<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Dimitrios Kateros on Medium]]></title>
        <description><![CDATA[Stories by Dimitrios Kateros on Medium]]></description>
        <link>https://medium.com/@dkateros?source=rss-92130c8ae4d6------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*K8oxDaKPFMWmny83CbDGjA.jpeg</url>
            <title>Stories by Dimitrios Kateros on Medium</title>
            <link>https://medium.com/@dkateros?source=rss-92130c8ae4d6------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 08:53:59 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@dkateros/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Navigating the agentic AI developments as software practicioners]]></title>
            <link>https://medium.com/@dkateros/navigating-the-agentic-ai-developments-as-software-practicioners-f1a4de7f69ad?source=rss-92130c8ae4d6------2</link>
            <guid isPermaLink="false">https://medium.com/p/f1a4de7f69ad</guid>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[agentic-ai]]></category>
            <category><![CDATA[software-development]]></category>
            <dc:creator><![CDATA[Dimitrios Kateros]]></dc:creator>
            <pubDate>Sun, 03 Aug 2025 17:36:26 GMT</pubDate>
            <atom:updated>2025-08-15T20:36:58.688Z</atom:updated>
            <content:encoded><![CDATA[<p>During the summer break, I found myself taking time to catch up with the AI sphere. I am usually more reserved in spending time with a new technology and tend to do it later on the hype cycle. As far as I can tell, AI agents are all the rage now, which is both expected and welcome. I suspect that the first killer application, chatGPT, was more than a mere model since day one. Seems like we are already shifting from the phase of high expectations and prompt magic, to figuring out how to use the technology in a more constructive manner.</p><p>As a software practicioner, I am really interested in the effect of this breakthrough to my field. Both in terms of creating and maintaining software, as well as in terms of how the practice is affected and what a career in software will look like.</p><p>We are already getting assistants that can help to prototype and write a lot of the boilerplate. Integrated in our IDEs nonetheless! For the time being, I find myself reluctant to check in the generated code as is, but I am more than happy to consciously copy pasting stuff for my pet projects, especially when using libraries and frameworks I am not that familiar with. The learning curve is definitely shortened, especially with regards to fighting language syntax or configuration details. The benefit is arguably higher than what stackoverflow did some 15 years ago.</p><p>Down the road, we might be getting infrastructure that can “do stuff” with as little as specifying goals and tools. This can effectively reduce the amount of software written, especially in the experimentation and validation space. Assuming this approach turns out to be robust and cost effective, we might even see enterprise grade systems being developed, at least in part, this way.</p><p>What I really like about the direction of the agentic AI development is how it builds on classic software engineering principles and patterns. Our experience from OOP, DDD, workdflow engines, distributed systems is aldready apparent in agentic AI frameworks. At the same time, I am looking forward to witnessing how we will integrate essential practices such as versioning, testing, profiling, monitoring in this class of systems.</p><p>When the dust settles, I do not think we will write and maintain less software or that we will need less engineers. Through my experience so far, I have come to a simpler realization. Things tend to get as fast paced and complex as we can handle them to be. Every once in a while we get to transform increased understanding to better practices, and it takes time for them to get widely understood and adopted. And then, the extra capacity steming from the improved practices will eventually be spent towards getting things more fast paced and more complex.</p><p>For the younger software practitioners out there:</p><ul><li>Write some of the boilerplate. Do some of the trial and error. We still need to understand and create crystallized knowledge before delegating to the AI assistants. This used to be facilitated organically in the workplace. Senior engineers would actively take time to grow the next generations. By example, by figuring out which tasks will help them grow, by reviewing and explaining tradeoffs.</li><li>Do the work. Design, develop and support systems. Access to structured knowledge has been democratized way beyond the apprenticeship paradigm. Still, the best way for the finding motivation and for the learning experience to matter, is to get in the uncomfortable position of promising something, doing the work, getting user feedback and being responsible for operations.</li><li>Social skills and connecting to people will continue to be as career defining as tech skills.</li></ul><p>Lastly, some info in terms of existing agentic AI frameworks that grabbed my attention while I was having a look around:</p><ul><li>Langchain (Python, there is also a Java port, langchain4j): <a href="https://github.com/langchain-ai">https://github.com/langchain-ai</a></li><li>Agent Development Kit, by Google (Python, Java): <a href="https://google.github.io/adk-docs/">https://google.github.io/adk-docs/</a></li><li>Embabel, by <a href="https://medium.com/u/de5947672915">Rod Johnson</a> (Kotlin, Java) <a href="https://github.com/embabel">https://github.com/embabel</a></li></ul><p>Enjoy the rest of the summer, with or without agents :-)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HLxypl1oD2vzA3X78pL1lA.png" /><figcaption>The image is AI generated and the agents look sus :D</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f1a4de7f69ad" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>