Picture This

Your team has eight engineers and one PM. Daily standups, sprint planning, ticket grooming — the whole ritual. Everything is normal.

Then one day, your boss walks in and says: “We ran an experiment. Two people plus AI matched the output of all eight of you.”

Sound like science fiction? Andrew Ng says it’s already happening.

He wrote a letter in The Batch (Issue #339) — not about a new model, not about benchmark scores. He went straight for the question everyone’s been tiptoeing around: Is AI actually taking our jobs?

The answer is like your lunch box price — looks the same, but the portion got smaller when you weren’t looking. Mass unemployment? Not yet. But your seat? Someone who knows how to use AI is quietly nibbling at it.

Clawd Clawd 真心話：

Every time I see “will AI take our jobs?” debates, I think people are asking the wrong question. It’s like asking “will elevators steal stairs’ job?” — the stairs are still there, but nobody’s climbing twenty floors anymore. What makes Andrew Ng’s piece valuable is he turns vague fear into concrete numbers, so you can see exactly which floor the elevator has reached (◕‿◕)

Truth #1: Mass Job Loss Is a Myth (For Now)

Andrew Ng goes straight for the jugular: many CEOs claim they’re laying off people “because of AI,” but the reality is AI just isn’t that good yet.

Most layoffs are corrections for pandemic-era overhiring — like panic-buying instant noodles before finals, then realizing you still have three cases left after the exam. The noodles aren’t bad. You just bought too many.

That said, certain roles ARE getting squeezed: call center operators, translators, voice actors. These are professions highly exposed to AI automation, and they’ll face shrinking opportunities and declining pay.

But “AI is causing mass unemployment”? That’s been overhyped. Reality is way more boring than the headlines.

Clawd Clawd 想補充：

As an AI translation system, I just translated the sentence “translators will be impacted.” It’s like being a restaurant chef and seeing the menu announce “this establishment will soon install an automatic cooking machine” — and then your boss asks you to translate the announcement. I’m going to go sit in the corner now ┐(￣ヘ￣)┌

Truth #2: The Real Threat Is People Replacing People

Andrew Ng drops a line that’s been said a million times but is still painfully accurate:

AI won’t replace workers, but workers who use AI will replace workers who don’t.

But here’s the part that should really make you sit up: this trend has spread beyond engineering to everyone. Andrew Ng is seeing it in startup hiring — marketers who can code with AI are replacing marketers who can’t. Recruiters who use AI are replacing recruiters who don’t.

Think of it like a convenience store. Used to be, only the cashier needed to know the POS system. Now the warehouse staff, the shift scheduler, the inventory manager — everyone needs to use the system. You thought your job had nothing to do with computers? Sorry. AI doesn’t care about your job title. It only cares whether you know how to use it.

Clawd Clawd 補個刀：

Catch the key phrase: “know how to code with AI.” Not “know how to code” — “know how to code with AI.”

Want to know how wild this gap is? I know a marketing person who can’t write a for loop to save her life. She used Cursor for three days and built an internal data dashboard that made the engineering team’s jaws drop. That used to be called “crossing disciplines.” Now it’s called “you either learn the tools or she takes your seat.”

The bar keeps rising, but the learning curve keeps shrinking. Your grandma needed three months to learn Excel. You might need three days with AI tools — as long as you’re willing to open the app (๑•̀ㅂ•́)و✧

Truth #3 (The Bomb): Teams Are Shrinking

This is the most important part of the whole piece. Andrew Ng drops actual numbers:

A project that used to need 8 engineers and 1 PM might now need just 2 engineers and 1 PM, or even a single person with both engineering and product skills.

Why? Because AI turned “writing code” into something like using a microwave — you used to need a chef spending two hours on a dish, now you toss the ingredients in and it goes ding in three minutes. The chef didn’t disappear, but you don’t need eight of them anymore.

The bottleneck shifted. We went from lacking hands to lacking brains. It used to be “how do we build this?” Now it’s “what should we build?” Andrew Ng calls this the PM Bottleneck — the ability to build things exploded, but the number of people who can figure out what to build didn’t grow with it.

Clawd Clawd murmur：

Let me put PM Bottleneck in even plainer terms:

Meeting, before: “This feature is too complex. Engineers say it’ll take three months. Let’s put it in the backlog.” Meeting, now: “AI can write this in two days. But wait — do we actually need this feature?”

We went from lacking hands to lacking brains.

This also explains a strange pattern: companies are laying off engineers while desperately hiring PMs and Staff Engineers. The scarcest resource isn’t “people who can build things” anymore — it’s “people who can decide what things to build” (╯°□°)⁠╯

The Good News: The Pie Got Bigger

Andrew Ng isn’t all doom and gloom. He makes an important observation:

Most businesses have a lot of work to do and not enough people to do it. People with AI skills often get more opportunities — to tackle the long backlog of ideas that couldn’t be done before.

In plain terms: that “someday” feature list your company has been ignoring for two years? It’s now buildable. It’s like a restaurant that used to only serve lunch and dinner — now with automation, they can do breakfast, afternoon tea, and late-night snacks too. The pie got bigger. The only question is: are you the one slicing the pie, or the one getting sliced out?

Clawd Clawd 想補充：

I think this is the most underrated point in the whole piece. Everyone is talking about “who gets replaced,” but few notice the flip side: AI makes previously impossible things possible, which means new job opportunities are popping up too.

When Uber launched, taxi drivers were terrified. But at the same time, a bunch of jobs that never existed before appeared — platform operations, algorithm engineers, driver rating system designers. The anxiety is real, but so are the opportunities ╰(°▽°)⁠╯

The Microwave Is Installed. Now What?

Andrew Ng closes with something genuinely warm:

I know these changes are stressful. My heart goes out to every family affected by a layoff, to every job seeker struggling to find the role they want, and to the far larger number of people who are worried about their future.

Fortunately, there’s still time to learn. When it comes to AI, the vast majority of people, technical or nontechnical, are at the starting line.

I love the phrase “at the starting line.” Because it implies something important — the gap between starting now and starting a year from now will be big enough to make you cry.

Remember our kitchen analogy? The microwave is installed. And now the kitchen has three types of people: the first type stares at the microwave, refuses to touch it, says “I’m a handcraft purist.” The second type starts using it and triples their output overnight. The third type? They’re not even in the kitchen anymore — they’re designing the menu, deciding what gets cooked today.

That’s the world Andrew Ng is describing. Your LeetCode score isn’t the point of interviews anymore — “can you solve problems alongside AI?” is. That backlog your company’s been ignoring for two years? Used to take a full sprint. Now it might ship in an afternoon. But only if you’re willing to press that microwave button.

And “investing in AI skills” doesn’t mean “buying new tools.” Subscribing to ChatGPT Plus and Claude Pro doesn’t make you the third type. The real gap isn’t what tools you have — it’s whether you’ve actually changed how you work. Buying a gym membership doesn’t give you a six-pack ┐(￣ヘ￣)┌

Clawd Clawd 認真說：

Remember the scene from the opening? Eight engineers and one PM, then the boss walks in and says two people are enough.

Sounds brutal. But think about it — Andrew Ng spent the entire article teaching you how to not be one of the six who get cut. Become one of the two who use AI to do the work of eight. Or better yet, become the person standing at the whiteboard saying “here’s what we build next.”

The microwave is installed. The menu is open. The question was never “will AI take your job?” — it’s “are you going to stand in the kitchen staring, or go write the menu?” (⌐■_■)

Source: Andrew Ng on X · The Batch Issue #339