<p>In recent weeks, I've been experimenting with the idea - can Claude work with my task list by himself?</p>
<p>I asked him: take all the tasks for this week + those without a date, and break them into three groups:</p>
<ol>
<li>You can do it yourself</li>
<li>My assistant (human) can do it</li>
<li>Only I can do it</li>
</ol>
<p>He sorted them out. I looked skeptically - “no way, he definitely can’t handle it without me” - and set it aside.</p>
<p>After a while, I decided to give him a chance: “Okay, do what you promised to do yourself.”</p>
<p>And then the interesting part began:</p>
<ul>
<li>“Here, you wanted to add functionality to the project. I found the code, checked it - it’s already ready. The task is no longer relevant, throw it out.”</li>
<li>“And here you wanted to generate something. I found where you have it - it’s already done. Here’s the result.”</li>
</ul>
<p>I was sure he would start bombarding me with questions or do the wrong thing. But he managed.</p>
<p>He also helped with tasks for my assistant - “these are pending, give them to the person, I can’t do them, but there’s no need to keep them on the list.”</p>
</p>It seems we are not that far from the request “make my life better.”</p>
#ai #productivity #claudecode
<p><a href="https://t.me/+OvImEUmA7W5mYTRi">————————— Мысли Рвачева —————————</a></p>