17 Things the Bloom Community Taught Me About AI

My first corporate talk

Today I gave my first talk to an outside company. And the audience was executives from 23 of Korea's top conglomerates and global companies. When Adobe first reached out, my reaction was, "It's already wild that I'd be in the room, and now they want me to speak?" Still, I saw it as a chance to grow a step, so I worked up my courage and got on stage.

Before the event, a lot of people told me, "The mood might be pretty serious," and, "Try to loosen things up a bit." I'm not a professional stand-up comedian, you know. I was honestly nervous about what kind of room it would be. But once I was in there, to be honest, it wasn't that different from the community events I usually run. People made eye contact and nodded along with the insights and questions I threw out. When I picked one person from each table and asked them to introduce themselves and share their thoughts, they all shared their views and experiences with real care. It reminded me once again that if you set the stage right, Korean people have a lot they want to say. I was also grateful that everyone really understood and related to my theme, going from AI Blue to Bloom.

One person recognized me and came up to talk first. They said they usually just follow me on LinkedIn and send other people from their team to our community events, but next time they'd make sure to come themselves. Meetings like that still amaze me, and they give me a lot of strength.

Last thing. When I thought of Adobe, I mostly thought of Photoshop and Premiere. I had no idea they were so actively building and selling CX orchestration services to enterprise customers. At big companies, customer data and work processes are so fragmented that actually pulling off AX (AI transformation) well is really hard. What Adobe does best is structure all of those pieces into a form AI can understand well. On top of that, it automates individual work, unifies workflows across the whole organization, and refines the context in detail. That part especially stuck with me.

I got to dust off a suit for the first time in a while. My sincere thanks to everyone at Adobe for giving me such a special experience.

The 17 insights

#1 FOMO is a thrilling marketing tool, but in the end it only lasts a moment. It might even be a poisoned chalice. That said, I don't yet see a clear solution that fully resolves AI Blue. Even so, people are more than satisfied just to talk and connect with people like them, and they gladly pay money and time for it.

#2 People aren't that interested in how to use AI well anymore. They're much more interested in cases where someone in another role or field used AI and actually grew a business. With the developer role possibly disappearing and the lines between roles blurring, even developers now care more about "How can I make money with AI?" In fact, more and more developers are switching careers into business roles like sales and marketing (DevRel).

#3 Vibe coding means anyone can code now, but if you're not good at it, nobody can build a product. Building a real product takes a huge amount of knowledge, and people no longer build up as much knowledge as they need. In the AI era, you have to study more and be smarter. We're heading toward a world where only the real experts survive.

#4 AI's speed makes us more anxious. AI gives you an answer in a few minutes at most, but when a human spends 30 minutes thinking something through like we normally do, we strangely start to feel like we're falling behind the AI. In the AI era, sometimes speed matters more, and sometimes a human's offbeat creative planning matters more. Telling the two apart seems much more important than it used to be.

#5 AI is just a tool. Going forward, I think three things will matter more and more.

1) Communication skills, meaning the ability to understand what the other person wants. (Most people are only interested in what they themselves are doing.)

2) Your learning curve, which I think requires humility and curiosity as preconditions.

3) Sharing (open source) is how you become number one. The hardest asset to get is influence, and you only earn reputation by sharing and giving first.

#6 An eye for good work becomes more important. But if we hand everything to AI, we gradually get duller and don't even notice it happening. We're the kind of beings who make mistakes and learn by being corrected. Rather than demanding instant answers from AI, it's better to use it by going back and forth with questions and answers and learning together. You might not know the process, but you have to understand the plan and the result, and we always have to be the ones responsible for the steering wheel.

#7 AI got smarter, and we're getting more worn out. Human attention has clear limits, and in fact our physical stamina rules our mind and heart. "Misaeng", the Korean drama, puts it this way: "If there's something you want to achieve, build your stamina first." Stamina matters more than anything. We can't forget that we're human, not AI.

#8 As individuals, we may lose the desire to get better at our work. The value of the work humans create is dropping, and vague worries make us shrink. Maybe we should build services that help people break out of the office, instead of services used in the office. For example, a service that helps people become solo influencers.

#9 AI is creating the most dangerous gap of our time. The reason is that the gap in how deeply people actually immerse themselves in AI, between the 1% of early adopters and the 99% of the public, is far more extreme than people think. Developer work was relatively structured, so it was the first to start getting replaced, and fast. Non-developer work is unstructured, so change seems slower, but a much bigger wave may be coming. The Bloom community hopes to give non-developers an early heads-up about the AI wave that developers hit first in 2025.

#10 We need crazy ideas that will change the world, beyond AI itself. AI does almost everything now, and you can see the end of the tech software business. It might be better to take AI as a kind of "spirit" that breaks fixed ideas, more than as a technology in itself. For example, in the US, a service that puts mirrors on satellites so you can "subscribe" to sunlight raised a big investment. As a psychology major, I even imagined a service that uses hypnosis to implant the experience of having immigrated, without you ever actually moving abroad.

#11 In the AI era, people with ADHD-type traits may actually rise. The biggest hurdle to AX is time. On the ground, people have almost no time to invest deeply in training. That's the reality. Juniors who multitask, handle lots of things quickly at once, and have low switching costs may well stand out more. They keep throwing one thing after another at AI, judge the results fast, and move straight on to the next task.

#12 We're more honest with AI. According to an Anthropic report, people in East Asian countries, where saving face matters a lot, answer AI surveys more honestly. Soldiers in the war in Ukraine say it was AI that pulled them out of hell. Humans becoming more honest with AI: I think this will only speed up.

Developed countries fear what they'll lose.

Developing countries look forward to what they'll gain.

The East worries about a crisis of relationships (I disappear).

The West worries about a crisis of control (what's mine gets taken).

#13 We have to make people's dopamine explode at a primal level. There's an American service that got famous for "cold cakes instead of cold emails." It actually sends custom cakes to potential customers. It has nothing to do with AI directly, but people love it, so it went viral in Silicon Valley and hit more than 2 billion won (roughly $1.4 million) in monthly revenue. Spotting people's primal reactions and turning them into a business is always very powerful.

#14 VC investment criteria have changed in the AI era. VCs used to ask whether the market was big enough and what the TAM was. Now, because of AI, the size of every market is unpredictable. Nobody knew the coding assistant market would be worth trillions of won. So they focus more on whether the founder is ambitious, persistent, and able to execute enough to change the world, and they invest in determinate optimism: people who design and build the future.

#15 In Silicon Valley, people don't hand their writing to AI. (Which means they hand it almost everything else.) Writing is more than a task. It's my mirror. Writing is how I express my thoughts and my sense of self, and if I hand it to someone else, my own thinking disappears. We also communicate with computers through writing, in the form of prompts. If we stop writing our own prompts (voice coding counts as prompting too), we may no longer really be the ones using AI.

#16 These days the moat is marketing and distribution, more than a technology gap. To put it in extreme terms, if you can't build a GTM channel, building the product is pointless. Try selling yourself before you sell your service. If you can't market even yourself, the easiest thing in the world to explain, you may not be able to market your product either.

#17 We don't consume products as they are. Who made it, and with what thinking and background, the brand's whole story, is part of the product. Don't agonize over whether to use AI. If AI builds it better than we do, just use it. Humans should focus on what comes before and after: the narrative, the branding, and the storytelling.

Originally written in Korean on Brunch, June 28, 2026.