This video is a detailed field report from September 2025 on how AI is fundamentally disrupting organizations, labor, and the very nature of work. It covers BFACTORY's team dissolution experiment, the emergence of 100x engineers, changes to organizational structures, contracts, and talent models driven by AI, and the coming 'cognitive revolution' — all told through firsthand experience with a warm perspective. The key messages center on the role of humans, inequality, and the importance of learning and 'unlearning (updating)' that we all must prepare for in an era that treats AI as far more than just a powerful tool.
1. Opening and Key AI Industry Issues in 2025
Hosts Chester and Seungjun greet each other after a long time apart in September 2025, noting the need to organize recent developments. As they say, "We've been too busy with company work, each on our own" — they're experiencing the pace of change firsthand. Google Korea has been actively encouraging the use of a new AI tool called 'Nano Banana,' and new coding AI tools like Gemini and Claude Code are rapidly transforming how software work gets done.
"At the start of this year, people said '2025 is the year of agents,' and now we're truly in an era where agents are pouring out everywhere."
In Korea specifically, discussions around Sovereign AI and agile organizational transformation are also emerging. They explain that at this moment when the entire industry is being shaken up, the transformation is most tangible in areas directly connected to actual work like coding and design.
2. The Broad Spectrum and Receptiveness of the AI Industry
AI industry professionals display an extremely wide spectrum from experts to complete beginners. The two hosts self-deprecatingly note that with understanding levels varying so widely among participants, 'we're all struggling together.'
"Few industries have such a stark knowledge gap from 0 to 100 as the AI field."
They go on to discuss visions like 'superintelligence' and 'Sustainable Abundance' following AI's popularization. Elon Musk's Tesla Master Plan Part 4 announcement is mentioned, painting a picture of a world where robots replace all physical labor, and the idealistic notion of "a world where everyone can just play" is heating up.
3. The MIT '95% of Companies Fail at AI' Controversy and the Essence of Success
They emphasize that the recently viral MIT report about "95% of corporate AI pilot projects failing" was misrepresented. The report actually focused not on the "95% failure rate" but rather on the decisive difference in the successful 5% and the real starting point for AI adoption.
"The key message is 'it has truly begun.' Even if 95% fail, the success of the remaining 5% is far more important."
The real success factors were not simple ChatGPT prompt-based approaches, but rather the adoption of substantive agentic systems and field-centered, automation-focused pilots rather than top-down initiatives. An interesting reality was also presented: regardless of official adoption, 90% of employees at most companies are independently using paid AI services — a phenomenon known as 'Shadow AI.'
4. BFACTORY's Team Dissolution and One-Person-One-Project Experiment
Host Chester shares the radical experiment at BFACTORY in early 2025 of dissolving teams and transitioning to a one-person-per-project system. Previously, product teams, AI teams, leaders, and PMs worked through consensus, but communication speed and wait times led to the conclusion that "the entire organization is limited by the slowest person."
"We realized communication was the bottleneck, so we dissolved the teams."
Even the '100x engineer' with exceptional AI expertise and productivity initially resisted adopting a new tool (Claude Code), but after trying it firsthand said "now all I need is Claude Code," and the entire company experienced rapid evolution.
5. 1x, 10x, 100x Talent - Defining New Talent Models for the AI Era
They propose that organizations can now categorize talent into 1x (average), 10x (outstanding), and 100x (overwhelming) levels. 100x engineers in particular are people who know how to "design AI itself as a harness and maximize compute leverage."
"AI is the greatest leverage of our era. It's bigger leverage than money."
What's significant is that many '10x' talents are deployed directly into production work through AI usage, skipping repetitive tasks and raising the productivity bar to 10x levels. Meanwhile, some past-era talent still insists on doing things the old way — 'it's uncomfortable,' 'I want to go at my own pace' — and they either resign or don't move at all.
6. How AI Is Changing Internal Contracts and the Talent-Company Relationship
100x engineers comfortable with AI have reached a point where a single person drives company-scale projects. As a result, they've come to realize that "salary and stock options aren't enough — stronger incentives and new contract structures are needed."
"We're entering an era where one person equals one company. Future talent contracts must evolve into new forms that don't confine people within corporate frameworks."
OpenAI's launch of an AI talent recruitment site (similar to LinkedIn) supports this trend. Furthermore, they emphasize that the core competency of the new era lies in 'business acumen' — the ability to read market opportunities — rather than technical skills alone.
7. The Evolution of the Junior-Senior Debate and the New Formula: 'Attitude x AI = Performance'
The debate emerges about whether seniors are even necessary with AI. In the past, experience and attitude had a multiplicative effect on performance, but now AI x skill is the multiplier, and attitude is no longer a multiplier but an additive (bias) factor.
"In the AI era, the performance formula isn't attitude times skill — it's the product of AI and skill, with attitude merely added on."
In other words, a person without skills won't produce results just by using AI — baseline competency must be above zero.
8. Productivity Gaps, Hiring Market Changes, and the Surge of Solo/Micro Startups
With the spread of AI tools, the bar for 'core talent' has risen from 10x to 100x, and hiring in CS (computer science) has dropped dramatically. "Startup teams used to default to 10 people, but now 1-2 people, or at most 4-5" has become the norm.
"The era when startup labor costs were 70% of total investment is over. Now 1-2 people do the work of dozens."
9. Sequoia Capital's 'Cognitive Revolution' and Industry Restructuring
Citing Sequoia Capital's '$10 trillion AI revolution' narrative, they discuss the revolutionary inflection point. Just as the steam engine emerged during the Industrial Revolution and assembly lines took hold 200 years later, this cognitive revolution will convert all knowledge work into AI services at a far faster pace.
"If the Industrial Revolution took 300 years to unfold, the cognitive revolution might be done in 10."
They foresee the SaaS-ification of knowledge work and predict that new moguls like Rockefeller and Carnegie will emerge from AI startups.
10. Organizational Dissolution and Maximized Productivity — The Emergence of the 'One-Person Unicorn'
"Organizations will dissolve, and hyper-efficient solo entrepreneurs will collaborate remotely." Instagram being acquired by 13 people for $1 billion is old news — now they forecast that "one-person unicorns are not out of the question."
"We're truly heading into an era where one person does the work of 100 or 1,000."
Through the abstraction of productivity (programming, frameworks, etc.), the time it takes for humans to build wealth keeps shrinking, and they emphasize that "now is the historic opportunity."
11. Mass Layoffs of Middle Managers and AI Stratification — The Possibility of 'AI Feudalism'
Citing Google's mass layoffs of middle managers as an example, they note the dramatic decrease in communication costs and the acceleration of the 'manager equals bottleneck' phenomenon. Translating this into class dynamics, they express concern about "a new class society where only the 10% who control AI computing resources work normally, while 90% subsist on taxes."
"Only the 10% who possess AI capabilities and computing resources become the class that truly works, while the remaining 90% become digital serfs."
They also note that Silicon Valley's fierce competition has paradoxically enabled more equal access to information.
12. Redefining Value and the Rise of 'Homo Ludens' — Attention and Subscribers as Capital
They argue that economic, education, and social systems themselves must be completely redesigned around 'new value standards' rather than labor-hour-based metrics. Human play, creativity, and an 'attention economy' where subscribers/followers become capital will be the fundamental unit of future society.
"In the future, your subscriber count will be your capital, and prosumer micro-economies will become the everyday standard."
13. Adaptation Fatigue and the Importance of Rapidly 'Unlearning'
They repeatedly emphasize that in the AI era, individuals' ability to quickly learn and 'unlearn and move on' is more important than anything.
"New observations keep emerging that render all existing understanding meaningless, so 'forgetting and relearning' is absolutely critical."
Only individuals who catch the flow of change and adapt quickly will survive.
14. Closing
The hosts emphasize that thinking and learning together during this period of enormous change is "the seed of a future society where everyone is happy, not a dystopia," and they promise to continue observing and discussing from their respective perspectives.
"In an era where direction keeps changing, constantly updating is exhausting. But isn't that exactly the 'opportunity' and the 'path to survival'?"
Conclusion
In 2025, AI is going beyond being a mere tool to 'rewrite' organizations, talent models, the economy, and society itself. The essence ultimately comes down to how we can combine 'uniquely human will and insight' to envision a better future. Amid the evolving gaps and waves of change, it's time for all of us to adapt through rapid 'learning and unlearning' and create new opportunities.