not Everything)

Eric floats the idea that if Twitter had been a Public Benefit Corporation (PBC) or had dual-class shares, it might’ve avoided the Elon Musk takeover and made better long-term decisions.

Ev responds with cautious agreement:

“Yeah, that would’ve changed that… If the board wasn’t legally required to maximize shareholder value, maybe it wouldn’t have happened.”

But he also adds:

“Most of the mistakes didn’t come from being forced to do something wrong—they came from unconscious pressure to make the numbers go up.”

So while he supports stronger governance structures, he believes psychological and cultural pressures are often more powerful than legal frameworks.

6. Cultural Inertia and the “Make Numbers Go Up” Problem

One of Ev’s key arguments:

“The minute you get money and employees, the job becomes: make the numbers go up. That’s the default operating system.”

This culture:

Suppresses creativity
Forces shallow metrics
Discourages risk
Ignores qualitative user value

It’s not just about shareholder pressure. Even in private companies, teams push for short-term wins over long-term trust, because everyone internalizes growth as success.

7. Real Stakes: Life or Death

Ev mentions a company making character-based AI that received a letter saying the product saved someone’s life, but then later learned of a user suicide tied to a similar product.

“If you're in that driver’s seat, I hope you’re thinking long and hard about how to do more good than harm.”

The juxtaposition reveals the moral weight of building AI products—especially those that interact with vulnerable users.

Ev calls out the “underappreciated community of trust and safety teams” in Silicon Valley:

“They work really hard on policy and best practices across companies… and still, we’re in the state we’re in.”

Even with coordinated efforts, the structure of the industry still rewards engagement over well-being. The implication: individual good actors can’t outpace systemic misalignment.

Lesson	Quote or Idea
Human Nature is the Root Variable	“We were exploiting human nature… if we didn’t, someone else would.”
You Can’t Engineer Morality	“Healthy dialogue was the goal from day one. But incentives overpower design.”
Governance Helps, But Isn’t Enough	“Even I wasn’t aware of all the ways we were pressured to chase metrics.”
Trust is Expensive to Build, Cheap to Lose	“The login modal boosted sign-ups… and destroyed sentiment.”
AI Builders Must Be Introspective	“I hope the people building this tech are thinking hard about impact.”

Ev doesn’t offer clean answers—but that’s the point. His message is this:

“Technology becomes what it wants to become, based on the systems it’s built inside of. If we want different outcomes with AI, we have to start by changing the systems—economic, cultural, and organizational—that shape what we build.”