Marketing and Sales Are the Future. Coding Jobs Will Disappear.

I’m a developer.

I’ve spent years learning how to code, build systems, fix bugs, deploy things, work with databases, APIs, cloud, all of it.

And I think coding as a job is absolutely doomed.

Not "changing."

Not "developers will become more productive."

Not "AI will be another tool."

I mean actually doomed.

If AGI happens, I see almost no reason companies will keep paying humans to write code.

Why would they?

Coding is basically the perfect thing to automate

Think about what developers actually do.

You get a requirement.

You understand it.

You break it down.

You choose an approach.

You write code.

You test it.

You fix bugs.

You deploy it.

You monitor it.

Then you change it again because product decided everything needs to work differently now.

None of this is some magical human-only activity.

It's mostly structured reasoning.

AGI is literally supposed to be good at reasoning.

So why would programming survive?

I don't think it will.

"But AI makes mistakes"

Yeah.

Today.

Humans make mistakes too.

A lot of them.

The difference is that an AI can potentially improve extremely fast, copy itself, work 24/7, remember everything, inspect massive codebases, test thousands of things in parallel, and never get tired.

Meanwhile I'm sitting there at 2 AM trying to figure out why some random value is undefined.

Who do you think wins long term?

"Senior engineers are safe"

I don't buy this either.

People say junior developers will disappear but seniors will survive because architecture is hard.

Okay.

Why would AGI be able to write code but suddenly become stupid when asked to design a system?

System design is reasoning.

Architecture is reasoning.

Debugging is reasoning.

Code review is reasoning.

Understanding tradeoffs is reasoning.

Reading requirements is reasoning.

If AI becomes genuinely better than humans at general reasoning, senior engineering is not some sacred protected zone.

It gets eaten too.

Staff engineer.

Principal engineer.

Architect.

Whatever title you want.

If the work is mostly done with your brain and a computer, eventually the computer probably wins.

The uncomfortable part: even being "really good" at coding may not matter

This is what I think developers don't want to hear.

You can spend the next five years becoming insanely good at React, Go, Kubernetes, system design, distributed systems, whatever.

And then some model might eventually be better than you at all of it.

Not a little better.

Embarrassingly better.

It might understand an entire codebase faster than you can understand one service.

It might find bugs before you even realize where to look.

It might generate ten architectures, simulate them, benchmark them, pick the best one and deploy it before you've finished your coffee.

At that point, your "10 years of experience" doesn't automatically mean much.

That's depressing.

But I think it's plausible.

Coding itself is becoming less valuable

This has already started.

Years ago, if you could build a full-stack application, that was impressive.

Now?

A random person with no engineering background can already get surprisingly far with AI tools.

And we're still early.

So imagine 10 years from now.

You tell an AI:

"Build me a CRM for logistics companies in India. It should support WhatsApp, invoices, lead scoring, user roles, analytics, payments, and mobile access."

And it just does it.

Not a shitty prototype.

A proper system.

Tested.

Secure.

Scalable.

Deployed.

Then what exactly is the value of manually knowing how to write the code?

It starts to feel like knowing how to assemble a calculator by hand.

Cool skill.

Not economically useful for most people.

We might be one of the last generations that learns coding for a job

I genuinely think about this.

Imagine explaining programming to a kid in 2045.

"So yeah, we used to manually type thousands of lines telling computers what to do."

They'd probably ask why.

And we'd say:

"Because the AI wasn't good enough yet."

That's kind of crazy.

They may look at our current coding jobs the same way we look at switchboard operators.

A real job.

Important at the time.

Then technology made it unnecessary.

Software will explode while software jobs collapse

This sounds contradictory but it isn't.

I think there will be more software than ever.

Way more.

Every company.

Every small business.

Every individual.

Everyone will have custom tools.

But those tools won't require armies of developers.

That's the part people miss.

More software does not automatically mean more software engineers.

If one AGI can do the work of 100 engineers, companies don't hire 100 engineers because they suddenly want more software.

They use the AGI.

Maybe they hire five people to direct it.

Then eventually maybe they need zero.

And companies will absolutely choose the cheaper option

People sometimes talk like businesses will preserve developer jobs out of some moral attachment to humans.

Come on.

If one company can replace 500 engineers with AI and cut costs by 90%, they'll do it.

Then their competitor has to do it too.

Then everyone does it.

That's how markets work.

If AI becomes cheaper, faster and better, human developers become an expensive inefficiency.

It sounds brutal because it is brutal.

So why do I think marketing and sales matter?

Because while building becomes easier, getting attention is still hard.

Let's say everyone can build a great SaaS in one day.

Cool.

Now there are 100,000 SaaS products.

Which one wins?

The one with distribution.

The one people trust.

The one people know.

The one that understands the customer.

The one that can tell a story.

The one that can sell.

This is why I'm increasingly interested in sales and marketing.

Building used to be the bottleneck.

I think it's going away.

Distribution becomes the bottleneck.

A great product nobody knows exists is worthless

Developers hate this.

We love to think:

"If the product is good, people will use it."

No.

People need to know it exists first.

There are probably thousands of amazing products right now with 12 users.

And there are mediocre products making millions because they were distributed well.

The future probably makes this even worse.

If AI makes high-quality products cheap, quality itself becomes less differentiating.

Everyone can have good software.

So now what?

Brand.

Trust.

Relationships.

Sales.

Marketing.

Timing.

Audience.

Distribution.

That stuff becomes insanely important.

But yes, AGI will probably eat sales too

I'm not going to pretend sales is permanently safe.

If AGI becomes truly superhuman, why wouldn't it be good at sales?

Why wouldn't it understand psychology?

Why wouldn't it negotiate?

Why wouldn't it personalize every pitch perfectly?

Why wouldn't it run 10 million marketing experiments simultaneously?

Eventually it probably does all of that too.

So this isn't:

"Developers should become salespeople and they'll be safe forever."

No.

If AGI becomes what people think AGI will become, there may be very few cognitive jobs that are safe forever.

Marketing and sales just look more useful during the transition.

The worst strategy is identifying yourself with one skill

This is probably the biggest thing I've changed my mind about.

I don't want to be "a coder."

That's too fragile.

I don't want my entire economic value tied to one activity computers are getting dramatically better at every year.

I'd rather be someone who knows how to solve problems.

Someone who can understand technology.

Understand people.

Sell.

Communicate.

Build businesses.

Make decisions.

Lead.

Adapt.

If coding matters, I'll code.

If AI does the coding, I'll use AI.

If something else replaces AI tools in twenty years, I'll learn that.

The output matters more than the tool.

This is why I'm pessimistic about learning to code as a career strategy

If someone told me today:

"I'm 15. I want to spend the next ten years becoming really good at manually writing software because I want a stable career until I'm 60."

I would be very nervous for them.

Learn coding because you want to understand computers?

Absolutely.

Learn it because you want to build things right now?

Yes.

Learn it because it teaches you how to think?

Sure.

But betting your entire 40-year career on humans being needed to manually produce code?

That feels insane to me.

The direction is too obvious.

I think programming salaries will get crushed before the jobs fully disappear

This might actually happen first.

You don't need AI to replace every developer immediately.

You just need one developer with AI to do the work that previously required five developers.

Now demand drops.

Then companies need fewer juniors.

Then fewer mid-level engineers.

Then fewer senior engineers.

Competition increases.

Salaries fall.

Teams shrink.

Expectations rise.

One engineer is expected to ship what an entire team used to ship.

Then AI gets better again.

Repeat.

You don't wake up one morning and every coding job is gone.

It gets squeezed slowly.

Honestly, that might be worse.

Junior developers are especially screwed

How do you become senior if nobody wants juniors?

That's the weird problem.

Companies already prefer experienced engineers.

AI reduces the amount of basic work juniors usually learn on.

So now companies can use AI for the easy tasks and hire a few experienced people for the harder ones.

Great for companies.

Terrible for someone trying to enter the industry.

Then eventually AI starts doing the harder work too.

That career ladder gets very weird very quickly.

And no, learning another framework isn't going to save you

This is something I have to remind myself too.

React isn't protection.

Java isn't protection.

Python isn't protection.

C++ isn't protection.

Cloud isn't protection.

Kubernetes isn't protection.

The AI can learn tools faster than we can.

So the answer can't just be:

"I'll learn more tools."

You need to become useful at a higher level.

I still code

This is the funny part.

After writing all of this, I'm still going to open VS Code tomorrow.

Because coding is valuable right now.

People still pay for software.

AI isn't AGI yet.

There is money to be made.

There are things to build.

There is a huge amount to learn.

I'm not going to quit a useful skill because I think it eventually disappears.

That would be like refusing to drive in 1990 because self-driving cars might exist one day.

But I'm also not going to lie to myself.

I don't think programming is a safe long-term identity.

My bet is adaptation

That's basically it.

I'm pessimistic about coding.

Very pessimistic.

I think AGI eventually destroys software engineering as a profession.

Not changes it.

Not "enhances" it.

Destroys it.

And eventually it probably comes for a huge number of other white-collar jobs too.

So I don't want to spend my life protecting one skill.

I want to be the kind of person who can keep moving when the world changes.

Right now I build with code.

Tomorrow I might build with AI.

Later I might build with something we don't even have a name for yet.

I don't care.

I just don't want to be standing there in ten years saying:

"But I spent my whole life learning how to code."

The machine won't care.