What Bloom Is

At lunch, like always, I opened my laptop and pulled up something on YouTube. I just wanted something to listen to while I ate, so I clicked on a video from BizCafe, a Korean business media channel. A few minutes in, I was slapping my knee.

I run Bloom, an AI community with 6,800 unique applicants, maybe the biggest in Korea. But I'm not a community expert, and our company didn't start out as a community events business. So I've kept things low-pressure: I invite speakers on topics I'm personally curious about, we chat in what we call a fireside chat, and then we set up roundtables so the people who came can talk among themselves.

But as the scale grew, as we repeated this almost 30 times, and as more companies started sponsoring us, I felt more responsible. I started polishing things more, paying more attention, and worrying more about what people think.

Of course, being mindful of attendees and partners is necessary and always a good thing. But I felt that unless I understood the core of why 300 to 500 people sign up every time, and why more than 150 actually show up, neither attendees nor partners would get a good experience. I knew I needed to think about this, and honestly I'd barely done it.

Then I watched this a16z video and something clicked. At first I expected the usual line about everyone needing to become an influencer. Instead they said that even government officials in Washington, DC don't read legacy media like the Washington Post anymore. They only follow three-hour podcasts and newsletters. Newsletters are pretty much dead in Korea, but this is the golden age of YouTube, so either way, new media beat legacy media a long time ago.

The key thing about new media isn't just being on social media. It's whether what a speaker says privately to friends matches what they say on stage in front of us or in a video. In a way, it's honesty. Legacy media is rigid and restrained, so it could never do this no matter what. That's why people got sick of it and walked away.

In Bloom's early days, someone like me, with basically no media training, would get up and talk and trade questions. Honestly, I asked because I was curious, but I wondered how closely anyone would listen. Now it makes sense. Since I'm not an expert in anything, I ask the kind of questions friends toss at each other, and that turns out to be rawer and more fun. And since I'm not a celebrity, I post my thoughts honestly on Discord or LinkedIn, and if people don't like it, I apologize and get on with my life. That attitude is apparently what makes it unusual, something you at least won't see at a traditional event.

I'd been worrying a lot about how to run events more professionally. I even wondered if I should take public speaking lessons. I don't know if this video is the answer, but a16z's point is that if interesting people talk about interesting topics, people will watch, even with no script and no polish. So I figured I just need to keep doing what I always do: find interesting topics, talk with interesting people about them, and create a raw atmosphere so the guest doesn't stiffen up.

I'm not sure this counts as a strategy, since it's basically what we already do. But it gave me the nerve to stick to my guns and push it further. If that's why people come, then that's what we'll do. If people are happy, we're happy and our partners are happy. So I'm going to try it.

Below is my summary of the video. It's a conversation about the new rules of new media between a16z cofounders Marc Andreessen and Ben Horowitz, and Erik and Gabby, who handle media at the firm.

People in Washington don't read the Washington Post

The story starts in Washington. At a gathering with a lot of Washington people, senators included, they asked what everyone there actually reads. The answer was a newsletter written by one person. New media, in other words.

This is surprising because everyone believed Washington was legacy media's last stronghold. Even inside the firm, they assumed that if old media still had an audience anywhere, it would be Washington. Even that turned out not to be true.

Say what you'd say over lunch with a friend

The first rule of new media, they said, is authenticity. More precisely, can you say on camera exactly what you'd say behind closed doors? The most common mistake founders make is being too buttoned-up and trying to look too media-trained.

Marc Andreessen brings up his experience from the '90s. There were no blogs and no YouTube, so if you ran a company you had to go through the established media, and everyone got media training. He called it the most painful experience in the world. The results were always strange. Important people would go on TV to say important things and all look like plastic dolls. The host used an anchor voice and the CEO picked the most harmless words possible.

The best part is how CEOs back then judged whether an interview or speech went well: by how little controversy they created. They'd walk off stage proud that they hadn't made a single headline. The assumption was that all news is bad news.

But when he hired the media trainer considered the best at the time, that person told him the opposite. A former producer at 60 Minutes, the trainer announced they wouldn't do classic media training. Instead they would do the one thing everyone underrates: get him to say in public exactly what he'd say over lunch with a friend.

Marc said it felt like his head cracked open. It was such obvious advice, and nobody had ever told him. The trainer's logic went like this. If you're on a stage or in an interview talking about a topic you don't know inside and out, what are you even doing there? Only talk about what you know deeply. If you know it deeply, you'll instinctively talk about it in an interesting way, and you'll naturally come across as an interesting person. Just like talking to a friend across the table.

There was one more piece. On the premise that the press is hostile, you don't answer the question you're asked. You always answer a question you made up yourself. Marc admits this part is a bit unnatural, but says that device, refusing the bad question and slipping in a good one, is what made the whole approach work.

After watching this field change for 30 years, his conclusion is that the best communicators today are doing exactly this. The rise of three-hour long-form podcasts, and the dead simple idea of interesting people having interesting conversations, used to be radical and are now common. And what that delivers is authenticity. Does the viewer feel they're meeting the real person?

Old media plays defense, new media plays offense

Traditional media gave itself two jobs, they said. One was fair, objective journalism: a voice that belongs to no side and shows both sides. The other was speaking truth to power, often summed up as comforting the afflicted and afflicting the comfortable.

The problem is that the second swallowed the first. Their diagnosis is that this flipped the media's character. It crossed the line between objectivity and activism and turned into a way of scaring you out of saying what you think.

Marc says he did a huge amount of traditional media from 1994 to 2017. There was the occasional hit piece, but 90 percent of the time he was glad he did it and felt he'd been given a chance to tell his story. For most of that period, people thought startups and tech were cool, fun and good for the country, and reporters and editors sincerely believed they should explain this well to readers.

Then, he says, things changed starting in 2017. Since then the side with an agenda has had the upper hand completely, and he doesn't think it's going back. So they tell founders there's basically no way left to get a story you'll like through traditional media. You might land one now and then, but you can't run a strategy on that. To people who cite individual success stories, he says that's the exception that proves the rule, and there are 99 others that went the other way.

So the new approach has two parts. One is going direct: telling your own story through your own channels and your allies' channels. The other is new voices and new media themselves.

Ben Horowitz adds a fun contrast. Even in the so-called golden age of media, there was only one way to watch a smart person talk for an hour. In the '80s, all you had was a late-night talk show that aired at midnight, and with no VCR you had to stay up until midnight to see it. Going from that to today's world of podcasts and newsletters is unbelievable progress. Even so, people still have a nagging sense that legacy media carries authority and prestige. He flatly says he doesn't believe that anymore.

The brand becomes a person, not a company

This was the part that struck me most. Ben compares old and new media structurally, and his breakdown is very clean.

Old media had very few channels and limited formats, and above all, the brand was the company. You could talk to the New York Times or CNN, but you'd be boxed into one quote or a very short interview, and you'd even end up speaking for a brand instead of yourself. So media strategy came down to getting your name out without irreparably tarnishing the company. A big rule of old media was that you could never take back what you said, so you had no choice but to play extreme defense.

New media is the opposite. Formats and channels are unlimited, and the brand is now a person. He gives an example. When people got angry at a certain space company, the one they were actually angry at was the founder, not the company. Is it Palantir, or is it Alex Karp? It keeps happening that way. So the companies winning at marketing are the ones whose brand is a person.

Then comes the rule of new media. It has to be a person, and that person has to be interesting. Otherwise you get buried. He adds one line that nails it: old media is dangerous because its only rule is don't be interesting. Being interesting is the worst thing you can do there.

There's a condition, though. That person has to be someone who's at the organization essentially permanently. A VP of marketing who'll be gone in three years absolutely won't work. a16z itself is the example. Technically Ben is the CEO, but the brand is closer to Marc, and Marc does far more media. Since the firm is named after both of them, outsiders can't tell the difference, and that's fine.

What happened when everything had to fit through a narrow straw

Marc adds an explanation from the history of technology, and I felt this was the smartest part of the conversation.

Until the 1930s, he says, companies didn't have corporate brands. It was the Ford Motor Company, the Edison Electric company. Just the name of the person on the door. Henry Ford would never have thought of naming his company anything other than Ford Motor Company. Then through the '30s and '40s, big companies with abstract names appeared, followed by a flood of consumer brands, and that led into the conglomerate era of stacking up as many brands as possible.

In his view, this happened because of the rise of centralized media. Before the 1930s, media was extremely decentralized. Every town had a dozen newspapers and small radio stations were everywhere. Then starting in the '30s and '40s, media became more centralized than at any point in history, something he describes as a once-in-a-thousand-years event.

Centralized media, he said, is like drinking everything through a very thin straw. There were only three networks and 24 hours in a day, so the airtime you could get was fixed. Same with space on the front pages of the three big newspapers. So a company's message had to be compressed down to the smallest possible unit to get through that narrow straw. That, he says, is what the corporate brand was. You had to shrink down to the atomic unit of a brand to get through.

The problem came next. People came to believe that was just how the world works. But it only made sense in a world of centralized media, and since that world is now unraveling and falling apart, a new approach works and is also necessary.

Can you last three hours?

The example that best shows this shift is the long-form podcast. Should someone running for president go on a three-hour podcast? Up through 2024 the answer was absolutely not, but since then it has become 100 percent yes. In fact, in the post-mortems of that election, not doing that appearance came up as a big factor, and people who work in that world now think the next candidate has to do it.

But going on means being able to talk about any topic for three hours. That's the new baseline. You have to be interesting, and you have to be someone who can pull it off. If you can't, that's a big marketing deficit, and you end up putting a ceiling on your own opportunities.

Here the founders' practical worries come up. Running the company is already exhausting, and now it feels like they've picked up a second job as spokesperson. Founders who have spent their whole lives in a lab or in front of a computer were never trained for this. The answer is cold: there's no way around it. It doesn't have to be the CEO, though. Someone else can take on the role as long as they're going to stay with the organization for a long time.

What to fight and what to let go

Someone asks what to do when people say bad things about you. Ben admits it honestly: when someone says something negative, he always wants to respond. The question is whether responding improves your position, or whether it's bait from someone hoping you'll amplify their dumb opinion for them. Marc jokes from the side that responding always makes things better, and everyone laughs.

Without discipline, he says, you'll spend your whole life just reacting, because the bigger you get, the more people come at you. The extreme bad case is replying one by one to accounts with fifty followers. It's a huge waste of time, and you end up amplifying people who have no audience.

It's different when a serious opponent takes serious aim at you. That's a real chance to build your brand, because you get to show your point of view and where you stand. Ben says a large part of the brand they've built came from responding to attacks. His favorite piece he ever wrote was a rebuttal from a fight with a major newspaper, and it became his most-read post at the time. People love a fight, so that one shot made his position jump. Just pick your opponents carefully so you don't end up building their brand for them.

Their view on being hated is clear too. People should both hate you and love you. You shouldn't be lukewarm and neutral, because neutral means boring. As you grow, jealousy and hate always follow, and past a certain level you'll never again get only favorable coverage. But that's a good sign. It means you did something important, left a mark on the world, and people care. If you stop what you were doing to make people like you, you wreck your marketing. In Marc's words, it's important to have the right enemies.

Build the team differently

Asked how to put together a marketing and media team in the new media era, Ben's first principle was firm: it's very hard for someone trained in old media to do new media. If you spent ten years in old media, its rules are baked into you like the laws of physics. Everything from how you vet reporters to how you handle Q&A is the opposite in the new world, and very few people can walk out of the old one. It looks like marketing, but it's a completely different skill set.

So what do you hire for? Whether they've done it. Have they built a brand and an audience with something? Do they know how to build an audience? That's the core, and without it, no number of other skills will help. The people they actually brought in were product managers, founders and investors, but they were people obsessed with the discourse. They listened to podcasts, wrote on their own, and already had a body of work that nobody had asked them to make.

So the core skill comes down to storytelling. Even setting marketing aside, can you create a story someone wants to read or hear? It's a real skill, the gap between the top and the average is huge, and people who can't do it never will. It also explains why people from the old marketing world lack it: their stories always depended on a protagonist like the CEO. A team with several good storytellers becomes a huge amplifier.

Here Marc points out an important difference in how people see things. Media people tend to think the world has only two things, journalism and propaganda, so if you do something yourself, it's propaganda and just marketing. They don't 100 percent agree with that split, because there's a positive reason to speak authentically: you want people to actually understand who you are, what you do, and the context you do it in. So storytelling here doesn't mean making things up. It means taking what's actually happening and giving it tension, a beginning and an end, so people follow it all the way through.

Marc puts this almost as a duty. The changes happening in tech right now are profound, really hard and complex to understand from the outside, and there's a huge amount of noise. So it's the responsibility of people in this industry to explain honestly what's actually going on.

Message before distribution

Gabby sums up the mistakes founders commonly make in two points, and both sting.

The first is getting overly absorbed in distribution and tactics before you've nailed the message. The timeline is so addictive that it's an easy trap. But distribution is just a multiplier on the message. If the message is wrong, you've amplified something unrelated to your business or something your audience doesn't need to hear. In the worst case, you've told everyone you care about that you're not a very interesting person.

The questions founders ask most often are how to go viral and how to get on a famous podcast. But if you think about it, you could go on that podcast and say the wrong thing, and that might be the worst outcome of all. This isn't just an old media problem. It happens exactly the same way in new media.

The second is focusing on the material instead of the outcome. Ironically, companies that are doing really well struggle with this the most, because they have too many interesting stories to tell. Lots of milestones, a great mission. But just because everything is true doesn't mean all of it is relevant or strategic for the business.

So she says to flip the order. Instead of laying out everything you could say and hoping people remember one thing, start from the outcome. Do you want to sell to a certain kind of enterprise customer? Do you want to hire engineers who hold certain beliefs about you and your role in the market? Then you can work backward to the message: who those people are, what they believe, what they need to believe about you, and what feels urgent and personal to them right now. Put strong distribution behind a message built that way, and that's the winning strategy.

Ben ties this back to the team. If you build a team you can't talk with about the message, you're hurting yourself. Tossing someone an empty box and saying "market this" seems easy but is actually hard. The team has to be invested in the message itself for every variation that comes after to be accurate, and it gets very hard if the team has no opinion about the message.

From the outside in, not the inside out

The best part of the conversation is the last section.

Marc first pokes fun at the two most common stories. A venture capitalist's self-marketing usually goes like this: I met a great founder, we went for a walk, he called me eight times, and in the end I convinced him to take my money. A startup's story usually goes like this: we're a new company, we're all super motivated, we have a new product, please try it. He said hearing stories like these makes him want to stab himself in the throat.

What he zeroes in on here is detail. Good storytellers have great details, and details really matter. Those stories are terrible because they have no details at all. It's just "I'm great, some founder liked me, and you'll probably like me too." Nobody cares about that.

Then there's fake humility on top. The moment you say you're humbled, you show you're not humble at all. Listing your criticisms is a humblebrag. It's boring, aimlessly self-centered, passive-aggressive, and above all not distinctive. Every startup can stare at its own navel, so that's not a story that sets you apart.

The problem is that telling a bigger story feels arrogant. People hesitate, thinking who are we to tell a bigger story, so they never widen the lens.

His conclusion is exactly the opposite. The story of you and your startup isn't inherently interesting, but there is almost certainly an interesting story your startup is connected to. And that's the cheat code. That story is something else, something hugely interesting happening in the world right now, and your company is connected to it.

The grand wizard of this, in his view, is Palantir's Alex Karp. He doesn't talk about Palantir in interviews. All he says about his company is two pieces of jargon nobody understands, so nobody knows what Palantir does, and it doesn't matter. When there's a really interesting topic, like the future of the US military, superintelligence or neurodiversity, he goes out to talk about that. And because he's Palantir's CEO, it works.

Another is Ryan Petersen of the logistics company Flexport. It's the difference between talking about freight and saying that during COVID the global supply chain was collapsing and we were all about to starve. So he became the guy riding in a helicopter on a news show, saying look at those ships, our kids are going to go hungry. He doesn't even need to say "use Flexport." That follows on its own.

It works even more directly in enterprise sales, he said. Am I important enough to meet the customer's CEO? Important enough to be in the room with the decision maker? Important enough to sit with top government officials or big-company CEOs? Being a small startup building something interesting won't get you there. But if you've been talking about big, important things happening in the world and might have answers to them, that's bait almost anyone finds hard to resist.

So here's the trick as he puts it. People naturally think from the inside out: starting with me, my company and my product, then moving out to the world. Don't think that way. First think about what the most interesting thing happening in the world right now is, then think about how it connects to you.

Ben adds where CEOs go off the rails. They get so absorbed in what they're doing that they only want to talk about themselves and don't pay attention to what's happening in the outside world. That's understandable when you're building something, but if you want to market, investing in understanding what's happening outside really matters.

Marc wraps up this section with investor relations, and the comparison is very powerful. Public companies are required by law to produce all kinds of disclosure filings, write annual letters and explain everything the company does. But what percentage of Palantir investors have read those filings, he asks. Basically zero. Meanwhile, what percentage have watched Alex Karp talk on YouTube? 100 percent. So take a fraction of the time that goes into those filings and spend it understanding the world well enough to tell a good story.

Not something you're born with, something you grow

The last thing they talked about was the most comforting part for me. This isn't an innate talent. It's a skill you can develop, learn and grow.

If you look up Alex Karp's old interviews, they say, he's completely different from now. Look at the old tweets or old interviews of the people considered great at new media today, and they're nowhere near their current level. Of course some people are naturals and some less so. The conversation ends with an example of someone who gave very restrained, old-media-style interviews in the '80s and later completely figured out new media, and the point that you can develop this even at that level.

Looking back over all of it, it comes down to one sentence. An interesting person talks about an interesting topic they know inside and out, just the way they would in private. No script, no polish needed. And the skill gets better the more you do it.

Originally written in Korean on Brunch, July 24, 2026.