<?xml version="1.0" encoding="UTF-8"?>
<feed xmlns="http://www.w3.org/2005/Atom">
  <title>Croftsware: Writing on Technology Leadership &amp; AI Strategy</title>
  <link href="https://www.croftsware.com/feed.xml" rel="self"/>
  <link href="https://www.croftsware.com/blog/"/>
  <id>https://www.croftsware.com/</id>
  <updated>2026-10-22T00:00:00.000Z</updated>
  <author><name>James Webster</name></author>
  <entry>
    <title>Why Not Just Build It?</title>
    <link href="https://www.croftsware.com/blog/why-not-just-build-it/"/>
    <id>https://www.croftsware.com/blog/why-not-just-build-it/</id>
    <published>2026-10-22T00:00:00.000Z</published>
    <updated>2026-10-22T00:00:00.000Z</updated>
    <summary>Anyone can now build the software. Nobody can build the years behind it. Nokia's $7.6bn jump, a 1999 antitrust judgment and nineteen thousand Steam games a year explain the three bills software makes you pay — and why AI only cuts the first.</summary>
    <category term="cheap software"/>
    <category term="build vs buy"/>
    <category term="AI coding"/>
    <category term="John Sutton"/>
    <category term="sunk costs"/>
    <category term="applications barrier to entry"/>
    <category term="Nokia"/>
    <category term="Windows Phone"/>
    <category term="membership CRM"/>
    <category term="Technology"/>
  </entry>
  <entry>
    <title>New Motor, Old Belts</title>
    <link href="https://www.croftsware.com/blog/new-motor-old-belts/"/>
    <id>https://www.croftsware.com/blog/new-motor-old-belts/</id>
    <published>2026-09-23T00:00:00.000Z</published>
    <updated>2026-09-23T00:00:00.000Z</updated>
    <summary>Cheap software does not redesign a firm. The mills only changed shape because the buildings wore out, and software never wears out. Until the old thing is cheap to throw away, you are bolting a new motor onto old belts.</summary>
    <category term="cheap software"/>
    <category term="electrification"/>
    <category term="Paul David"/>
    <category term="line shaft"/>
    <category term="depreciation event"/>
    <category term="free disposal"/>
    <category term="organisational redesign"/>
    <category term="Technology"/>
  </entry>
  <entry>
    <title>Nobody Gets Fired</title>
    <link href="https://www.croftsware.com/blog/nobody-gets-fired/"/>
    <id>https://www.croftsware.com/blog/nobody-gets-fired/</id>
    <published>2026-07-15T00:00:00.000Z</published>
    <updated>2026-07-15T00:00:00.000Z</updated>
    <summary>Every founder believes the better product wins. In business-to-business it does not, and the reason is not that buyers are stupid. They are answering a different question from the one you are pitching. On the uncertainty tax, the safe incumbent, and why the challenger wins not by proving superiority but by making the change short enough to survive.</summary>
    <category term="B2B sales"/>
    <category term="positioning"/>
    <category term="behavioural economics"/>
    <category term="Rory Sutherland"/>
    <category term="Alchemy"/>
    <category term="Jonah Berger"/>
    <category term="Catalyst"/>
    <category term="Bent Flyvbjerg"/>
    <category term="How Big Things Get Done"/>
    <category term="uncertainty tax"/>
    <category term="challenger brand"/>
    <category term="buyer psychology"/>
    <category term="migration speed"/>
    <category term="project duration"/>
    <category term="software migration"/>
    <category term="digital transformation"/>
    <category term="no one ever got fired for buying IBM"/>
    <category term="SaaS Strategy"/>
  </entry>
  <entry>
    <title>The Memory That Won't Compress</title>
    <link href="https://www.croftsware.com/blog/the-memory-that-wont-compress/"/>
    <id>https://www.croftsware.com/blog/the-memory-that-wont-compress/</id>
    <published>2026-06-30T00:00:00.000Z</published>
    <updated>2026-06-30T00:00:00.000Z</updated>
    <summary>The vendors say agent memory is solved - extract the facts, embed them, retrieve on demand. But the institutional memory that actually stops an agent repeating your worst mistake is felt history, and felt history resists embedding by construction.</summary>

  </entry>
  <entry>
    <title>4HWW Was Right — We Just Needed Better Agents</title>
    <link href="https://www.croftsware.com/blog/4hww-was-right/"/>
    <id>https://www.croftsware.com/blog/4hww-was-right/</id>
    <published>2026-06-22T00:00:00.000Z</published>
    <updated>2026-06-22T00:00:00.000Z</updated>
    <summary>Tim Ferriss was right that most of your work can be automated — and wrong about what you would do with the freed time. On agents, the anti-hustle case, and the half of the job that was always the point.</summary>
    <category term="AI agents"/>
    <category term="four hour work week"/>
    <category term="Tim Ferriss"/>
    <category term="automation"/>
    <category term="one-person company"/>
    <category term="anti-hustle"/>
    <category term="human judgment"/>
    <category term="fractional CTO"/>
    <category term="AI &amp; Work"/>
  </entry>
  <entry>
    <title>Agents Give You Cadence</title>
    <link href="https://www.croftsware.com/blog/agents-give-you-cadence/"/>
    <id>https://www.croftsware.com/blog/agents-give-you-cadence/</id>
    <published>2026-05-21T00:00:00.000Z</published>
    <updated>2026-05-21T00:00:00.000Z</updated>
    <summary>The conventional sell of AI agents is intelligence, or speed. The thing they actually give you, for the kind of business I run, is cadence. Reliable, repeatable, undistracted rhythm. The boring half of work finally has a home.</summary>

  </entry>
  <entry>
    <title>The Non-Anxious CTO</title>
    <link href="https://www.croftsware.com/blog/the-non-anxious-cto/"/>
    <id>https://www.croftsware.com/blog/the-non-anxious-cto/</id>
    <published>2026-05-14T00:00:00.000Z</published>
    <updated>2026-05-14T00:00:00.000Z</updated>
    <summary>Every AI-for-CTOs piece in 2026 is about managing anxiety. The work I keep coming back to is older. Edwin Friedman, writing thirty years ago, told leaders to be the non-anxious presence in their systems. The infrastructure question turns out to be the embodied form.</summary>
    <category term="non-anxious presence"/>
    <category term="Edwin Friedman"/>
    <category term="differentiation"/>
    <category term="CTO leadership"/>
    <category term="AI strategy"/>
    <category term="organisational anxiety"/>
    <category term="family systems"/>
    <category term="Leadership"/>
  </entry>
  <entry>
    <title>The Memory You Own</title>
    <link href="https://www.croftsware.com/blog/the-memory-you-own/"/>
    <id>https://www.croftsware.com/blog/the-memory-you-own/</id>
    <published>2026-04-24T00:00:00.000Z</published>
    <updated>2026-04-24T00:00:00.000Z</updated>
    <summary>The second brain was a personal knowledge system for humans to browse. Version two has to be built for machines to query - and most of us are building it as a feature inside someone else's platform, which turns out to be the most insidious lock-in yet.</summary>

  </entry>
  <entry>
    <title>The Sum of All Tokens</title>
    <link href="https://www.croftsware.com/blog/the-sum-of-all-tokens/"/>
    <id>https://www.croftsware.com/blog/the-sum-of-all-tokens/</id>
    <published>2026-04-14T00:00:00.000Z</published>
    <updated>2026-04-14T00:00:00.000Z</updated>
    <summary>AI lets you replace paid software with custom tools that fit your problem. But the cost shifts to tokens plus human insight — and insight is the scarce half.</summary>
    <category term="AI tools"/>
    <category term="SaaS replacement"/>
    <category term="LLM economics"/>
    <category term="build vs buy"/>
    <category term="custom software"/>
    <category term="token cost"/>
    <category term="AI &amp; Software"/>
  </entry>
  <entry>
    <title>The Clean Interface</title>
    <link href="https://www.croftsware.com/blog/the-clean-interface/"/>
    <id>https://www.croftsware.com/blog/the-clean-interface/</id>
    <published>2026-04-13T00:00:00.000Z</published>
    <updated>2026-04-13T00:00:00.000Z</updated>
    <summary>A weekend of AI-assisted home energy optimisation revealed the inversion: the sophisticated parts of software are robust now; the simple parts are fragile.</summary>
    <category term="AI engineering"/>
    <category term="fragile interfaces"/>
    <category term="automation"/>
    <category term="system design"/>
    <category term="brittleness"/>
    <category term="vibe coding"/>
    <category term="AI &amp; Software"/>
  </entry>
  <entry>
    <title>The Long Obedience</title>
    <link href="https://www.croftsware.com/blog/the-long-obedience/"/>
    <id>https://www.croftsware.com/blog/the-long-obedience/</id>
    <published>2026-04-07T00:00:00.000Z</published>
    <updated>2026-04-07T00:00:00.000Z</updated>
    <summary>Why staying in one place for fifteen years turned out to be a competitive advantage - and what happens when AI meets patience.</summary>
    <category term="long-term focus"/>
    <category term="engineering culture"/>
    <category term="AI patience"/>
    <category term="sheepCRM"/>
    <category term="expertise"/>
    <category term="founder lessons"/>
    <category term="Engineering Leadership"/>
  </entry>
  <entry>
    <title>Fifteen Years of sheepCRM</title>
    <link href="https://www.croftsware.com/blog/fifteen-years-of-sheepcrm/"/>
    <id>https://www.croftsware.com/blog/fifteen-years-of-sheepcrm/</id>
    <published>2026-03-31T00:00:00.000Z</published>
    <updated>2026-03-31T00:00:00.000Z</updated>
    <summary>How a favour for a friend became a membership CRM serving over a hundred organisations. The sheepCRM story — from Access replacement to focused product.</summary>
    <category term="SaaS"/>
    <category term="membership CRM"/>
    <category term="founder story"/>
    <category term="product focus"/>
    <category term="niche positioning"/>
    <category term="sheepCRM"/>
    <category term="fifteen years"/>
    <category term="SaaS Strategy"/>
  </entry>
  <entry>
    <title>The End of Simplicity</title>
    <link href="https://www.croftsware.com/blog/the-end-of-simplicity/"/>
    <id>https://www.croftsware.com/blog/the-end-of-simplicity/</id>
    <published>2026-02-10T00:00:00.000Z</published>
    <updated>2026-02-10T00:00:00.000Z</updated>
    <summary>When AI collapses the cost of building features and agents replace human users, what happens to our instinct for simplicity? On abundance and discipline.</summary>
    <category term="AI agents"/>
    <category term="software complexity"/>
    <category term="simplicity"/>
    <category term="fractional CTO"/>
    <category term="product design"/>
    <category term="feature bloat"/>
    <category term="AI &amp; Software"/>
  </entry>
  <entry>
    <title>The Spreadsheet Moment</title>
    <link href="https://www.croftsware.com/blog/the-spreadsheet-moment/"/>
    <id>https://www.croftsware.com/blog/the-spreadsheet-moment/</id>
    <published>2026-01-15T00:00:00.000Z</published>
    <updated>2026-01-15T00:00:00.000Z</updated>
    <summary>AI code generation is doing for software what VisiCalc did for financial modelling — putting creation into the hands of people who understand problems.</summary>
    <category term="AI code generation"/>
    <category term="software development"/>
    <category term="vibe coding"/>
    <category term="LLM"/>
    <category term="productivity"/>
    <category term="VisiCalc"/>
    <category term="no-code"/>
    <category term="AI &amp; Software"/>
  </entry>
  <entry>
    <title>The Three Mountains</title>
    <link href="https://www.croftsware.com/blog/the-three-mountains/"/>
    <id>https://www.croftsware.com/blog/the-three-mountains/</id>
    <published>2026-01-06T00:00:00.000Z</published>
    <updated>2026-01-06T00:00:00.000Z</updated>
    <summary>A simple model for operational maturity — where you actually are, not where you think you are. From a primary school framework to a decade of consulting.</summary>
    <category term="operational maturity"/>
    <category term="organisational change"/>
    <category term="data strategy"/>
    <category term="change management"/>
    <category term="technology leadership"/>
    <category term="fractional CTO"/>
    <category term="Operational Maturity"/>
  </entry>
  <entry>
    <title>The Accidental Operating System</title>
    <link href="https://www.croftsware.com/blog/the-accidental-operating-system/"/>
    <id>https://www.croftsware.com/blog/the-accidental-operating-system/</id>
    <published>2025-11-18T00:00:00.000Z</published>
    <updated>2025-11-18T00:00:00.000Z</updated>
    <summary>Holacracy was criticised as too rigid for humans. It might have been building the organisational architecture AI agents need.</summary>
    <category term="Holacracy"/>
    <category term="AI agents"/>
    <category term="organisational design"/>
    <category term="governance"/>
    <category term="structure"/>
    <category term="agent architecture"/>
    <category term="Organisational Design"/>
  </entry>
</feed>
