The world is too loud. Read what matters.

The a16z Show

Perfect broadband strategy lost to good-enough rough web pages

Disney, Microsoft, and Comcast bet on perfect set-top-box broadband, but lost to crude web technology simple enough that hobbyists could code it themselves.

Internet historyBusiness modelsPlatform competitionEntrepreneurshipAdvertisingTechnology cycles

The video won't play here. Listen to the audio instead:

This 2018 replay covers the internet's history from Netscape to iPhone. There's no new material, but the structural patterns about how the next wave of tech opportunity emerges still hold today.

The argument · tap a timestamp to hear it

7:21

Good-enough technology defeats the perfect solution

Bill Gates, Barry Diller, and others bet on the "Information Superhighway"—a top-down, perfectly-engineered set-top-box broadband scheme led by Comcast, Disney, and Microsoft. They got the timeline for broadband adoption right, but what actually prevailed was the crude Mosaic browser and amateur websites, because it was simple enough that college hobbyists could compete with Disney on equal footing. Gates didn't misjudge the future; he lost the bet on the path forward: he thought it was too geeky for ordinary people to use.

— Brian McCullough
11:30

eBay taught strangers to trust each other

eBay is often overlooked, but the author thinks it may be the most representative internet company. First, it taught ordinary people in 1999 to hand over their credit card information to a complete stranger. Second, it scaled a non-hierarchical, reputation-based trust system (star ratings) in a way that could grow infinitely—Uber and Airbnb later extended this same logic. Third, it was the first successful company that "owns no physical assets and whose business model is what users do on the platform"—this is the ancestor of every social media giant's valuation logic today.

— Brian McCullough
14:34

Napster was about convenience, not piracy

After interviewing the Napster team, the author found they were saying from 1998-99 onward: users want unlimited choice and instant gratification, not free music. If the record companies had been willing to negotiate instead of sue immediately, the media industry might look completely different—Napster could have pioneered video streaming, and YouTube and Netflix might never have existed. YouTube eventually won the music industry's approval because Google had already proven through AdWords that it could monetize; Napster had no leverage at the time.

— Brian McCullough
17:34

Google's most profitable business wasn't invented by Google

Google solved search, which was the first miracle. But its ad model actually came from GoTo/Overture; Google just made it better—by introducing quality score, letting advertisers pay less for more relevant ads while Google earned more from higher-CTR cheap ads. The entire industry consensus (including portals like Excite) was that too-accurate search would make users leave the site and destroy the ad model. Larry and Sergey did the opposite: they insisted on the best search with no business model, even once considered selling a $100,000-a-box enterprise search appliance.

22:37

Not knowing what works made the bubble inevitable

The early internet bubble's frenzy wasn't purely because of fraud—it was because nobody knew what doing business on the internet actually meant. Products, business models, user habits were all unknown. But around 1998, there was a structural shift: some VCs realized you didn't need a company that could succeed long-term—just one that could be packaged and taken public. When "get to IPO and cash out" started outcompeting long-termism, the system began rewarding that play. Combined with baby boomers hitting peak wealth accumulation and Wall Street capital needing a home, the bubble became structurally inevitable.

29:58

Mobile internet's low-hanging fruit is already gone

The author thinks we're in a "flat period" now. Around 2012, something new seemed to arrive every few days, but that boom came from one special condition: a sixteen-person team making a chat app with minor differences could race to a billion users. Now most categories are crowded. The next generation of founders won't win by "claim a category and scale"—they'll need to deliver qualitative difference within a category, the way the auto industry moved past "you can have any color as long as it's black."

— Brian McCullough
36:11

Big platforms kill innovation by copying, not buying

The two discuss a structural concern: in the Web 2.0 decade, Facebook could rise partly because Microsoft was reeling from antitrust and afraid to acquire, the internet bubble had just burst and giants were distracted, leaving ten years during which no one could eat these new companies. Now it's different. Snapchat invented an entirely new media form, but Facebook didn't even need to acquire it—it copied Stories directly into Instagram, and leveraged distribution to cap Snapchat's growth.

41:18

Only hobbyists and government fund decade-long bets

Most companies operate on quarterly or annual cycles; those willing to bet for two or three years already count as visionary. Only academia, DARPA-type government projects, and hobbyists have ever bet on ten-year timescales. The PC industry was built by hobby clubs like the Homebrew Computer Club. To figure out where the next infrastructure will emerge, find the people who don't think it'll make money but are investing anyway.

— Brian McCullough

In their own words · checked verbatim

The lesson, we're almost saying it, it's obvious. The lesson is sometimes just good enough technology is good enough.

Brian McCullough7:21

today we live in a world for media of unlimited selection and instant gratification

Brian McCullough14:34

It was the froth on the cappuccino of the longest bull market in history.

Brian McCullough23:37

it's almost like we're being held back by the fact that we have been so successful in tech and tech has taken over the world

startups try to find distribution for their innovation faster than the incumbents attach innovation to their distribution

Chris Dixon39:18

always look for the people that are doing this even though they don't think there's a reasonable expectation they'll get paid for it

Brian McCullough40:18

running Amazon was almost like this test that he was running to himself

Brian McCullough45:20

Figures

1999 CD industry revenue$20 billion15:34
Amazon stock price when Microsoft could have acquired it$5 per share36:11

Glossary

quality score
A metric in Google's ad system that adjusts the cost-per-click based on ad relevance.
Web 2.0
The era when websites shifted from read-only content to user-generated platforms like Wikipedia and Flickr.
low-hanging fruit
Early business opportunities on new platforms that are easy to copy and scale, often in emerging categories.

How to listen

Who it's for

Founders and investors trying to identify where the next infrastructure opportunity will emerge, and anyone interested in how early internet companies' business models evolved.

Skip

The opening small talk and the first two minutes of book recommendations can be skipped.