Imp is a full port of DSPy to the BEAM

(github.com)

53 points by mpweiher7 hours ago

2 comments

  • dang7 hours ago
    Related. Others?<p><i>If DSPy is so great, why isn&#x27;t anyone using it?</i> - <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=47490365">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=47490365</a> - March 2026 (120 comments)<p><i>DSPy – Programming–not prompting–LMs</i> - <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=42343692">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=42343692</a> - Dec 2024 (45 comments)<p><i>Betting on DSPy for Systems of LLMs</i> - <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=41213561">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=41213561</a> - Aug 2024 (18 comments)<p><i>DSPy: Framework for programming with foundation models</i> - <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=37417698">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=37417698</a> - Sept 2023 (52 comments)
    • visarga7 hours ago
      We moved from structured decoding to tool calls and agents a long time ago. It was more useful when models were too weak to hold syntax reliably.
      • vatsachak7 hours ago
        Yeah. Myself, a PhD in math, have less attention to detail than models.<p>Now their main weakness is creativity, planning and decision making
      • genxy7 hours ago
        Techniques in DSPy are still extremely valuable. Your comment is akin to saying, we don&#x27;t need feedback because open loop control is so good now. It is until it isn&#x27;t. I don&#x27;t need a deep understanding of the underlying theory because my gut reaction is always right, until it isn&#x27;t.
  • jauntywundrkind4 hours ago
    Are there TS or rust ports? Please? :)<p>Edit, follow-up:<p>DSRr in rust: <a href="https:&#x2F;&#x2F;github.com&#x2F;krypticmouse&#x2F;DSRs" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;krypticmouse&#x2F;DSRs</a>