<script data-pm-proxy="intercept"></script><?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:itunes="http://www.itunes.com/dtds/podcast-1.0.dtd" xmlns:googleplay="http://www.google.com/schemas/play-podcasts/1.0"><channel><title><![CDATA[Tiny Frequencies]]></title><description><![CDATA[Tiny Frequencies explores the limitations of AI products and is published 2x a month by Anita. Lucky for you, she asks "why?" a lot as a product manager and former machine learning engineer. This is her answer.]]></description><link>https://tinyfreqs.substack.com</link><image><url>https://substackcdn.com/image/fetch/$s_!mKrC!,w_256,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Ftinyfreqs.substack.com%2Fimg%2Fsubstack.png</url><title>Tiny Frequencies</title><link>https://tinyfreqs.substack.com</link></image><generator>Substack</generator><lastBuildDate>Thu, 03 Sep 2026 20:06:15 GMT</lastBuildDate><atom:link href="/__u/tinyfreqs.substack.com/feed" rel="self" type="application/rss+xml"/><copyright><![CDATA[Anita Mehrotra]]></copyright><language><![CDATA[en]]></language><webMaster><![CDATA[tinyfreqs@substack.com]]></webMaster><itunes:owner><itunes:email><![CDATA[tinyfreqs@substack.com]]></itunes:email><itunes:name><![CDATA[Anita]]></itunes:name></itunes:owner><itunes:author><![CDATA[Anita]]></itunes:author><googleplay:owner><![CDATA[tinyfreqs@substack.com]]></googleplay:owner><googleplay:email><![CDATA[tinyfreqs@substack.com]]></googleplay:email><googleplay:author><![CDATA[Anita]]></googleplay:author><itunes:block><![CDATA[Yes]]></itunes:block><item><title><![CDATA[Your business probably doesn't need an LLM]]></title><description><![CDATA[And some recs for what to invest in instead]]></description><link>https://tinyfreqs.substack.com/p/your-business-probably-doesnt-need</link><guid isPermaLink="false">https://tinyfreqs.substack.com/p/your-business-probably-doesnt-need</guid><dc:creator><![CDATA[Anita]]></dc:creator><pubDate>Tue, 25 Nov 2025 23:27:12 GMT</pubDate><content:encoded><![CDATA[<p>The solutions that will enable companies to solve 80-90% of the problem are simple solutions, not complex ones. In most cases, generative AI is not the correct solution for problems the business needs to prioritize tackling.</p><p>More complex models do not always lead to better business outcomes. I&#8217;m loath to quote Ernest Hemingway for his general character a&#8230;</p>
      <p>
          <a href="/__u/tinyfreqs.substack.com/p/your-business-probably-doesnt-need">
              Read more
          </a>
      </p>
   ]]></content:encoded></item><item><title><![CDATA[What LLMs can't do well today]]></title><description><![CDATA[What is distribution shift?]]></description><link>https://tinyfreqs.substack.com/p/what-llms-cant-do-well-today</link><guid isPermaLink="false">https://tinyfreqs.substack.com/p/what-llms-cant-do-well-today</guid><dc:creator><![CDATA[Anita]]></dc:creator><pubDate>Sun, 09 Nov 2025 19:08:02 GMT</pubDate><enclosure url="https://substackcdn.com/image/fetch/$s_!-KiX!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6e6f815e-1a5e-42c0-8867-cca78d331f32_1322x1016.png" length="0" type="image/jpeg"/><content:encoded><![CDATA[<p>I&#8217;ve historically been skeptical about mainstream popularity of different ML / AI trends, convinced that it&#8217;s overblown, and that most problems in industry are probably solvable with far less complex ML solutions &#8212; if one is even needed at all. In 2011, it was &#8220;big data&#8221; and Hadoop, then in 2016 it was natural language processing, and since early 2023 i&#8230;</p>
      <p>
          <a href="/__u/tinyfreqs.substack.com/p/what-llms-cant-do-well-today">
              Read more
          </a>
      </p>
   ]]></content:encoded></item><item><title><![CDATA[So you're using a chat bot as your therapist]]></title><description><![CDATA[Two reasons why, since you are 100% not alone]]></description><link>https://tinyfreqs.substack.com/p/so-youre-using-a-chat-bot-as-your-therapist</link><guid isPermaLink="false">https://tinyfreqs.substack.com/p/so-youre-using-a-chat-bot-as-your-therapist</guid><dc:creator><![CDATA[Anita]]></dc:creator><pubDate>Mon, 27 Oct 2025 21:00:58 GMT</pubDate><enclosure url="https://substackcdn.com/image/fetch/$s_!Ep8S!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9c3a1664-0d4b-4459-bc16-08bea72d4809_660x330.png" length="0" type="image/jpeg"/><content:encoded><![CDATA[<p>On a balmy Thursday night this April, at the New York City classique Old Town Bar, I found myself deep in discussion with some girl friends about the state of the world (burning), relationships (very hard, always entertaining), and the business of art (changing-ish). Slowly - and then all of a sudden - it became clear that each of them was using a chat &#8230;</p>
      <p>
          <a href="/__u/tinyfreqs.substack.com/p/so-youre-using-a-chat-bot-as-your-therapist">
              Read more
          </a>
      </p>
   ]]></content:encoded></item><item><title><![CDATA[I recently learned about system-dependent paths]]></title><description><![CDATA[This is a thing from thermodynamics, but it's also a thing for building the right product!!!]]></description><link>https://tinyfreqs.substack.com/p/i-recently-learned-about-system-dependent</link><guid isPermaLink="false">https://tinyfreqs.substack.com/p/i-recently-learned-about-system-dependent</guid><dc:creator><![CDATA[Anita]]></dc:creator><pubDate>Mon, 15 Sep 2025 01:34:32 GMT</pubDate><content:encoded><![CDATA[<p>I recently learned about the idea of a &#8220;system-dependent path&#8221; A.K.A. &#8220;path-dependent process.&#8221; It is a concept in physics and seems fancy, but really is just what it sounds like: the specific path followed by a system depends on the conditions of that system. And what this means is that the final state depends on <em>the specific sequence of events taken b&#8230;</em></p>
      <p>
          <a href="/__u/tinyfreqs.substack.com/p/i-recently-learned-about-system-dependent">
              Read more
          </a>
      </p>
   ]]></content:encoded></item><item><title><![CDATA[Three-ish flavors of product management cultures]]></title><description><![CDATA[From my research so far]]></description><link>https://tinyfreqs.substack.com/p/three-flavors-of-product-management</link><guid isPermaLink="false">https://tinyfreqs.substack.com/p/three-flavors-of-product-management</guid><dc:creator><![CDATA[Anita]]></dc:creator><pubDate>Thu, 04 Sep 2025 04:38:22 GMT</pubDate><content:encoded><![CDATA[<p>As you know, I work in product management, and the thing that is really interesting about this craft is that:</p><ol><li><p>It is still pretty opaque, so</p></li><li><p>not all product managers (PMs) work the same way.</p></li></ol><p>For me, this is kind of fun? Like we are <em>still</em> figuring things out. It reminds me a lot of working as a data scientist when the industry was nascent and people would as&#8230;</p>
      <p>
          <a href="/__u/tinyfreqs.substack.com/p/three-flavors-of-product-management">
              Read more
          </a>
      </p>
   ]]></content:encoded></item><item><title><![CDATA[What is useful genAI tool usage?]]></title><description><![CDATA[Pls be smart when using genAI tools as a leader]]></description><link>https://tinyfreqs.substack.com/p/what-is-useful-genai-tool-usage</link><guid isPermaLink="false">https://tinyfreqs.substack.com/p/what-is-useful-genai-tool-usage</guid><dc:creator><![CDATA[Anita]]></dc:creator><pubDate>Tue, 26 Aug 2025 01:20:55 GMT</pubDate><content:encoded><![CDATA[<p>In April, the press blew up when Shopify&#8217;s CEO&#8217;s memo about <a href="https://x.com/tobi/status/1909251946235437514?lang=en">reflexive AI use</a> was leaked. Shortly afterwards, there was an internal acceleration to measure genAI tool usage, by doing things like tracking the number of <a href="https://x.com/karpathy/status/1886192184808149383?lang=en">vibecoded pull requests</a> put out by product managers (&#8220;PMs&#8221;) and other non-engineers. </p><p>Engineering thought-leadership has <a href="https://martinfowler.com/bliki/OutcomeOverOutput.html">been against metri&#8230;</a></p>
      <p>
          <a href="/__u/tinyfreqs.substack.com/p/what-is-useful-genai-tool-usage">
              Read more
          </a>
      </p>
   ]]></content:encoded></item><item><title><![CDATA[ChatGPT's a smart product]]></title><description><![CDATA[The product team paid a lot of attention to the overall experience, rather than solely riding on the improvement of its underlying model]]></description><link>https://tinyfreqs.substack.com/p/a-tiny-thought-chatgpt-is-a-smart</link><guid isPermaLink="false">https://tinyfreqs.substack.com/p/a-tiny-thought-chatgpt-is-a-smart</guid><dc:creator><![CDATA[Anita]]></dc:creator><pubDate>Sat, 22 Jun 2024 12:49:00 GMT</pubDate><content:encoded><![CDATA[<p>If your company is building a product that uses an LLM, you might at some point find yourself advocating to continue investing in it simply because it means your company is building a generative AI product. This would be entirely understandable: after all, you saw OpenAI release one that became <a href="https://www.reuters.com/technology/chatgpt-sets-record-fastest-growing-user-base-analyst-note-2023-02-01/">very popular very quickly</a>.</p><p>But what you really should be pay&#8230;</p>
      <p>
          <a href="/__u/tinyfreqs.substack.com/p/a-tiny-thought-chatgpt-is-a-smart">
              Read more
          </a>
      </p>
   ]]></content:encoded></item><item><title><![CDATA[Career ladders for data scientists]]></title><description><![CDATA[A set of open-source career ladders for data scientists and orgs, alike]]></description><link>https://tinyfreqs.substack.com/p/data-science-career-ladders</link><guid isPermaLink="false">https://tinyfreqs.substack.com/p/data-science-career-ladders</guid><dc:creator><![CDATA[Anita]]></dc:creator><pubDate>Wed, 15 Jun 2022 17:43:00 GMT</pubDate><content:encoded><![CDATA[<p>Several years ago, I stumbled into data science while the field in industry was nascent. Since then, I&#8217;ve navigated my career via trial and error. Today, I&#8217;m sharing the below <strong>data science career ladders </strong>(linked to at the bottom of this article), a culmination of my hard-knock learnings and research, with the hope that they are useful to the broader com&#8230;</p>
      <p>
          <a href="/__u/tinyfreqs.substack.com/p/data-science-career-ladders">
              Read more
          </a>
      </p>
   ]]></content:encoded></item><item><title><![CDATA[How to leverage a data science / machine learning background as a product manager]]></title><description><![CDATA[Early in my product management career, I was told I had a unique point of view because of my data science background.]]></description><link>https://tinyfreqs.substack.com/p/how-to-leverage-a-data-science-background</link><guid isPermaLink="false">https://tinyfreqs.substack.com/p/how-to-leverage-a-data-science-background</guid><dc:creator><![CDATA[Anita]]></dc:creator><pubDate>Thu, 07 Oct 2021 12:56:00 GMT</pubDate><enclosure url="https://substackcdn.com/image/fetch/$s_!1ygS!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fba440f60-cc6e-400c-94b2-18c1790a1756_1200x769.png" length="0" type="image/jpeg"/><content:encoded><![CDATA[<p>Early in my <a href="https://a16z.com/2012/06/15/good-product-managerbad-product-manager/">product management</a> career, I was told I had a unique point of view because of my data science background. I was skeptical&#8202;&#8212;&#8202;and hard-pressed to find guidance in existing product management literature.</p><p>Since then, I&#8217;ve learned that product managers (PMs) with a background in data science have a superpower that can turbo-charge their team&#8217;s imp&#8230;</p>
      <p>
          <a href="/__u/tinyfreqs.substack.com/p/how-to-leverage-a-data-science-background">
              Read more
          </a>
      </p>
   ]]></content:encoded></item><item><title><![CDATA[A model for widening the AI innovation funnel]]></title><description><![CDATA[A couple of years ago, a friend shared that he was a big Container Store fan &#8212; he believed all things tended towards chaos, and there was something incredibly relieving about having a place to organize your life.]]></description><link>https://tinyfreqs.substack.com/p/a-model-for-widening-the-ai-innovation</link><guid isPermaLink="false">https://tinyfreqs.substack.com/p/a-model-for-widening-the-ai-innovation</guid><dc:creator><![CDATA[Anita]]></dc:creator><pubDate>Tue, 02 Feb 2021 01:58:00 GMT</pubDate><enclosure url="https://substackcdn.com/image/fetch/$s_!k8r9!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa19c40e1-a045-4751-b00c-645171e1e842_523x353.png" length="0" type="image/jpeg"/><content:encoded><![CDATA[<p>A couple of years ago, a friend shared that he was a big Container Store fan&#8202;&#8212;&#8202;he believed all things tended towards chaos, and there was something incredibly relieving about having a place to organize your life. That was the image that stuck with me in the first part of 2020, as I was building out a set of processes to <strong>facilitate artificial intelligenc&#8230;</strong></p>
      <p>
          <a href="/__u/tinyfreqs.substack.com/p/a-model-for-widening-the-ai-innovation">
              Read more
          </a>
      </p>
   ]]></content:encoded></item></channel></rss>