Table of Contents
AI that can code sounds like a dream come true: “Ask what you need, and the AI tool will write the code for you.“
Engineers’ first reactions were something like, “This is exactly what I was looking for!” — and the fear in the back of their heads was, “This is it. We are out of jobs.“
To me, the first time I tried GitHub Copilot was a bad experience… too many retries/reasking the same questions. That was last year. This year, 2025, is way better.
I had the same experience with Cursor, but GitHub Copilot has been better for me so far.
I remember thinking, “I don’t need to code anymore. I will just ask the AI to write the code, and I can just review it,” and that has been my experience so far.
Asking the right questions is not the same as coding
This is where things get interesting… before trying the first AI-powered code assistant (or whatever you want to call it), I was prompting and building tools with OpenAI models for almost two years. That’s two years of experience learning how to ask the right questions!
Does your in-house team have the same experience? Most likely not. So you are asking a coder to become a questioner. As you might have already guessed, one thing is to know how to code, another completely different thing is to understand how to formulate a question.
If you are a manager, you most likely know that asking for things is one of your hardest-earned skills. A skill that is very different from coding.
Should Managers code instead of Engineers?
Well, wouldn’t that be amazing? Fire your entire development team — just ask the AI to build whatever you were asking your former engineers to build.
Isn’t that the direction we are heading in anyway? Just take the leap and get there before competitors.
There is a catch: any competitor with well-trained prompting engineers, former Senior Engineers, will beat you by a factor of 10x or more.
AI is dragging your In-House Team
Now you know what I mean by this. Your team is most likely spending more time debugging now than they used to when they were coding. After all, they are living their worst nightmare: fixing other people’s code, in this case, machine-generated code!
Ask any engineer: the best possible experience is to build something from scratch. The worst possible experience is fixing someone else’s code.
That’s exactly what you are putting them through right now: throwing AI-supposedly assistance tools at them, and they are in an infinite loop of asking > reviewing > debugging > fixing > asking.
It used to be simpler before: writing > debugging > fixing.
My Prediction: You will lose patience with your In-House Team
Startup leaders are freezing hiring because they believe the cost of having local talent in an office will be compensated by the productivity gain from using AI tools.
They are using AI as an excuse to revive the old dream of “if we are together in the same room, we are more productive“.
The Reality is Different
In-house engineers were traditionally slower than remote engineers because they needed to follow the company’s rules (aka office rules.) That means whenever they wanted to learn something new, they needed to do it in their own time.
Remote teams were usually composed of Senior Engineers who wanted more freedom and were usually assigned to non-core projects, where they could make decisions, among them using new technologies.
As with any new technology, whoever has the most ramp-up time will usually leverage the technology better. Guess who has the upper hand on using AI tools in today’s market?
Do you want to stay ahead?
If, supposedly, remote engineers have more knowledge about how to use AI tools to increase productivity, what would happen?
I don’t think managers will back down on their decision to freeze hiring and focus on local talent. They are not wrong that AI will make them more productive.
The question is how? The how will be provided by remote engineers, who, once more, will help startups, but this time, it will not be so much about lending an extra coding hand but teaching how to leverage AI to increase speed.
Ready to hire your next AI-Enhanced Remote Engineer?
Through my company, InTheValley, we use AI the right way every day for our startup clients.
Check out our Remote Staff Augmentation service.
- AI isn’t helping your in-house team… it’s paralyzing them - 03/28/25
- Engineers Refuse Office Comeback - 02/03/25
- Self-Discipline - 01/27/25