How are we blooming in the AI era?

What was discussed

Nam Sung-pil, CEO of AB180, which hosted the event, said he spends 90% of his day with AI and had been feeling a fair amount of Blue, which is exactly why this gathering resonated with him. The heart of his short welcome was "inspiration." Doing anything takes will, will comes from inspiration, and inspiration ultimately comes from interacting with other people. However good the tools get, the desire to actually use them comes from people, a line that neatly summed up why this meetup exists.

Lee Dong-gun founded MyRealTrip in 2012, at 27. He prided himself on reading mobile faster than anyone, but he decided a trip costing several million won would never be paid for on a phone, and that was his big miss. The smaller latecomers could only build apps, which turned out to be their weapon, and they caught up within 10 months. Looking back, he said "my imagination as a CEO was pitiful," and resolved that when the next wave came he would imagine it as destructively as possible.

That next wave arrived at the end of 2022 with GPT-3.5. Convinced once he saw a travel scenario in every AI demo, he pulled the team together over a weekend and shipped a live service within 48 hours of the announcement. It made enough noise that the 9 o'clock news and Microsoft headquarters came calling, but the traffic drained away in three or four days. It was too early to hand a high-involvement decision like travel to an AI that hallucinated heavily, and the conclusion was that "we had bolted on a feature, but the way we worked hadn't changed."

After that he focused less on building things and more on changing execution structures and the organization itself. An AI Lab in 2024, an AI Champion in every team in 2025, and then in 2026 MyRealTrip declared that it had moved past "AI literacy" to working as an "AI native" company. His definition of an AI-native organization has three parts: work simply doesn't run without AI, a small elite team delivers bigger impact, and it is the kind of company that grinds to a halt when Claude goes down, not AWS. Evaluation follows the same logic, not metrics like token usage, but how much AI has improved fundamental numbers such as contribution margin and conversion rate.

The high point of the keynote was seven real cases from seven teams. LuckyGlide, which finds the cheapest airfare; MRT Biz, which books business travel through Slack; Korean Foodies, which curates restaurant reviews, all of them running in live products and operations. One principle ran through all of them: "the person who recognizes the problem and the person who solves it must not be separated." After watching an AI Lab that built things on everyone else's behalf degenerate into an outsourcing shop, they landed on this: whoever spots the problem should solve it themselves, and the organization's job is to remove the barriers in the way: engineering, AI, cost. The Q&A produced an equally candid answer: "the biggest bottleneck in AX is, surprisingly, the CEO himself."

What the 16 tables across two rounds of roundtables had in common was a focus on the essence of the work rather than the performance of the tools. LG H&H pointed out that output has exploded but connecting it to outcomes is a separate matter; Rebellions offered that "tokens are the cheapest they will ever be"; and several tables converged on the keyword "harness." The more effortlessly AI digests everything available online, the more paradoxically precious the density of an offline room becomes, a place to share what is in your head, refine it, and build on it together. That was why 100 people showed up on a Saturday morning.