<p>I played around with the new subagents from Anthropic - it's really cool and exactly what was missing - creating a shell within which you can easily (without code) write your own agents and conveniently call them. Surprisingly, it seems to be my first real reason to move from OpenAI ChatGPT towards Anthropic.</p>
<p>What I managed to test:</p>
<ol>
<li>Content summarization when I throw in a link (doesn't matter how)</li>
<li>Downloading videos via the terminal</li>
</ol>
<p>What I expect:</p>
<ol>
<li>The emergence of repositories where people share their agents</li>
<li>A closer integration with browser use and computer use agents.</li>
</ol>
<p>From ideas: create an interface for a Telegram bot on top of Anthropic Claude Code, so that all of this can be used from a mobile phone in one place.</p>
<p>If you don't know where to start - I recommend this video - short, clear, and to the point <a href="https://www.youtube.com/watch?v=6MLyc3jYisQ">https://www.youtube.com/watch?v=6MLyc3jYisQ</a>;. I especially liked the idea of using Claude Code in conjunction with a text file editor (VSCode, Cursor, etc.) to track what it does.</p>
<p>π Source: <a href="https://docs.anthropic.com/en/docs/claude-code/sub-agents">https://docs.anthropic.com/en/docs/claude-code/sub-agents</a></p>;
<p>#anthropic #claude #ai #agi #agent #</p>
<p>βββββββββ<br>ΠΡΡΠ»ΠΈ Π Π²Π°ΡΠ΅Π²Π°<br>βββββββββ</p>
