<< Go back to Posts

Thoughts About Agents




From Completion to Agents

Discovery

In my current company, all developers have a GitHub Copilot licence.

I was very happy with chat completion; it helped me write code with fewer typos, reusing the patterns I have developed so far. I didn’t use the agent mode. I was unaware of it. For me, it was just a “copilot chat” to ask questions, nothing more. Maybe when I got the licence, the agent mode did not exist.

In November 2025, I went to the GitHub Universe Summit in Paris. There, we got some talks about “Agents will be your new teammates”, and some broken demos (thanks to the unreliable network of the room!).

At the end of this day, my conclusion was twofold:

  • “Agents seem to be the new thing.” GitHub is trying to sell its product, OK. But it seems to be the new direction.
  • “Still, agents are incomplete: you can design a coding agent and a reviewer agent, but you cannot let them work alone.”

In the same period, some developers “hit” the GitHub Copilot Premium request limit. It represented 1% of the developers, so not that much. So, I gave it a try, to test how “useful” it could be.

Playing with Agents

The first time I opened the Agent window, I was like “What can I ask? I know what I need to do, but it is minor edits here and there, so it is difficult to specify them.”

After some trials, I found out how to ask and what to ask. The “what” comes with practice. The “how” is easier to get; it is more a question like “how to comply with the tool limitations”:

  • Always provide the context: This could be by writing your specs in a Markdown file, or providing the class the agent needs to use in your script.
  • Be specific: About the libraries, the expectations, the dos and don’ts

The World is Changing

I read Hacker News several times a week. There are always new articles about “how Claude Code helps to do X”. Or articles about the new way of working. When reading these articles, I feel like I am a very late adopter. However, when looking around me, colleagues and friends, very few developers already use agents, within GitHub Copilot, Cursor, or Claude Code. They are still asking their questions to ChatGPT (the free version), nothing more. They haven’t tested agents yet, and so are not willing to pay for a Claude Code subscription over $100.

From a rough estimate, I would say that only 5% of the developers are actively using agents for coding.

Coding with agents is different:

  • We do not want coders anymore, i.e., people typing on their keyboard one letter at a time that makes sense to the compiler or the interpreter at the end of the day. Agents code better than you; they are faster, make fewer typos, and handle more exceptions.
  • We need software developers: people able to convert high-level requirements (from PO or others) into coding directions. More like an architect than a pure developer.
  • Software developers need to develop “management skills”. No one is born a manager. This means several things:
    • Learning how to ask, how to specify a task so that the outcome of the agent is predictable
    • Understanding what the agent’s capabilities are (where they are good or not, the level of abstraction they can handle alone)
    • Learning to delegate, i.e., even if they know they can do the job alone, they should learn when to keep it for themselves and when to delegate the job to agents

Learning how to Replace Ourselves

Computer science evolves toward more abstract concepts:

  • We got assembly
  • We got compilers and C
  • We got garbage collection, with no need to handle memory
  • We got high-level languages making code easy to read, easy to write
  • We got libraries for anything, making it easy to interact with Bluetooth, LLMs, without the need to understand all the complexity (we still need a basic understanding of the processes happening though)

Now, we got LLMs and agents. We need to figure out how to use them.

This year, I was pretty happy to learn how to play with CI/CD. Thanks to that, I saved a lot of time. I also learned to adapt my code so it fits in a Docker container to be built by Jenkins. The tool did not adapt to me. But in the end, I gained a level of automation that I will add to my home projects.

Software Developper will not Disappear

  • You can vibe code a tool easily with no coding knowledge at all. Yet, you still need some knowledge to run the tool, integrate it, do additional configuration. We need people to make the bridge between code, networks, databases, and more.
  • When the codebase increases, you can still ask an agent to do the job. However, this will be more costly, as the agent will need to read your files and figure out what to do. Getting someone able to read the code, able to organize it—this person can point out to the agent which files it needs to look at and edit, making it more efficient.
  • Without human guidance, agents will take decisions that will make the final product better or not, more efficient or not. At some point, we need someone able to manage the ship, to guide it so it always stays safe from the rocks.

Yet, their number might decrease a bit.

This might be related to “bullshit jobs” at some point. Programming jobs with no extra value.


>> You can subscribe to my mailing list here for a monthly update. <<