Three Things AI Still Can’t Do

Technology moves fast. Humans move differently.

AI keeps getting better. It writes, summarizes, searches, and builds in seconds. So the question keeps coming back. What can humans still do?

That question came up in a talk about work and AI. Bob Sternfels, the global managing partner of McKinsey, shared a clear answer.

At CES 2026, Sternfels talked about how AI is already saving his firm millions of work hours. Searching. Summarizing. Building charts. Machines are fast. No argument there.

But he was clear about one thing. Even the best AI models still fall short in three human areas.

First is aspiration. AI can suggest goals. Humans decide what is worth chasing. Aspiration comes from desire, limits, fear, and hope. AI does not want anything.

You see this in simple choices. A student chooses to become a teacher even if another job pays more. AI can list higher-paying careers. Only a human can decide what kind of life feels meaningful. Aspiration is choosing direction, not just results.

Second is judgment. AI calculates. Humans choose. Judgment appears when there is no clean answer.

A school head decides not to punish a teacher who made a mistake, but to guide them instead. The policy says one thing. The situation says another. AI follows rules. Humans decide when mercy, fairness, or patience matters more.

Even small moments count. Choosing not to repost a viral story because it may hurt someone. No algorithm rewards that. That is judgment.

Third is creativity. Not remix creativity. Real creativity.

A writer starts a story with no clear plan, just a feeling. A musician breaks structure and risks sounding wrong at first. AI works from patterns that already exist. Humans can step outside patterns and try something new.

Sternfels was not saying AI is useless. He was saying it changes the game. It removes busy work. It forces people to lean harder into what makes them human.

Machines can assist. They can speed things up. But aspiration, judgment, and real creativity still come from people.

Even with Predictive Quantum Research, machines can only project what might happen. They cannot decide what should matter, what choice is right, or when to create something new.

That part remains human.

PQR (Predictive Quantum Research) • Darem Placer
Generation Alpha Bets includes PQR. Soon on Bandcamp.

⌨ ᴛʸᵖⁱⁿᵍ ᴏᵘᵗ ᵒᶠ ᵗʰᵉ ʙˡᵘᵉ ᵈᵃʳᵉᵐ ᵐᵘˢⁱᶜ ᵇˡᵒᵍ

AI vs AI: Before You Sign Anything

Contracts today are often written with AI. Reading them without help may no longer be enough before you sign.

For those who’d rather listen.

Almost everything today comes with a contract, even the simplest things. A small service, a quick signup, a basic agreement. Then suddenly you are handed a document that is more complicated than the thing you are agreeing to.

That is not accidental. Many modern contracts are no longer written purely by humans. They are assisted by AI, built from templates, optimized for risk, and filled with language designed to survive future changes. Polite tone. Clean structure. Harmless-looking clauses that quietly cover a lot of ground.

So if the contract was likely drafted with AI tools, why would you read it without help?

Depending on AI in this situation is not laziness. It is balance.

This is not about letting AI decide for you. It is about using AI as a lens. A pattern spotter. A way to test language that was designed to be flexible, expandable, and protective of the other side.

A simple way to do this is to scan or photograph the contract, upload it to an AI you trust, and ask direct questions. What is the worst-case scenario for me here? Which clauses quietly favor the other side? What parts could be used differently if technology changes? You are not asking AI to decide for you. You are using it to spot risks before you sign.

Modern contracts are no longer built around specific situations. They are built around concepts. Access. Use. Distribution. Derivatives. Training. Reuse. These words do not expire. They adapt.

The real danger is not AI. The danger is speed reading. Contracts today are written slowly and strategically, but signed quickly by people who assume the document matches the simplicity of the product.

A simple rule still applies. If you cannot explain a clause back to yourself in plain language, you did not understand it. And if you did not understand it, signing becomes a gamble.

The side asking you to sign is already inside an AI-built safe zone. Using AI before you sign is not about beating them. It is about finding where you stand, what you gain, and where the balance actually is.

This is why AI vs AI makes sense in this era. They wrote the maze with tools. You bring your own tools to read it.

Depending on AI here is not surrender. It is awareness. The signature is still yours, but now your eyes are open.

⌨ ᴛʸᵖⁱⁿᵍ ᴏᵘᵗ ᵒᶠ ᵗʰᵉ ʙˡᵘᵉ ᵈᵃʳᵉᵐ ᵐᵘˢⁱᶜ ᵇˡᵒᵍ