The Rise of ADHD-Type Talent in the AI Era
Every company runs at a different AI temperature
The surprising part: everyone at the table works at a household-name tech company, and yet the temperature differs this much from company to company.
Company A rolled Claude out company-wide and actively encourages it, wired into the messenger, the issue tracker, and the wiki, with other AI tools open as well. Company B bans Claude and allows only one specific coding tool. People know that attaching AI to the messenger is a new world, but the company blocks it, so they cannot even have the experience. Content and creative teams resist adoption outright, which complicates things further.
The internal gap is extreme too. A colleague from the next team quietly asked what AI they should use, and it turned out they were only using a chatbot. With almost no internal training, everyone learns alone, so the usage gap between employees is enormous; the higher the title, the more likely someone still works by hand. The best users are too busy with their own work to teach, and their know-how is personal, so it does not spread. The gap widens instead of closing.
Company C is entirely team by team. The customer support team adopted well and handles repetitive work efficiently, while the data team still analyzes everything manually on the grounds that internal data is too unstructured for AI. If this is the spread inside the tech industry, what does the rest look like? And is this temperature gap quietly becoming the future competitiveness gap?
The biggest hurdle is time
The most common line of the night: we know AI is good, but really using it takes three or four hours of deep focus, and that time does not exist.
At Company A, meetings run wall to wall from morning to evening. The FOMO is huge, but there is physically no time to dig in. Meanwhile executives post daily in the leadership channel about how many lines of code they wrote with AI, and when working-level staff say they have no time, the response is: do you have time at home? Internal lectures keep getting scheduled, and nobody has time to attend them either. A vicious cycle: no trigger, no time, no experience. And yet the gap between those who somehow made the time and those who did not is already enormous.
Walking into work feeling covered
The representative power user was from Company B, using a coding tool almost as a secretary. All work data lives in organized local folders; a scheduled job reports each morning only what needs doing that day. When a meeting ends, the notes summarize themselves, get distributed to owners, and update the wiki, fully automated. That alone cut the workload sharply.
The operations case was striking too. Entering country-by-country data used to be fully manual. Now AI does it, instructed to mark everything one of three ways: certain, estimated, or blank if unknown, color-coded so reliability is visible at a glance.
When I get to work and the tool is open in the side window, I feel so covered.
Like a trustworthy colleague permanently seated next to you. People said earlier that they cannot use AI for lack of time, but invest the time once to set it up, and time is what you get back.
Revenge of the eighteen-page document
Does everything improve as more people use AI well? Not quite. At Company A, juniors now stamp out product requirement documents with AI, and the volume has ballooned: eighteen-page documents keep arriving. The problem is the senior who has to review and catch everything is drowning.
AI raised productivity and the verification cost exploded, an irony. Reviews could be answered with AI too, but people have not dug deep enough to get there yet. Building with AI and verifying with AI has arrived, and the seniors caught in the middle are suffering most.
The ideal employee profile is shifting
The most memorable story of the meetup came here. The person from Company A said their strength had always been deep immersion: pick one thing and dig to the bottom. But the deep part is exactly what AI now does better.
Instead, multitasking juniors who switch contexts quickly are standing out: throw things at AI, judge the results fast, move to the next thing. That does not make depth meaningless. It changes where depth goes. The depth of execution moves to AI; the depth that remains for people is deciding what to do and judging what comes back.
Join Bloom
Bloom builds offline rooms where people and technology meet. We run them in Seoul, and now beyond it.
- 💬 Discord community, where events are announced first
- ▶️ YouTube, full talks from past events
- 📸 Instagram, photos from our events
- 🎟️ Upcoming events
