The AI Era’s Great Compression: Redesigning Organizations, Incentives, and Culture

At the 2025 Luo Han Hall – Peking University National Development Institute Digital Economy Conference, Professor Li Jin of Hong Kong University introduced the “Great Compression Effect”, a framework describing how AI reshapes modern enterprises across three core dimensions: skills, motivation, and collaboration.

Skill Compression: AI Turns “Low‑Performers” into “Average‑Performers”

Li argues that AI systematically narrows the ability gap between novices and experts by scaling tacit knowledge. AI learns from massive data to capture experts’ subtle communication tricks, efficient problem‑solving approaches, and complex business processes, then distributes them as reproducible, scalable models, acting like an tireless mentor.

He cites the example of Yum! China (formerly KFC China). Previously, training a competent store manager required three to four years; with AI‑driven demand forecasting and other analyses, the training cycle shrank to 18 months, allowing more employees to handle roles that once demanded composite talent.

Research from Erik Brynjolfsson et al., Generative AI at Work , aligns with this observation. In experiments with over 5,000 customer‑service agents, AI assistants raised average efficiency by 14% (more issues resolved per hour). The weakest agents saw a 35% productivity jump, while top performers showed little change, illustrating that AI helps “low‑performers” quickly catch up to “average‑performers”. Additionally, the learning curve flattened: new hires reached the performance level that previously took 6–10 months in just 2–3 months with AI assistance.

The essence of skill compression is converting the scarce asset of experience into an on‑demand resource, destabilizing traditional hierarchies that valued seniority and accumulated expertise.

Motivation Compression: The “80‑Point Curse”

Beyond skills, Li highlights a deeper impact on motivation. In a study of artists, he found that after introducing generative AI, most creators stopped working after 60 minutes, well before exhausting their allotted time. About 30% of illustrators produced lower‑quality work because few were willing to invest additional hours to improve from an 80‑point to an 85‑point outcome.

Li labels this phenomenon the “80‑point curse”, explaining why AI‑era outputs often reach “good” but rarely achieve “excellent”. When achieving 80 % quality becomes effortless, the marginal return of striving for higher excellence diminishes, prompting many to adopt a rational “slack‑off” stance. Leaders thus face a stark question: “Do we need many 80‑point results or a few 95‑point results?”

Collaboration Compression: From “Three Stooges” to “One Zhuge Liang”

The combined skill and motivation compression triggers a third transformation—compressed collaboration mechanisms—most visibly the erosion of middle management. As AI handles routine tasks and offers strong decision support, the traditional role of middle managers as information relayers and process overseers weakens.

Li again references Yum! China’s “MEGA” super‑store‑manager program. A single high‑performing manager can now oversee 2.5 to 5 stores, compared with the previous one‑store‑per‑manager model. The company’s leadership remarked, “We used to prefer ‘three stinky shoemakers’; now we prefer ‘one Zhuge Liang.’” This shift mirrors Scott Belsky’s “Collapsing the Talent Stack” theory, which posits that AI enables empowered individuals to accomplish tasks that once required multiple experts, allowing teams to become smaller and more agile.

Former Alibaba chief strategist Zeng Ming also advocated for “organizational sharpness”, emphasizing that future organizations will rely on compact, high‑impact “special‑forces” teams capable of rapidly tackling open‑ended problems.

Future organizational sharpness, or impact, lies in the ability to swiftly solve open problems. Continuous creation demands close human‑machine collaboration, and the emerging optimal form may be tightly‑coupled special‑forces teams. - Zeng Ming, former Alibaba chief strategist

Overall, the Great Compression suggests a move toward flatter, more agile, and elite‑oriented structures.

Where Will Future Experts Come From?

Skill compression automates many entry‑level, repetitive tasks, undermining traditional apprenticeship pathways. Tasks such as junior lawyers reviewing contracts or junior analysts cleaning data are increasingly performed by AI, eroding the low‑risk environments where novices learn through mistakes.

In software engineering, tools like Cursor, Windsurf, Claude Code, and CodeX are advancing to the point where freshly graduated programmers may find few entry opportunities, raising the question: if young professionals lose the chance to err and learn, how will future experts emerge?

Leadership’s New Challenges

1. Redefine Talent: From Skills to Traits

As AI assumes more analytical work, Li argues that personality traits—e.g., extroversion vs. introversion—may outweigh raw technical ability. Qualities such as curiosity, proactivity, empathy, and perseverance become highly valuable because AI cannot replicate them.

2. Re‑engineer Incentives: Guard Against the “80‑Point Curse”

Traditional output‑based KPIs risk backfiring under motivation compression. Leaders should design reward systems that recognize boundary‑pushing, excellence‑seeking behavior rather than merely task completion, thereby encouraging outcomes that exceed the easy 80‑point plateau.

3. Strengthen Organizational Culture: Psychological Safety and Belonging

When AI makes performance highly visible and quantifiable, cultural factors can be sidelined. Li stresses that cultures fostering psychological safety and belonging motivate employees to embrace AI rather than resist it, making collaboration, trust, and innovation essential buffers against the negative effects of compression.

Conclusion: Finding New Value Peaks in a Compressed World

The Great Compression starts with skill compression—AI accelerates novice growth. It then triggers motivation compression—easy 80‑point performance dampens the drive for excellence. Together they compress collaboration mechanisms, foreshadowing the decline of traditional middle management and the rise of “super‑individuals.”

Organizations must confront two pivotal questions: how to motivate and develop talent capable of achieving 100‑point performance, and how to construct new growth pathways for young professionals as classic career ladders crumble.

Leaders who deeply understand and proactively manage the Great Compression will redesign structures, incentives, and culture to locate new high‑value plateaus amid the compression.