<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"><channel><title>Sukarma Insights</title><link>https://sukarma.uk/insights/</link><description>Perspectives on AI and public services from Sukarma.</description>
<item><title>The story behind our name and our symbol</title><link>https://sukarma.uk/insights/ai-without-fear/</link><guid>https://sukarma.uk/insights/ai-without-fear/</guid><pubDate>Wed, 18 Nov 2026 09:00:00 GMT</pubDate><description>Why we are called Sukarma, why our symbol is a ripple, and the five promises behind them.</description></item>
<item><title>Introducing Sukarma</title><link>https://sukarma.uk/insights/out-of-stealth/</link><guid>https://sukarma.uk/insights/out-of-stealth/</guid><pubDate>Tue, 17 Nov 2026 09:00:00 GMT</pubDate><description>Three years ago we set out to build a company for the moment AI started writing software. That moment has arrived. Here is what we built while we were quiet, and why we are focusing on public services.</description></item>
<item><title>This week in AI: telling real from generated</title><link>https://sukarma.uk/insights/week-telling-real-from-generated/</link><guid>https://sukarma.uk/insights/week-telling-real-from-generated/</guid><pubDate>Fri, 13 Nov 2026 09:00:00 GMT</pubDate><description>Google has expanded its SynthID Detector worldwide and the EU's transparency rules for AI-generated content are coming into effect. Public bodies should be clear about where they use AI.</description></item>
<item><title>We don't build models. We build with them.</title><link>https://sukarma.uk/insights/we-dont-build-models/</link><guid>https://sukarma.uk/insights/we-dont-build-models/</guid><pubDate>Tue, 10 Nov 2026 09:00:00 GMT</pubDate><description>Foundation models are becoming the most powerful general-purpose infrastructure ever built. The winning strategy for almost every organisation is to use them brilliantly, not to compete with them.</description></item>
<item><title>This week in AI: evidence from health care</title><link>https://sukarma.uk/insights/week-ai-in-health-evidence/</link><guid>https://sukarma.uk/insights/week-ai-in-health-evidence/</guid><pubDate>Fri, 06 Nov 2026 09:00:00 GMT</pubDate><description>A study of a medical AI in primary care was published in The Lancet, labs reported progress on drug development, and the UK commissioned a review of NHS productivity and AI. Health is where AI evidence is maturing fastest.</description></item>
<item><title>Prototype before you procure</title><link>https://sukarma.uk/insights/prototype-before-you-procure/</link><guid>https://sukarma.uk/insights/prototype-before-you-procure/</guid><pubDate>Tue, 03 Nov 2026 09:00:00 GMT</pubDate><description>AI has collapsed the cost of building working software. Public buyers can now ask suppliers to show, not tell, and suppliers can afford to.</description></item>
<item><title>This week in AI: what the Budget means for public-sector AI</title><link>https://sukarma.uk/insights/week-budget-ai-package/</link><guid>https://sukarma.uk/insights/week-budget-ai-package/</guid><pubDate>Fri, 30 Oct 2026 09:00:00 GMT</pubDate><description>The Autumn Budget on 28 October included a package of AI measures, from more funding for the AI Security Institute to local AI partnerships with mayoral authorities. Here are the parts most relevant to public services.</description></item>
<item><title>This week in AI: open standards for agents</title><link>https://sukarma.uk/insights/week-open-standards-for-agents/</link><guid>https://sukarma.uk/insights/week-open-standards-for-agents/</guid><pubDate>Fri, 23 Oct 2026 09:00:00 GMT</pubDate><description>Anthropic has donated Agent Skills to the Agentic AI Foundation, joining the Model Context Protocol. Shared standards make it easier to switch tools and keep what you have built.</description></item>
<item><title>Agentic delivery, explained without the hype</title><link>https://sukarma.uk/insights/agentic-delivery-explained/</link><guid>https://sukarma.uk/insights/agentic-delivery-explained/</guid><pubDate>Tue, 20 Oct 2026 09:00:00 GMT</pubDate><description>Agentic is the word of the moment. Here is what it actually means in our delivery practice, where it helps, and where humans remain firmly in charge.</description></item>
<item><title>This week in AI: AI on the defenders' side</title><link>https://sukarma.uk/insights/week-ai-and-cyber-defence/</link><guid>https://sukarma.uk/insights/week-ai-and-cyber-defence/</guid><pubDate>Fri, 16 Oct 2026 09:00:00 GMT</pubDate><description>Anthropic launched a cyber mission and widened its verification programme, and Google is testing a new model for cyber defence. For most organisations, the basics still matter most.</description></item>
<item><title>This week in AI: small models do most of the work</title><link>https://sukarma.uk/insights/week-small-models-matter/</link><guid>https://sukarma.uk/insights/week-small-models-matter/</guid><pubDate>Fri, 09 Oct 2026 09:00:00 GMT</pubDate><description>Anthropic released Claude Haiku 5.5 and OpenAI's GPT-6 reached all ChatGPT users. The flagships get the headlines, but small, fast models handle most everyday tasks.</description></item>
<item><title>Accessibility is not a phase</title><link>https://sukarma.uk/insights/accessibility-is-not-a-phase/</link><guid>https://sukarma.uk/insights/accessibility-is-not-a-phase/</guid><pubDate>Tue, 06 Oct 2026 09:00:00 GMT</pubDate><description>Public sector websites and apps must meet accessibility regulations. With AI-assisted testing there is no longer any excuse for treating accessibility as a remediation project.</description></item>
<item><title>This week in AI: skills, not just models</title><link>https://sukarma.uk/insights/week-skills-not-just-models/</link><guid>https://sukarma.uk/insights/week-skills-not-just-models/</guid><pubDate>Fri, 02 Oct 2026 09:00:00 GMT</pubDate><description>Anthropic is putting $100 million into training 10,000 engineers, and Barclays is scaling Claude across its operations. The bottleneck for AI is increasingly people's skills and confidence.</description></item>
<item><title>This week in AI: new flagships, released carefully</title><link>https://sukarma.uk/insights/week-careful-releases/</link><guid>https://sukarma.uk/insights/week-careful-releases/</guid><pubDate>Fri, 25 Sep 2026 09:00:00 GMT</pubDate><description>Anthropic released Claude Opus 5.5 and Google introduced Gemini 4 Argon, which went first to trusted cyber defenders. Staged releases are becoming normal for the most capable models.</description></item>
<item><title>Reading the UK Government AI Playbook as a supplier</title><link>https://sukarma.uk/insights/ai-playbook-for-suppliers/</link><guid>https://sukarma.uk/insights/ai-playbook-for-suppliers/</guid><pubDate>Tue, 22 Sep 2026 09:00:00 GMT</pubDate><description>The AI Playbook for the UK Government sets out ten principles for safe and effective AI. Here is how we turn each one into something concrete in our delivery.</description></item>
<item><title>This week in AI: new model numbers, same question</title><link>https://sukarma.uk/insights/week-new-model-numbers/</link><guid>https://sukarma.uk/insights/week-new-model-numbers/</guid><pubDate>Fri, 18 Sep 2026 09:00:00 GMT</pubDate><description>OpenAI has started rolling out its GPT-6 models and Google has new real-time voice models. A bigger version number is not a reason to switch. Your own tests are.</description></item>
<item><title>This week in AI: planning for economic change</title><link>https://sukarma.uk/insights/week-planning-for-economic-change/</link><guid>https://sukarma.uk/insights/week-planning-for-economic-change/</guid><pubDate>Fri, 11 Sep 2026 09:00:00 GMT</pubDate><description>Anthropic published scenarios for how AI could reshape the economy. For employers, including public bodies, the useful response is practical: redesign jobs, invest in skills and look after people starting out.</description></item>
<item><title>The Procurement Act 2023 is an opportunity for AI-native SMEs</title><link>https://sukarma.uk/insights/procurement-act-sme-opportunity/</link><guid>https://sukarma.uk/insights/procurement-act-sme-opportunity/</guid><pubDate>Tue, 08 Sep 2026 09:00:00 GMT</pubDate><description>The new procurement regime, live since February 2025, is designed to open public contracts to smaller and more innovative suppliers. Here is what that means for buyers who want AI delivered well.</description></item>
<item><title>This week in AI: back to school, and the trouble with detectors</title><link>https://sukarma.uk/insights/week-back-to-school/</link><guid>https://sukarma.uk/insights/week-back-to-school/</guid><pubDate>Fri, 04 Sep 2026 09:00:00 GMT</pubDate><description>As schools, colleges and universities return, the debate about AI and assessment continues. The evidence points away from detection tools and towards better-designed tasks.</description></item>
<item><title>This week in AI: putting models in scientists' hands</title><link>https://sukarma.uk/insights/week-ai-for-science/</link><guid>https://sukarma.uk/insights/week-ai-for-science/</guid><pubDate>Fri, 28 Aug 2026 09:00:00 GMT</pubDate><description>Anthropic is giving 10,000 scientists free access to Claude, following grants for rare-disease research. Programmes like these are where AI's long-term value may show up first.</description></item>
<item><title>This week in AI: agents need fences</title><link>https://sukarma.uk/insights/week-agents-need-fences/</link><guid>https://sukarma.uk/insights/week-agents-need-fences/</guid><pubDate>Fri, 21 Aug 2026 09:00:00 GMT</pubDate><description>Recent reports of AI agents reaching systems they were never meant to touch are a useful reminder. The fix is old-fashioned: least privilege, sandboxes and a person approving anything that matters.</description></item>
<item><title>This week in AI: capable models get cheaper</title><link>https://sukarma.uk/insights/week-capable-models-get-cheaper/</link><guid>https://sukarma.uk/insights/week-capable-models-get-cheaper/</guid><pubDate>Fri, 14 Aug 2026 09:00:00 GMT</pubDate><description>New model tiers from Anthropic and OpenAI bring near-flagship ability at lower prices, and let you choose how much thinking a task gets. That changes the sums for public services.</description></item>
<item><title>This week in AI: the EU AI Act deadline that moved</title><link>https://sukarma.uk/insights/week-eu-ai-act-deadline-moves/</link><guid>https://sukarma.uk/insights/week-eu-ai-act-deadline-moves/</guid><pubDate>Fri, 07 Aug 2026 09:00:00 GMT</pubDate><description>The high-risk rules in the EU AI Act were due to apply on 2 August. A late change pushed most of them to December 2027. Here is what changed and why we are carrying on regardless.</description></item>
<item><title>A quarter of big releases, and a pause for EU rules</title><link>https://sukarma.uk/insights/a-quarter-of-big-releases/</link><guid>https://sukarma.uk/insights/a-quarter-of-big-releases/</guid><pubDate>Wed, 24 Jun 2026 09:00:00 GMT</pubDate><description>GPT-5.5, Claude Fable 5 and Gemini 3.5 arrived, the UK published a compute roadmap, and the EU moved to delay its high-risk AI rules.</description></item>
<item><title>From strategy to delivery</title><link>https://sukarma.uk/insights/from-strategy-to-delivery/</link><guid>https://sukarma.uk/insights/from-strategy-to-delivery/</guid><pubDate>Thu, 19 Mar 2026 09:00:00 GMT</pubDate><description>The UK government reported on a year of its AI action plan and started piloting AI assistants in public services, while India hosted the AI Impact Summit.</description></item>
<item><title>Agents get skills, and open standards get a home</title><link>https://sukarma.uk/insights/agents-get-skills-and-standards/</link><guid>https://sukarma.uk/insights/agents-get-skills-and-standards/</guid><pubDate>Tue, 16 Dec 2025 09:00:00 GMT</pubDate><description>Gemini 3 and Claude Opus 4.5 arrived, Agent Skills let teams package know-how for AI, and the Model Context Protocol moved to a neutral foundation.</description></item>
<item><title>GPT-5, an Olympiad milestone and the first EU rules for model makers</title><link>https://sukarma.uk/insights/gpt-5-and-the-maths-olympiad/</link><guid>https://sukarma.uk/insights/gpt-5-and-the-maths-olympiad/</guid><pubDate>Tue, 30 Sep 2025 09:00:00 GMT</pubDate><description>Models reached gold-medal standard at the International Mathematical Olympiad, OpenAI released GPT-5 and open-weight models, and EU obligations for model providers began.</description></item>
<item><title>Coding agents go mainstream</title><link>https://sukarma.uk/insights/coding-agents-go-mainstream/</link><guid>https://sukarma.uk/insights/coding-agents-go-mainstream/</guid><pubDate>Wed, 18 Jun 2025 09:00:00 GMT</pubDate><description>Claude 4, OpenAI's o3 and Codex, and Google's I/O announcements made AI agents that carry out whole tasks part of everyday software work.</description></item>
<item><title>The Procurement Act arrives, and government publishes its AI Playbook</title><link>https://sukarma.uk/insights/the-procurement-act-arrives/</link><guid>https://sukarma.uk/insights/the-procurement-act-arrives/</guid><pubDate>Thu, 27 Mar 2025 09:00:00 GMT</pubDate><description>A big quarter for UK public-sector AI. The AI Opportunities Action Plan, the AI Playbook and the new procurement regime all landed within weeks of each other.</description></item>
<item><title>A common way to connect AI to everything else</title><link>https://sukarma.uk/insights/a-common-way-to-connect-ai-to-systems/</link><guid>https://sukarma.uk/insights/a-common-way-to-connect-ai-to-systems/</guid><pubDate>Tue, 17 Dec 2024 09:00:00 GMT</pubDate><description>AI researchers won Nobel Prizes, the Model Context Protocol offered an open standard for connecting models to data, and the year ended with a run of new releases.</description></item>
<item><title>Models that think before they answer</title><link>https://sukarma.uk/insights/models-that-think-before-answering/</link><guid>https://sukarma.uk/insights/models-that-think-before-answering/</guid><pubDate>Tue, 24 Sep 2024 09:00:00 GMT</pubDate><description>OpenAI's o1 introduced models that reason step by step, the EU AI Act came into force, and the Procurement Act got a new start date.</description></item>
<item><title>Models learn to listen and speak</title><link>https://sukarma.uk/insights/models-learn-to-listen/</link><guid>https://sukarma.uk/insights/models-learn-to-listen/</guid><pubDate>Wed, 26 Jun 2024 09:00:00 GMT</pubDate><description>GPT-4o made real-time voice conversation practical, Claude 3.5 Sonnet raised the bar for coding, and governments met again in Seoul.</description></item>
<item><title>Government guidance catches up, and models read whole archives</title><link>https://sukarma.uk/insights/government-guidance-catches-up/</link><guid>https://sukarma.uk/insights/government-guidance-catches-up/</guid><pubDate>Thu, 21 Mar 2024 09:00:00 GMT</pubDate><description>The UK published its first framework for generative AI in government, models gained million-token memories, and the European Parliament adopted the AI Act.</description></item>
<item><title>Bletchley, bigger context windows and a busy autumn</title><link>https://sukarma.uk/insights/bletchley-and-a-busy-autumn/</link><guid>https://sukarma.uk/insights/bletchley-and-a-busy-autumn/</guid><pubDate>Thu, 14 Dec 2023 09:00:00 GMT</pubDate><description>The UK hosted the first AI Safety Summit, models learned to read much longer documents, and Europe reached a deal on the AI Act. Here is what it means for the work we do.</description></item>
<item><title>Why we think AI will change how software is built</title><link>https://sukarma.uk/insights/why-ai-will-change-how-software-is-built/</link><guid>https://sukarma.uk/insights/why-ai-will-change-how-software-is-built/</guid><pubDate>Tue, 19 Sep 2023 09:00:00 GMT</pubDate><description>Our first note as a new company. Language models can already write useful code. The open question is how organisations that answer to auditors can trust what they produce.</description></item>
</channel></rss>
