The AI Shortcut | Saif Shines <meta name="title" content="The AI Shortcut | Saif Shines"> <meta name="description" content="The AI Shortcut">  <meta property="og:type" content="website"> <meta property="og:url" content="https://saifshines.dev/thoughts/the-ai-shortcut-0f9d2a75/"> <meta property="og:title" content="The AI Shortcut | Saif Shines"> <meta property="og:description" content="The AI Shortcut">  <link rel="alternate" type="application/rss+xml" title="Saif Shines" href="/rss.xml">  <script> const theme = (() => { if (typeof localStorage !== "undefined" && localStorage.getItem("theme")) { return localStorage.getItem("theme"); } return window.matchMedia("(prefers-color-scheme: dark)").matches ? "dark" : "light"; })(); document.documentElement.classList.toggle("dark", theme === "dark"); </script>  <script type="module">function e(){document.querySelectorAll(".animate").forEach((t,a)=>{setTimeout(()=>t.classList.add("show"),a*150)})}document.addEventListener("DOMContentLoaded",e);document.addEventListener("astro:after-swap",e);</script>  <script type="module">function e(){document.documentElement.classList.toggle("scrolled",document.documentElement.scrollTop>100)}document.addEventListener("scroll",e);</script>

Casual conversations of my friends working in software industry, either in design, engineering or marketing; all of them feel a push to use AI in pursuit of productivity gains.

Most probably you too.

Before we felt the push:
- Time budget for a project: 2 weeks
- Time took for 75% completion: 1 week
- Time took for 95% completion: another 1 week

After the AI push:
- Time budget for a project: 1 week
- Time took for 75% completion: 2 days
- Time took for 95% completion: 1.5 weeks

Rest of 5% is ever-existing spillovers irrespective of AI use.

Observations:
- That 75% completion, has certainly become a joy to do.
- Growing magic bullet fallacy about AI systems within mid-management.
- Almost every one skips struggle from 75% to 95%

Maybe the joy is in those first leaps, the spark of creation. But I wonder, without the grit from 75% to 95%, are we missing a deeper learning?