<?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 Marcelo Lobato Haro on Medium]]></title>
        <description><![CDATA[Stories by Marcelo Lobato Haro on Medium]]></description>
        <link>https://medium.com/@marcelo.lobato.haro_688?source=rss-2f5c200a0e21------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*wgbdoWxbeP6JXaiXD9h-pA.jpeg</url>
            <title>Stories by Marcelo Lobato Haro on Medium</title>
            <link>https://medium.com/@marcelo.lobato.haro_688?source=rss-2f5c200a0e21------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 05:07:46 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@marcelo.lobato.haro_688/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[Claude has arrived]]></title>
            <link>https://medium.com/@marcelo.lobato.haro_688/claude-has-arrived-dd9c33c4cae0?source=rss-2f5c200a0e21------2</link>
            <guid isPermaLink="false">https://medium.com/p/dd9c33c4cae0</guid>
            <category><![CDATA[ai-tools]]></category>
            <category><![CDATA[claude]]></category>
            <dc:creator><![CDATA[Marcelo Lobato Haro]]></dc:creator>
            <pubDate>Sun, 21 Jul 2024 23:57:09 GMT</pubDate>
            <atom:updated>2024-07-21T23:57:09.127Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/910/1*asc6Er-LvZ6Q3pyRJbqyOQ.jpeg" /></figure><p>Today, most of us developers are already using generative AI tools in our projects (I did not use it to write this article). We do it to get help in all areas of technology such as: writing specific and understandable code, optimizing SQL statements, creating containers, etc. Of course, AI tools display the answer according to the prompt we provide; therefore, our prompt should be as descriptive as possible. In this short article I will explain why I consider Claude to be one of the best tools that has appeared.</p><p><strong>A not so complex scenario, I think</strong><br>The scenario is the following: I wanted to replicate a server for development with a Docker container based on RHEL Linux, installed with Python 3.11 and Oracle 19c with full client (not the instant client because I needed the wallet manager functionality present only in the full client); the installation was to be done via console because on a physical server or with GUI interface the Oracle full client installer presents no inconvenience, but in a container you must use commands for a silent installation.<br>I read and reviewed my prompt and thought it was specific enough, so I asked the question to four tools: the well-known ChatGPT, Google’s Gemini, Microsoft’s Copilot and the new Claude.</p><p><strong>The results</strong><br>I’ll start with Gemini, the answer it gave me was quite general and pretty much described the steps I provided at the prompt. Secondly, I used Copilot, the response was not as expected because it generated a dockerfile with Oracle instant client and when interacting in the conversation requesting more information to install the full client it gave me the same initial alternative. Third, I used ChatGPT and it gave me a good answer but when I applied it I got errors which I used as new entries in the chat but finally I could not solve them with the new answers; and finally as I had heard about Claude’s release I used it and this tool really hit the nail on the head and gave me the solution.<br>It is worth mentioning that I interacted with all four tools and provided feedback on the warnings and error messages I got in the process with the instructions, and all except Claude ended up falling into a loop at the end. In contrast, Claude interacted with me in a polite way and regenerated his response with new configurations; this interaction was really satisfying and surprising to me, it gives me another approach to what an AI can perform.<br>This was the first time I worked with Claude and obviously it will not be the last, if you decide to try Claude I left you the link:<br><a href="https://claude.ai/login">https://claude.ai/login</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=dd9c33c4cae0" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>