What LinkedIn Offshore Development Pitches Tell Me About How the Industry Has Changed

The Declining Value of Adding More People to Write Code

I still occasionally receive sales messages on LinkedIn from overseas development companies.

“We can provide highly skilled engineers at a low cost.”

“We can reduce your development costs by X percent.”

There was a time when pitches like these made perfect sense. From the 2000s through the 2010s, offshore development was often a highly rational choice. Hiring 100 people in a country with lower labor costs was cheaper than hiring 100 people in Japan. It was only natural that many companies outsourced development to China, Vietnam, and India.

But whenever I see one of these messages now, I wonder whether the sender is still selling to a world that disappeared a few years ago.

I am not arguing against offshore development. There are many excellent engineers around the world.

What has changed is not the people, but the nature of the work.

In the past, adding more people generally meant producing more code. That made it worthwhile to assemble a large team of people who could write it cheaply.

Today, the situation is different. Whether you use Claude Code, Codex, Cursor, or another tool, once the design is reasonably clear, AI can generate a substantial amount of code in a short time.

The output still needs to be reviewed, and human judgment remains essential. Even so, we no longer live in a simple world where adding more developers automatically makes development faster.

AI Makes Communication Costs More Visible

Working with AI reveals an interesting imbalance. The time required to write code has fallen dramatically, but the work that comes before coding has barely become any shorter.

What does the client actually want to build?

How should this exception be handled?

Is this specification really the right one?

The heart of development now lies in this process of trial and error: repeatedly discussing the problem with AI, refining the design, and sometimes deciding that another approach would be better after all.

Adding an offshore team to that process can easily create a game of telephone.

• I talk with the client.
• I write the specifications.
• I send them to the overseas team.
• The team interprets them.
• A programmer implements them.
• The result comes back.
• I review it, and the conversation begins with, “That isn’t quite what I meant.”

This problem is not new. But now that AI has made implementation so much faster, the waiting and communication overhead stand out more than ever.

In many cases, I can finish sooner by working directly with AI and refining the design and implementation as I go.

Why “Offshore Teams Can Use AI Too” Misses the Point

A common response is that overseas development companies can simply use AI as well.

That is half right and half wrong.

AI reduces the time spent typing code. It does not reduce the need for domain knowledge. It does not replace design experience. It does not eliminate the work of building a shared understanding with the client.

AI can multiply the capabilities of a good designer, but it does not understand the design on that person’s behalf.

We often hear that AI will replace engineers. That is not what I see. If anything, AI seems to be widening the gap between people who can design systems and those who cannot.

Design Ability Matters More Than Cheap Labor

I do not expect offshore development to disappear. There will still be demand for operations and maintenance, large-scale testing, and services developed locally for overseas markets.

However, I expect fewer projects to be sent abroad solely because labor is cheaper than it is in Japan. AI has multiplied the productivity of each individual.

If three people can now complete work that once required ten, the question of which country should supply the missing headcount becomes far less important.

What I Now Want to Hear in a Sales Pitch

That is why offshore development pitches on LinkedIn now leave me with a strange feeling.

“We can develop it for less” used to be a powerful selling point. But that is no longer what I want to hear.

I want to ask:

• What development process have you built around AI?
• How do you approach design?
• In the age of AI, how do you close the distance between your team and the client?

Those questions are far more interesting to me.

If an offshore company intends to handle only the coding, I am unlikely to outsource the work to it. Seeing companies continue to sell little more than additional headcount feels slightly melancholy.

When steam locomotives first appeared, perhaps coach drivers felt something similar as they proudly told prospective customers, “Our carriage runs faster than it did yesterday.”