from <a href="/en/2023-02-14%20Organize%20the%20top%20page%20leads">2023-02-14 Organize the top page leads</a>
An example of how optimizing for short-term rewards can weaken
<ul>
<li><blockquote>
<a href="https://twitter.com/tsukammo/status/1624677427669770241?s=20&t=ICqRkCEN-0YDZcWlR0WuPQ">@tsukammo</a>: I&#39;m having trouble explaining why <a href="/en/Life%20Optimization">Life Optimization</a> doesn&#39;t work, <a href="/en/%20game%20tree"> game tree</a> <a href="/en/search">search</a>.
</blockquote>
</li>
<li><blockquote>
 <img src="https://gyazo.com/597878edc889a3c2489d01be73177041/thumb/1000" alt="image">
</blockquote>
<ul>
<li><blockquote>
<a href="https://twitter.com/tsukammo/status/1624680298947678209?s=20&t=ICqRkCEN-0YDZcWlR0WuPQ">@tsukammo</a>: This is what happens with an evaluation function based on direct rewards alone, so a common &quot; lifehacks&quot; are optimizing the <a href="/en/evaluation%20function">evaluation function</a> with &quot;<a href="/en/curiosity">curiosity</a>&quot; or &quot;prepare a reward by chopping in small steps&quot;.
</blockquote>
</li>
<li><blockquote>
Yeah, I know all that. I just don&#39;t.
</blockquote>
</li>
</ul>
</li>
<li><a href="/en/Trade-offs%20between%20use%20and%20exploration">Trade-offs between use and exploration</a>
</li>
</ul>
<hr>
This page is auto-translated from <a href="https://scrapbox.io/nishio/%E7%9F%AD%E6%9C%9F%E7%9A%84%E5%A0%B1%E9%85%AC%E3%81%AB%E6%9C%80%E9%81%A9%E5%8C%96%E3%81%99%E3%82%8B%E3%81%A8%E5%BC%B1%E3%81%8F%E3%81%AA%E3%82%8B%E4%BE%8B">/nishio/短期的報酬に最適化すると弱くなる例</a> using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at <a href="https://twitter.com/nishio_en">@nishio_en</a>. I&#39;m very happy to spread my thought to non-Japanese readers.