Why The Future Of Work Rewards Taste Over Output

​Elton Chan is the Co-Founder of Second Talent, a solution that connects global tech leaders with AI-native engineering talent across Asia.

For most of the industrial era, work meant output. You were valued for what you produced, how fast you produced it and how much you produced. That logic built the modern org chart.

But with generative AI making raw output nearly free, the model is breaking. A competent first draft of almost anything now appears in seconds, whether it's a memo, a model, a marketing plan or a block of code.

When the act of producing becomes abundant, its scarcity value falls. What remains scarce, and therefore valuable, is the human taste to decide whether the output is any good, whether it fits the strategy and whether it is the right thing to make at all.​

This is the real shift leaders need to understand. The future of work does not reward those who can produce the most, but those who can judge the best.

In the World Economic Forum's Future of Jobs Report 2025, analytical thinking ranks as the single most sought-after core skill, with seven in ten companies calling it essential in 2025.

As the report notes, organizations are facing growing complexity in decision-making and the need for critical problem-solving in a data-heavy world. In other words, the ability to generate is being solved by AI, but this increases the need for those who can analyze, weigh and decide based on what's generated.

The same report projects a churning labor market, with the WEF estimating a net 78 million new jobs by 2030 once creation and displacement are netted out, with enough roles based on human taste and judgment to more than replace those lost to AI.

The reason is uncomfortable but important to understand: AI is fluent and frequently wrong, and it almost never tells you when it is bluffing.

The clearest evidence comes from a domain where accuracy is non-negotiable. A Stanford-led study of purpose-built legal research tools found that even specialized, retrieval-grounded systems hallucinated on a meaningful share of queries, with error rates between 17% and 33%.

These are not toy chatbots. They are professional tools built specifically to be reliable, and they still produced confident, plausible, wrong answers often enough to matter.

That is the whole point: When output is cheap and unreliable, the bottleneck moves to verification. The person who can look at a polished, persuasive AI answer and say “this citation does not support that claim” is now worth more than the person who simply generated the answer. ​

There is a more subtle reason leaders should care.

Stanford's Institute for Human-Centered Artificial Intelligence's 2025 analysis of economic studies that track worker output found that AI boosts productivity and, in most cases, narrows the gap between lower- and higher-skilled workers, letting less experienced employees reach results that once required an expert.

But it carries a trap. If AI lets a junior employee produce senior-level output, the visible product no longer signals real capability. A great deliverable might reflect deep understanding, or it might reflect a good prompt and zero comprehension.

For leaders, this means you can no longer manage by inspecting output alone. You have to manage for judgment, which is harder to see and harder to fake.

First, change what you measure. If you still reward volume, speed and visible activity, you are optimizing for the one thing machines now do for free. Start measuring decision quality. Did this person catch the flaw? Did they ask the better question? Did they kill the bad idea early?

Second, train evaluators, not just producers. We spent decades teaching people to make things. The new core skill is critiquing things, spotting the confident error, knowing when a plausible answer is the wrong one. Build your organization's analytical and critical thinking muscles deliberately.

Third, protect the space for judgment. Taste needs time, context and the freedom to say no. An organization that floods its people with AI-generated output but gives them no room to scrutinize it has simply automated the production of plausible mistakes. The leader's job is to slow the right moments down.

The numbers tell a coherent story. Adoption is high. According to McKinsey's 2025 State of AI survey, 88% organizations report regularly using generative AI in at least one business function, up from 78% a year earlier.

However, even​ with most organizations now using generative AI, the skills employers prize most are analytical and critical thinking, precisely because the technology is powerful but unreliable. The gap between those two facts is where employee value now lives.

Output has been commoditized, but taste has not. The leaders who understand this will stop asking their people to produce more and start helping them decide better. That is the work that is left, and it is the work that matters.​​

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?