<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Carlos Casalicchio</title>
    <description>The latest articles on DEV Community by Carlos Casalicchio (@ccasalicchio).</description>
    <link>https://dev.to/ccasalicchio</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F1253522%2F1a704fe8-0d12-41dd-8304-f8743b57c544.png</url>
      <title>DEV Community: Carlos Casalicchio</title>
      <link>https://dev.to/ccasalicchio</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/ccasalicchio"/>
    <language>en</language>
    <item>
      <title>We just published a detailed investigation into whether AI agent skills can push</title>
      <dc:creator>Carlos Casalicchio</dc:creator>
      <pubDate>Sun, 19 Jul 2026 16:41:54 +0000</pubDate>
      <link>https://dev.to/ccasalicchio/we-just-published-a-detailed-investigation-into-whether-ai-agent-skills-can-push-1de3</link>
      <guid>https://dev.to/ccasalicchio/we-just-published-a-detailed-investigation-into-whether-ai-agent-skills-can-push-1de3</guid>
      <description>&lt;p&gt;We just published a detailed investigation into whether AI agent skills can push front-end from "competent" to "genuinely beautiful" — with blind testing to prove it.&lt;/p&gt;

&lt;p&gt;Key findings:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Built 8 skills (7 knowledge + 1 process)&lt;/li&gt;
&lt;li&gt;First evaluation: tie on correctness (we were measuring the wrong thing)&lt;/li&gt;
&lt;li&gt;Blind test: skilled agent won both tasks at high confidence&lt;/li&gt;
&lt;li&gt;The process skill (render-critique-refine loop) was the breakthrough&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The article covers the full methodology, results, and what we changed based on the findings.&lt;/p&gt;

&lt;p&gt;Full write-up: &lt;a href="https://splatdev.com/blog/how-ai-agent-skills-produce-measurably-better-front-end-blind-tested/" rel="noopener noreferrer"&gt;https://splatdev.com/blog/how-ai-agent-skills-produce-measurably-better-front-end-blind-tested/&lt;/a&gt;&lt;/p&gt;

</description>
    </item>
    <item>
      <title>We published the article most "AI made our code better" posts skip — the honest</title>
      <dc:creator>Carlos Casalicchio</dc:creator>
      <pubDate>Sat, 18 Jul 2026 16:41:45 +0000</pubDate>
      <link>https://dev.to/ccasalicchio/we-published-the-article-most-ai-made-our-code-better-posts-skip-the-honest-1o8p</link>
      <guid>https://dev.to/ccasalicchio/we-published-the-article-most-ai-made-our-code-better-posts-skip-the-honest-1o8p</guid>
      <description>&lt;p&gt;We published the article most "AI made our code better" posts skip — the honest comparison.&lt;/p&gt;

&lt;p&gt;Results:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Correctness: tie (~90% vs ~95%)&lt;/li&gt;
&lt;li&gt;Quality (blind test): skills won both tasks at high confidence&lt;/li&gt;
&lt;li&gt;Cost: 24% more tokens, 33% more time&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The insight: "Correctness is solved. Craft is not." The durable gains come from verification loops, not longer instruction files.&lt;/p&gt;

&lt;p&gt;Full article: &lt;a href="https://splatdev.com/blog/ai-agent-skills-for-front-end-the-gains-the-gaps-and-an-honest-comparison/" rel="noopener noreferrer"&gt;https://splatdev.com/blog/ai-agent-skills-for-front-end-the-gains-the-gaps-and-an-honest-comparison/&lt;/a&gt;&lt;/p&gt;

</description>
    </item>
    <item>
      <title>We just published research on how AI agent skills perform across model tiers. Ke</title>
      <dc:creator>Carlos Casalicchio</dc:creator>
      <pubDate>Fri, 17 Jul 2026 16:41:39 +0000</pubDate>
      <link>https://dev.to/ccasalicchio/we-just-published-research-on-how-ai-agent-skills-perform-across-model-tiers-ke-n70</link>
      <guid>https://dev.to/ccasalicchio/we-just-published-research-on-how-ai-agent-skills-perform-across-model-tiers-ke-n70</guid>
      <description>&lt;p&gt;We just published research on how AI agent skills perform across model tiers.&lt;/p&gt;

&lt;p&gt;Key finding: Knowledge skills are a bigger win on cheaper models. The correctness lift roughly triples from frontier to smallest model.&lt;/p&gt;

&lt;p&gt;But there's a nuance: the "taste" (design system, anti-pattern guardrails) transfers down-tier, but the verification loop (render, critique, fix) doesn't — it needs a browser and enough model capability.&lt;/p&gt;

&lt;p&gt;We added an automated quality gate to address this.&lt;/p&gt;

&lt;p&gt;Full article: &lt;a href="https://splatdev.com/blog/do-ai-agent-skills-help-weaker-models-more-we-measured-it/" rel="noopener noreferrer"&gt;https://splatdev.com/blog/do-ai-agent-skills-help-weaker-models-more-we-measured-it/&lt;/a&gt;&lt;/p&gt;

</description>
    </item>
    <item>
      <title>We just published an in-depth look at how AI-enhanced .NET development is transf</title>
      <dc:creator>Carlos Casalicchio</dc:creator>
      <pubDate>Thu, 16 Jul 2026 16:41:34 +0000</pubDate>
      <link>https://dev.to/ccasalicchio/we-just-published-an-in-depth-look-at-how-ai-enhanced-net-development-is-transf-1ib0</link>
      <guid>https://dev.to/ccasalicchio/we-just-published-an-in-depth-look-at-how-ai-enhanced-net-development-is-transf-1ib0</guid>
      <description>&lt;p&gt;We just published an in-depth look at how AI-enhanced .NET development is transforming software agencies.&lt;/p&gt;

&lt;p&gt;Key findings from 30+ agency deployments:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;35% faster project delivery&lt;/li&gt;
&lt;li&gt;27% cost reduction&lt;/li&gt;
&lt;li&gt;57% fewer bugs&lt;/li&gt;
&lt;li&gt;58% increase in developer productivity&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The article covers intelligent code generation, automated testing, performance optimization, and security analysis.&lt;/p&gt;

&lt;p&gt;Full write-up: &lt;a href="https://splatdev.com/blog/why-agencies-are-switching-to-ai-enhanced-net-development/" rel="noopener noreferrer"&gt;https://splatdev.com/blog/why-agencies-are-switching-to-ai-enhanced-net-development/&lt;/a&gt;&lt;/p&gt;

</description>
    </item>
    <item>
      <title>We just published an in-depth look at how AI-powered case management is transfor</title>
      <dc:creator>Carlos Casalicchio</dc:creator>
      <pubDate>Wed, 15 Jul 2026 16:04:26 +0000</pubDate>
      <link>https://dev.to/ccasalicchio/we-just-published-an-in-depth-look-at-how-ai-powered-case-management-is-transfor-569e</link>
      <guid>https://dev.to/ccasalicchio/we-just-published-an-in-depth-look-at-how-ai-powered-case-management-is-transfor-569e</guid>
      <description>&lt;p&gt;We just published an in-depth look at how AI-powered case management is transforming small law firms.&lt;/p&gt;

&lt;p&gt;The technical angle: we built CaseOrbit AI to automate document classification, deadline tracking, and client communication for legal practices. Customer data shows 10+ hours/week saved per firm on average.&lt;/p&gt;

&lt;p&gt;Key features:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Automatic document classification and tagging&lt;/li&gt;
&lt;li&gt;Intelligent deadline tracking with urgency-based reminders&lt;/li&gt;
&lt;li&gt;Client communication automation via smart chatbots&lt;/li&gt;
&lt;li&gt;Integration with existing practice management tools&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Full article with implementation roadmap: &lt;a href="https://splatdev.com/blog/how-ai-case-management-saves-small-law-firms-10plus-hours-weekly/" rel="noopener noreferrer"&gt;https://splatdev.com/blog/how-ai-case-management-saves-small-law-firms-10plus-hours-weekly/&lt;/a&gt;&lt;br&gt;
(Results vary by firm. CaseOrbit customer data.)&lt;/p&gt;

&lt;p&gt;Would love feedback from anyone working on legal tech or AI automation!&lt;/p&gt;

</description>
    </item>
    <item>
      <title>We just published a detailed investigation into whether AI agent skills can push</title>
      <dc:creator>Carlos Casalicchio</dc:creator>
      <pubDate>Tue, 14 Jul 2026 16:04:14 +0000</pubDate>
      <link>https://dev.to/ccasalicchio/we-just-published-a-detailed-investigation-into-whether-ai-agent-skills-can-push-42bb</link>
      <guid>https://dev.to/ccasalicchio/we-just-published-a-detailed-investigation-into-whether-ai-agent-skills-can-push-42bb</guid>
      <description>&lt;p&gt;We just published a detailed investigation into whether AI agent skills can push front-end from "competent" to "genuinely beautiful" — with blind testing. Built 8 skills (7 knowledge + 1 process). First eval: tie. Blind test: skilled won both tasks. Process skill was the breakthrough. &lt;a href="https://splatdev.com/blog/how-ai-agent-skills-produce-measurably-better-front-end-blind-tested/" rel="noopener noreferrer"&gt;https://splatdev.com/blog/how-ai-agent-skills-produce-measurably-better-front-end-blind-tested/&lt;/a&gt; #ai #agentskills #frontend&lt;/p&gt;

</description>
      <category>agents</category>
      <category>ai</category>
      <category>frontend</category>
      <category>testing</category>
    </item>
    <item>
      <title>We published the article most "AI made our code better" posts skip — the honest</title>
      <dc:creator>Carlos Casalicchio</dc:creator>
      <pubDate>Mon, 13 Jul 2026 16:04:08 +0000</pubDate>
      <link>https://dev.to/ccasalicchio/we-published-the-article-most-ai-made-our-code-better-posts-skip-the-honest-65</link>
      <guid>https://dev.to/ccasalicchio/we-published-the-article-most-ai-made-our-code-better-posts-skip-the-honest-65</guid>
      <description>&lt;p&gt;We published the article most "AI made our code better" posts skip — the honest comparison. Results: Correctness tie (~90% vs ~95%), Quality (blind test): skills won both tasks at high confidence, Cost: 24% more tokens, 33% more time. The durable gains come from verification loops, not longer instruction files. &lt;a href="https://splatdev.com/blog/ai-agent-skills-for-front-end-the-gains-the-gaps-and-an-honest-comparison/" rel="noopener noreferrer"&gt;https://splatdev.com/blog/ai-agent-skills-for-front-end-the-gains-the-gaps-and-an-honest-comparison/&lt;/a&gt; #ai #agentskills #frontend&lt;/p&gt;

</description>
      <category>agents</category>
      <category>ai</category>
      <category>frontend</category>
      <category>programming</category>
    </item>
    <item>
      <title>We just published research on how AI agent skills perform across model tiers. Ke</title>
      <dc:creator>Carlos Casalicchio</dc:creator>
      <pubDate>Sun, 12 Jul 2026 16:04:00 +0000</pubDate>
      <link>https://dev.to/ccasalicchio/we-just-published-research-on-how-ai-agent-skills-perform-across-model-tiers-ke-46nj</link>
      <guid>https://dev.to/ccasalicchio/we-just-published-research-on-how-ai-agent-skills-perform-across-model-tiers-ke-46nj</guid>
      <description>&lt;p&gt;We just published research on how AI agent skills perform across model tiers. Key finding: Knowledge skills are a bigger win on cheaper models — the correctness lift roughly triples from frontier to smallest. Nuance: taste transfers down-tier, but the verification loop needs a browser and enough model capability. &lt;a href="https://splatdev.com/blog/do-ai-agent-skills-help-weaker-models-more-we-measured-it/" rel="noopener noreferrer"&gt;https://splatdev.com/blog/do-ai-agent-skills-help-weaker-models-more-we-measured-it/&lt;/a&gt; #ai #agentskills #evaluation&lt;/p&gt;

</description>
      <category>agents</category>
      <category>ai</category>
      <category>llm</category>
      <category>machinelearning</category>
    </item>
    <item>
      <title>We just published an in-depth look at how AI-enhanced .NET development is transf</title>
      <dc:creator>Carlos Casalicchio</dc:creator>
      <pubDate>Sat, 11 Jul 2026 16:03:54 +0000</pubDate>
      <link>https://dev.to/ccasalicchio/we-just-published-an-in-depth-look-at-how-ai-enhanced-net-development-is-transf-189n</link>
      <guid>https://dev.to/ccasalicchio/we-just-published-an-in-depth-look-at-how-ai-enhanced-net-development-is-transf-189n</guid>
      <description>&lt;p&gt;We just published an in-depth look at how AI-enhanced .NET development is transforming software agencies. Key findings from 30+ agency deployments: 35% faster delivery, 27% cost reduction, 57% fewer bugs, 58% increase in productivity. Covers intelligent code generation, automated testing, performance optimization, and security analysis. &lt;a href="https://splatdev.com/blog/why-agencies-are-switching-to-ai-enhanced-net-development/" rel="noopener noreferrer"&gt;https://splatdev.com/blog/why-agencies-are-switching-to-ai-enhanced-net-development/&lt;/a&gt; #dotnet #ai #softwaredevelopment&lt;/p&gt;

</description>
    </item>
    <item>
      <title>We just published an in-depth look at how AI-powered case management is transfor</title>
      <dc:creator>Carlos Casalicchio</dc:creator>
      <pubDate>Fri, 10 Jul 2026 16:03:49 +0000</pubDate>
      <link>https://dev.to/ccasalicchio/we-just-published-an-in-depth-look-at-how-ai-powered-case-management-is-transfor-fjo</link>
      <guid>https://dev.to/ccasalicchio/we-just-published-an-in-depth-look-at-how-ai-powered-case-management-is-transfor-fjo</guid>
      <description>&lt;p&gt;We just published an in-depth look at how AI-powered case management is transforming small law firms. The technical angle: we built CaseOrbit AI to automate document classification, deadline tracking, and client communication. Key features: automatic document tagging, intelligent deadline tracking with urgency-based reminders, client communication automation. Full article: &lt;a href="https://splatdev.com/caseorbit-ai-time-savings" rel="noopener noreferrer"&gt;https://splatdev.com/caseorbit-ai-time-savings&lt;/a&gt; (Results vary by firm.) #legaltech #ai #casemanagement #saas&lt;/p&gt;

</description>
    </item>
  </channel>
</rss>
