John Crickett

This account is verified. Learn more

@johncrickett

Helping software engineers become better software engineers by building projects. With or without AI.

South West, England codingchallenges.fyi Joined April 2010

1.3K Following 15K Followers 8.2K Posts

Latest posts 5

John Crickett @johncrickett 10h

My prediction for software engineering over the next 3 to 5 years: the industry will come to regret not hiring junior developers.

Organisations are treating junior engineers as a cost they can eliminate with AI, ignoring the fact that developing engineers is an investment in talent we’ll need tomorrow.

In the next 3 to 5 years, the losers will be companies that eliminate junior positions to cut costs. They'll find themselves competing for a shrinking, and therefore, increasingly expensive, pool of experienced engineers.

The winners will be companies that continue investing in junior engineers, and who continue to provide them opportunities to grow, with clear retention strategies.

And for a while, the engineering consultancies that invest in training junior engineers will benefit massively. As those who struggle to staff their projects pay the consultancies a premium for experienced talent.

John Crickett @johncrickett 12h

Steve’s coding agent had fixed the bug. All the tests passed. Job done.

Except Steve wasn't sure.

He wanted to check. So he opened the code in his IDE, followed a few dependencies and then stepped through it in the debugger.

There it was. The value of that variable seemed wrong.

But Steve didn’t know this code. The agent had just written it.

He wanted the agent to investigate, but the agent couldn't see what Steve was seeing. It couldn't access the debugger.

Instead, Steve would have to explain the problem, describing the state to the agent via chat.

He wished the agent could use the debugger itself. It would be easier if it could set breakpoints, inspect variables, and follow the call stack.

And if it found something interesting, Steve could pick up the same debugging session in the IDE.

Neither would have to explain what the IDE showed them.

This post is the third in a new series where I'm using microfiction, stories of approximately 150 words, to bring software engineering ideas, challenges, and lessons to life.

Why? Because I've challenged myself to become a better storyteller.

This story was inspired by my earlier post: x.com/johncrickett/s…

John Crickett Verified account @johncrickett 2d

If AI agents write most, if not all of our code, do we still need an IDE?

Maybe not. But why would we want to give a powerful tool like an IDE even if we only use it 10% of the time?

As agents write more code, I think our ability to understand, review, debug and verify that code becomes even more important. After all, we're still accountable for the software we ship.

And the powerful tools IDEs already provide, from language servers and code navigation to debuggers and test runners, could be used by coding agents to make them more effective.

We’re already seeing this happen with the JetBrains IDEs. Their IDEs allow us to work with multiple agents, provide the agents access to the tooling in the IDE and provide a great environment for us to read and review the code when we want to.

So, I think the future of IDEs is as a shared development environment, where software engineers and agents work together, use the same tools and context, and where we can choose between different agents and models.

I don't want to choose between an IDE and a coding agent. I want both.

In this video, I explain why I think that's a much more interesting future for software development.
What do you think?

This post was sponsored by @jetbrains. I highly recommend checking out their IDEs: jb.gg/k1zcf3 and the new JetBrains Air for managing agents and verifying their work: jb.gg/z4axr0

View quoted post on X

John Crickett @johncrickett 13h

Better is too subjective and context dependant to make this defensible.

Sometime better is simpler code.
Sometimes better is code delivered faster.
Sometimes better is more performant code.
Sometimes better is more maintainable code.
Sometimes better is code that can be formally verified.

Many times it's a tradeoff.

Glitchbyte Verified account @0xglitchbyte 1d

If ai writes better code than you, you have skill issues.

Its that simple.
View quoted post on X

John Crickett @johncrickett 2d

If AI agents write most, if not all of our code, do we still need an IDE?

Maybe not. But why would we want to give a powerful tool like an IDE even if we only use it 10% of the time?

As agents write more code, I think our ability to understand, review, debug and verify that code becomes even more important. After all, we're still accountable for the software we ship.

And the powerful tools IDEs already provide, from language servers and code navigation to debuggers and test runners, could be used by coding agents to make them more effective.

We’re already seeing this happen with the JetBrains IDEs. Their IDEs allow us to work with multiple agents, provide the agents access to the tooling in the IDE and provide a great environment for us to read and review the code when we want to.

So, I think the future of IDEs is as a shared development environment, where software engineers and agents work together, use the same tools and context, and where we can choose between different agents and models.

I don't want to choose between an IDE and a coding agent. I want both.

In this video, I explain why I think that's a much more interesting future for software development.
What do you think?

This post was sponsored by @jetbrains. I highly recommend checking out their IDEs: jb.gg/k1zcf3 and the new JetBrains Air for managing agents and verifying their work: jb.gg/z4axr0

7:01

John Crickett @johncrickett 4d

If Anthropic is merging more than 30,000 code changes per day, why aren't the GitHub issues being assigned and addressed?

This is a great question on r/ClaudeAI.

I think there's a bigger question:

Why aren't Anthropic triaging this with a software factor?

There's a risk that the issues could be used for prompt injection, but you'd have thought if any organisation has the expertise to build a software factory that can review the issues, filter out the risky ones and assign the ones that matter to either another agent or a human for further review fix, it would be Anthropic.

Building such a software factory and letting us see it work would make a better marketing case study than claiming to be merging 30,000 PRs per day!