<?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: Zakhar Garan (Гаран Захар)</title>
    <description>The latest articles on DEV Community by Zakhar Garan (Гаран Захар) (@zakhar88).</description>
    <link>https://dev.to/zakhar88</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%2F4250140%2F34eff7d3-846f-4a91-93cf-7d4b8e8e9da4.png</url>
      <title>DEV Community: Zakhar Garan (Гаран Захар)</title>
      <link>https://dev.to/zakhar88</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/zakhar88"/>
    <language>en</language>
    <item>
      <title>Fake stories about AI making money on its own</title>
      <dc:creator>Zakhar Garan (Гаран Захар)</dc:creator>
      <pubDate>Tue, 17 Nov 2026 15:28:07 +0000</pubDate>
      <link>https://dev.to/zakhar88/fake-stories-about-ai-making-money-on-its-own-1o98</link>
      <guid>https://dev.to/zakhar88/fake-stories-about-ai-making-money-on-its-own-1o98</guid>
      <description>&lt;p&gt;Have you seen those fake stories about AI making money on its own? And lots of people believe them.&lt;/p&gt;

&lt;p&gt;I came across a popular Facebook post: an agent supposedly ran a blog in the background for a whole year, it became so popular that a client came to order a sponsored post for $150, the agent negotiated with him itself and even opened a bank account at Monobank on its own. That last part is the funniest: Monobank might want to review its security rules for dealing with agents.&lt;/p&gt;

&lt;p&gt;And at the end of this fantastic story, where the client supposedly did pay the $150, the author of course invites you to a webinar where he will tell you how to earn the same way.&lt;/p&gt;

&lt;p&gt;What worries me most is that the comments are full of interested people. We are probably in for many more schemes and AI-assisted scam attempts. Be wary of stories that end with an offer to buy something.&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://zakhargaran.com/en/blog/fake-ai-income-stories/" rel="noopener noreferrer"&gt;zakhargaran.com&lt;/a&gt;. I'm Zakhar Garan, a Senior iOS Engineer and iOS Tech Lead from Ukraine and the creator of &lt;a href="https://zakhargaran.com/en/home-rate/" rel="noopener noreferrer"&gt;Home Rate&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>security</category>
      <category>career</category>
    </item>
    <item>
      <title>AI and video editing: how easy it has become</title>
      <dc:creator>Zakhar Garan (Гаран Захар)</dc:creator>
      <pubDate>Tue, 17 Nov 2026 15:27:53 +0000</pubDate>
      <link>https://dev.to/zakhar88/ai-and-video-editing-how-easy-it-has-become-59jd</link>
      <guid>https://dev.to/zakhar88/ai-and-video-editing-how-easy-it-has-become-59jd</guid>
      <description>&lt;p&gt;I used to run two YouTube channels, one about films and one about iOS development. To learn video editing I took a 40-hour course plus dozens of hours of lessons on voice processing. And now? Now AI does it all.&lt;/p&gt;

&lt;p&gt;I came across a video about Claude + CapCut, and it turns out AI can do everything: you just explain how to cut the video and which effects to add, give it the files, images or music, and it edits the whole thing itself.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why CapCut is not the best choice
&lt;/h2&gt;

&lt;p&gt;CapCut currently has no MCP, a special “port” for connecting AI, so an agent works with it very inefficiently. One of the best video editors, the free DaVinci Resolve, has not only MCP but also many built-in AI tools. Through MCP the agent works far more economically and can use those built-in tools too, for example for voice processing.&lt;/p&gt;

&lt;h2&gt;
  
  
  A tip about thumbnails
&lt;/h2&gt;

&lt;p&gt;I would not generate video thumbnails entirely with AI. There are so many of them now, and they look so alike, that in my opinion they put viewers off.&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://zakhargaran.com/en/blog/ai-video-editing/" rel="noopener noreferrer"&gt;zakhargaran.com&lt;/a&gt;. I'm Zakhar Garan, a Senior iOS Engineer and iOS Tech Lead from Ukraine and the creator of &lt;a href="https://zakhargaran.com/en/home-rate/" rel="noopener noreferrer"&gt;Home Rate&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>mcp</category>
      <category>claude</category>
      <category>productivity</category>
    </item>
    <item>
      <title>Fable, Opus, Sonnet or Haiku? How not to overpay for AI without losing quality</title>
      <dc:creator>Zakhar Garan (Гаран Захар)</dc:creator>
      <pubDate>Tue, 17 Nov 2026 15:27:41 +0000</pubDate>
      <link>https://dev.to/zakhar88/fable-opus-sonnet-or-haiku-how-not-to-overpay-for-ai-without-losing-quality-4ieo</link>
      <guid>https://dev.to/zakhar88/fable-opus-sonnet-or-haiku-how-not-to-overpay-for-ai-without-losing-quality-4ieo</guid>
      <description>&lt;p&gt;Anthropic has released Claude Fable 5.2, so which model should you pick for which task?&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Opus 5.5 is recommended as the default for complex code, at medium effort. If the result is weak, raise the effort level first and only then switch models: it is cheaper, and the cache is preserved.&lt;/li&gt;
&lt;li&gt;Fable 5.2 is for large, fuzzy tasks where a mistake is expensive, such as rewriting the architecture of a whole product. Its tokens cost twice as much as Opus, but it plans on its own and tests as it goes.&lt;/li&gt;
&lt;li&gt;Sonnet suits well-defined tasks where speed and budget matter most, such as a bug that is easy to reproduce. It costs half as much as Opus.&lt;/li&gt;
&lt;li&gt;Haiku is for large volumes of routine work: sorting customer requests or writing short summaries.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;You can combine models. For example, Fable can lead the migration of a whole project to a new framework version and hand off updates of individual packages to Sonnet subagents.&lt;/p&gt;

&lt;p&gt;Are you already combining different models in your work?&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://zakhargaran.com/en/blog/fable-opus-sonnet-or-haiku/" rel="noopener noreferrer"&gt;zakhargaran.com&lt;/a&gt;. I'm Zakhar Garan, a Senior iOS Engineer and iOS Tech Lead from Ukraine and the creator of &lt;a href="https://zakhargaran.com/en/home-rate/" rel="noopener noreferrer"&gt;Home Rate&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>claude</category>
      <category>llm</category>
      <category>productivity</category>
    </item>
    <item>
      <title>How to save AI tokens</title>
      <dc:creator>Zakhar Garan (Гаран Захар)</dc:creator>
      <pubDate>Tue, 17 Nov 2026 15:27:25 +0000</pubDate>
      <link>https://dev.to/zakhar88/how-to-save-ai-tokens-36pd</link>
      <guid>https://dev.to/zakhar88/how-to-save-ai-tokens-36pd</guid>
      <description>&lt;p&gt;Working with AI agents every day, I noticed that what drives token usage most is not the model but how you frame the task. Here is what helps me spend less.&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;The more detailed the prompt, the less time the AI spends hunting for details. If you know exactly where the fix goes, say so, so the agent does not waste tokens searching for files or functions.&lt;/li&gt;
&lt;li&gt;In Claude and Cursor the cache lives for 30–60 minutes, so it is better to finish a task in one go rather than come back to it hours later.&lt;/li&gt;
&lt;li&gt;New task, new chat. The previous context will not get in the way, and there is room for the new one.&lt;/li&gt;
&lt;li&gt;Simpler models for simple changes. You can ask the AI to delegate easy tasks to simpler subagents.&lt;/li&gt;
&lt;li&gt;Keep only short, genuinely necessary instructions in CLAUDE.md or AGENTS.md.&lt;/li&gt;
&lt;li&gt;Watch your spending: /usage and /context in Claude Code, /status in Codex.&lt;/li&gt;
&lt;li&gt;And it works surprisingly well to ask the AI itself how to save tokens: it will analyse your chats and give recommendations.&lt;/li&gt;
&lt;/ul&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://zakhargaran.com/en/blog/how-to-save-ai-tokens/" rel="noopener noreferrer"&gt;zakhargaran.com&lt;/a&gt;. I'm Zakhar Garan, a Senior iOS Engineer and iOS Tech Lead from Ukraine and the creator of &lt;a href="https://zakhargaran.com/en/home-rate/" rel="noopener noreferrer"&gt;Home Rate&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>productivity</category>
      <category>claude</category>
      <category>cursor</category>
    </item>
  </channel>
</rss>
