{
  "version": "https://jsonfeed.org/version/1",
  "title": "predictions on Coté",
  "icon": "https://avatars.micro.blog/avatars/2025/42/8457.jpg",
  "home_page_url": "https://cote.io/",
  "feed_url": "https://cote.io/feed.json",
  "items": [
      {
        "id": "http://cote.micro.blog/2026/05/29/failure-is-normal.html",
        "title": "50%+ failure is normal",
        "content_html": "<span data-author=\"cote\">\n<blockquote>\n<a href=\"https://www.theregister.com/ai-ml/2026/05/28/most-generative-ai-and-custom-model-projects-will-be-a-bust-gartner/5247633?ref=cote.io\">Analyst firm Gartner thinks</a> at least half of all generative AI projects \"will overrun their budgeted costs due to poor architectural choices and lack of operational know-how,\" and most organizations that try to build custom models \"will abandon their efforts due to costs, complexity and technical debt in their deployments.\"</blockquote>\n<p><em>Yes, and</em> this matches decades software project <s>failure</s>success studies from the Standish Group.</p>\n<p>The success rate of projects has held steady forever:</p>\n<figure>\n<a href=\"https://cote.pizza/success/\">\n  <img src=\"https://cdn.uploads.micro.blog/3446/2026/photo-upload-screenshot-tape-machine-2026-05-29-at-10.03.592x.jpg\" width=\"600\" alt=\"Standish Group CHAOS Report stacked bar chart, 1994-2015, showing share of IT projects per year as successful, challenged, or failed. Success rates hold at roughly 26-35% across all years.\"></a>\n  <figcaption>Sources: <a href=\"https://www.swqual.com/verification_validation.html?ref=cote.io\">2009 study</a> and <a href=\"https://www.infoq.com/articles/standish-chaos-2015/?ref=cote.io\">2015 study overviews</a>.</figcaption>\n</figure>\n<p>That chart ends 11 years ago, but I haven't heard a lot of reports that the numbers have changed much over the years...case in point the opening quote!</p>\n<p>In IT and software, very few projects are successful on the iron triangle of budget, schedule, and quality<sup><a href=\"#fn-quality\" id=\"fnref-quality\">1</a></sup>.</p>\n<p>You can take this to mean that expectations were unrealistic, or that there is just genuine failure. I favor the first. I'm more of a \"I'm not late to this meeting, it was just scheduled at the wrong time\" kind of guy.</p>\n<p>There's some kind of <a href=\"https://cote.pizza/shit-people-say/#jevons-paradox\">Jevons Paradox</a> thing here. Each time we optimize how we make software, we then take on <em>more</em> challenging and difficult tasks, likely causing setbacks again. To me, this is what accounts for the low success rate.</p>\n<p>If you don't want to do new things and try to do better, you could get those success numbers up probably by just continuing to do what works.</p>\n<p>But, you add in a new technology, and while you're figuring it out, it feels like you're failing.</p>\n<p>Back in the digital transformation days, we'd be clever and say: you're not failing, you're <em>learning</em>.</p>\n<p>🔗 <a href=\"https://www.theregister.com/ai-ml/2026/05/28/most-generative-ai-and-custom-model-projects-will-be-a-bust-gartner/5247633?ref=cote.io\">Most generative AI and custom model projects will be a bust: Gartner</a></p>\n</span>\n<div class=\"footnotes\">\n<hr>\n<ol>\n<li id=\"fn-quality\">I think of quality as more than \"bug free.\" It also includes \"does the software solve our problem, are the features done well,\" e.g., did we get something useful? <a href=\"#fnref-quality\">↩</a>\n</li>\n</ol>\n</div>\n<!--\nAnnotations: 0,2241 SHA-256 0000000000000000000000\n@Coté <cote.io>: 0,2241\n&AI <Diane>:\n0% AI-written\n-->\n",
        "date_published": "2026-05-29T10:16:00+02:00",
        "url": "https://cote.io/2026/05/29/failure-is-normal.html",
        "tags": ["tech","link","enterpriseai","gartner","predictions","roi"]
      }
  ]
}