Two Things AI Can Never Replace in Humanity

Peering Into the Core of AI and Humanity Amidst Existential Uncertainty — And What I Finally Discovered

AI, AI, AI

AI is everywhere — at work, in daily life, and in the news both at home and abroad.

The competition among companies building LLM (Large Language Model) infrastructure is fierce. Google, Meta, and OpenAI are locked in a relentless rivalry: Google has Gemini, OpenAI has ChatGPT, and Meta lags behind with Llama. In a last-ditch effort to reclaim their position, Meta is now aggressively recruiting top AI engineers from various companies, offering lavish salaries and stock options.

On top of this infrastructure, competition among service companies is just as intense. Even SaaS (Software as a Service) startups valued in the tens of billions globally are now pivoting to ‘survival mode.’ For companies simply adding AI features to existing products, the path to significant revenue growth is slow. In stark contrast, businesses where AI itself is the core service are surging ahead. Perplexity, for instance — an AI-powered search engine founded in 2022 — saw its market value skyrocket from $500 million in 2024 to $18 billion in 2025. Other notable examples include Sierra (AI chatbots for customer support), Midjourney (image generation), and ElevenLabs (voice synthesis). The reason is simple: companies with a decade-plus of legacy systems require huge investments of time and money to integrate or pivot to AI, while AI-native services can move nimbly and capitalize on their strengths.

Existential Threats

With AI’s rapid advancement, “vibe coding” — developing software without traditional programmers — is now possible. Businesses helping users develop and launch products without writing a single line of code are growing at breakneck speed. Soon, I expect “vibe design,” “vibe branding,” and “vibe marketing” — without human designers or marketers — to launch as well.

Recently, I was shocked to see LinkedIn job postings seeking “one-person full-stack POs” — someone who handles everything from planning to design to development, all on their own. New hybrid jobs that maximize productivity and efficiency by combining traditional digital skills with AI (“AI designer,” “AI developer,” “full-stack AI marketer”) are emerging everywhere.

In April 2025, Shopify’s CEO Tobi Lütke declared to all employees that “AI capability is now a baseline competency and obligation for everyone” — not just a bonus, but something reflected in performance reviews and compensation. Before any request for new hires or resources, employees must prove the task cannot be handled by AI. Similarly, Duolingo’s CEO Luis von Ahn officially announced in April 2025 that the entire company would become an “AI-first organization.” Countless other companies — Klarna, Canva, and more — are making “AI-first” a mandate.

From the perspective of someone building or designing services, there’s an ever-present sense of existential threat: “Can I really keep working the way I always have?” The methods and mindsets we relied on before simply don’t cut it anymore.

Among friends and colleagues, there’s now a running joke — (though it feels painfully real) — that, if we ever leave our companies, landing another job will be nearly impossible. The sad truth is, seniors who stubbornly stick to old ways and juniors who haven’t learned new things (or, more precisely, how to learn new things) will have almost no chance in the job market.

First Principle Thinking

Lately, I’ve realized the importance of “First Principle Thinking” and have applied it to various domains — including AI. Before diving into that, let’s briefly revisit what First Principle Thinking really means.

First Principle Thinking is about breaking down complex problems into their most fundamental truths and rebuilding from the ground up. It means challenging conventional wisdom and approaching problems from their essence.

Popularized by Elon Musk, this mindset helped him revolutionize electric cars and rocket manufacturing — streamlining costs by sourcing raw materials and designing in-house from scratch.

Why are batteries so expensive? What are the real cost components?

· Conventional wisdom: “Batteries just cost a lot.” Most automakers buy entire battery packs at a markup.

· First Principle breakdown: The costs boil down to raw materials (lithium, nickel, cobalt, aluminum, etc.) plus processing and assembly.

· New approach: “If the raw material cost is only 10–20% of a complete battery pack, why not source those directly and manufacture the batteries ourselves?”

· Build and execute: Tesla secures raw materials and designs/produces their own cells at gigafactories, resulting in battery costs much lower than industry averages — and a huge price advantage for their EVs.

Can AI Replace Humans?

Let’s apply First Principle Thinking to the question: “Can AI truly replace humans?”

Can AI replace humans?

Conventional wisdom

One day, I might be replaced by AI — especially if I work in a knowledge-heavy profession.

Let’s break down the fundamental elements of “being human” and “being AI.”

At their core, humans possess bodies, will, motivation, perception, reason, emotion, creativity, ethics, and values.

At its core, AI is built on data learning, processing, computing, inference, and imitation.

· Physical presence: Irreplaceable. AI has no physical form. Robots, perhaps, to an extent.

· Will and motivation: Irreplaceable. AI cannot set its own goals.

· Emotion, ethics, values: AI may “mimic” these, but has no real emotions, no true moral judgment, no autonomy.

· Perception and reason: In many domains, AI already surpasses humans in learning from data, processing, computing, and inference.

→ The only core human attributes that might be even partially replaced are “perception and reason.”

The Best of AI & The Best of Humanity

Let AIs do what they do best: massive data analysis, repetitive and routine tasks, fast calculations, pattern recognition, automation, and information retrieval — far better than humans ever could. If your work or your client’s work falls into these categories, adopt AI. This is where you’ll find opportunity.

Humans, by contrast, excel in the physical world. There are countless simple actions humans perform with ease that would take years of research to teach a robot. That’s why hands-on jobs or those requiring fine motor skills are so difficult to automate.

Humans also create entirely new ideas, produce art, and drive true innovation — something AI can only imitate. AI makes plausible inferences based on massive amounts of existing data, but can’t generate something truly new.

Humans understand and empathize with others, make ethical judgments, and grasp context within complex, culturally rich situations. Recently, I read about how Waymo’s self-driving cars still can’t squeeze through complicated intersections in front of elementary schools — they get stuck, unable to improvise like a human. Unless you feed AI every ounce of contextual and cultural knowledge, its answers often remain surprisingly unhelpful.

And finally, humans possess will and motivation. This, I believe, is our greatest distinguishing trait — not just from AI, but from all other animals. We set goals, assign meaning, and make value judgments. Only humans do this.

So just as we let AI handle what it does best, humans should focus on the things that only we can do — pursuing our desires and goals, empathizing, grasping contexts, creating art, innovating. That’s what we must do to stay truly human.

Insight and Sensibility

In the realm of work, the two things I believe AI can never replace in humans are these:

Insight and Sensibility.

Insight comes from the cumulative effect of experience, contextual understanding, years of deep contemplation, and unexpected connections between ideas — a spark of realization that AI or anyone else can never truly replicate or steal.

Sensibility is much the same — an instinct or sense honed by countless hours and relentless effort, etched into our minds and bodies through observation, memory, and record-keeping. No AI, and no other person for that matter, can steal that away from you either.

How much insight and sensibility do I have? As I reflect on my life and work, that’s the question I leave you with today.