<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Adil Soukaini | Full-Stack Developer on Medium]]></title>
        <description><![CDATA[Stories by Adil Soukaini | Full-Stack Developer on Medium]]></description>
        <link>https://medium.com/@adilsoukaini?source=rss-7f614efc1b10------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*wKG1xqjVGFVLLgN2IYFGXA.jpeg</url>
            <title>Stories by Adil Soukaini | Full-Stack Developer on Medium</title>
            <link>https://medium.com/@adilsoukaini?source=rss-7f614efc1b10------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 20:58:15 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@adilsoukaini/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Code ownership and IP when outsourcing software development]]></title>
            <link>https://medium.com/@adilsoukaini/code-ownership-and-ip-when-outsourcing-software-development-e67bceccae5c?source=rss-7f614efc1b10------2</link>
            <guid isPermaLink="false">https://medium.com/p/e67bceccae5c</guid>
            <category><![CDATA[intellectual-property]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[outsourcing]]></category>
            <dc:creator><![CDATA[Adil Soukaini | Full-Stack Developer]]></dc:creator>
            <pubDate>Wed, 04 Nov 2026 08:00:19 GMT</pubDate>
            <atom:updated>2026-11-11T16:10:39.222Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*89LGkwUK3m8kdqL6.png" /></figure><p>Paying someone to write your code does not, by itself, make you the owner. In the US, the UK and France, copyright in software written by a contractor stays with the contractor unless a written contract transfers it to you; only software written by employees in their job belongs to the employer by default. So when you outsource, you need three things: a written assignment of rights, the code in a repository your company controls, and a check of the open-source licences inside it.</p><p><em>This article is general information, not legal advice. Rules differ by country and by contract; have a lawyer review agreements that matter to your business.</em></p><h3>Who owns the code by default</h3><h3>United States</h3><p>The US Copyright Office explains that a “work made for hire” belongs to the employer or commissioning party. It covers <a href="https://www.copyright.gov/circs/circ30.pdf">a work prepared by an employee within the scope of employment</a>, or a specially commissioned work, but only if it falls into one of nine listed categories (such as a contribution to a collective work, a translation or a compilation) and both parties sign a written agreement. Software is not one of the nine categories. In practice, code written by an independent contractor stays with the contractor unless they assign the copyright to you in writing.</p><h3>United Kingdom</h3><p>The UK Intellectual Property Office is blunt: when you commission a work, <a href="https://www.gov.uk/guidance/ownership-of-copyright-works">“the first legal owner of copyright is the person or organisation that created the work and not you the commissioner, unless you otherwise agree it in writing”</a>. Employees are different: their employer is the first owner of work made in the course of employment.</p><h3>France</h3><p>Article L113–9 of the French Intellectual Property Code gives the employer the economic rights in <a href="https://www.legifrance.gouv.fr/codes/article_lc/LEGIARTI000006278890">software and its documentation created by employees in the course of their duties or following the employer’s instructions</a>. That rule does not cover freelancers or agencies. As Bpifrance Création puts it, <a href="https://bpifrance-creation.fr/encyclopedie/trouver-proteger-tester-son-idee/proteger-son-idee/protection-juridique-dun-logiciel">ordering specific software from a provider does not automatically make the client the rights holder; a rights-assignment contract is needed</a>. And French law is strict about the form of that contract: under Article L131–3, <a href="https://www.legifrance.gouv.fr/codes/article_lc/LEGIARTI000006278958">each right assigned must be mentioned separately, and the scope of use must be defined by extent, purpose, place and duration</a>. A one-line “all rights belong to the client” clause may not be enough.</p><h3>What the contract should cover</h3><ol><li><strong>Assignment, not just a licence.</strong> State that the provider assigns to you the rights in all code, documentation and designs created for you. Under French law, list the rights (reproduction, modification, adaptation, distribution, commercial exploitation) and define territory and duration.</li><li><strong>When it happens.</strong> Usually on payment of each invoice. That is fair to the provider, and it means you should pay on time and keep the paperwork.</li><li><strong>Pre-existing material.</strong> Providers reuse their own libraries, templates and tools. You do not need to own those, but you need a permanent, transferable licence to use them in your product.</li><li><strong>Subcontractors.</strong> If the provider uses other developers, they must have assigned their rights to the provider first, or the chain breaks.</li><li><strong>Third-party and open-source components.</strong> The provider should list them with their licences and confirm that they are compatible with how you use the product.</li><li><strong>Deliverables at the end.</strong> Full source code, build and deploy instructions, documentation, and the credentials for every account they created on your behalf.</li><li><strong>Confidentiality and data.</strong> Your code and data stay confidential; personal data is handled under a processor agreement (see <a href="https://shipfastdev.com/blog/how-to-choose-a-remote-development-partner">how to choose a remote development partner</a>).</li></ol><h3>Own the repository, not just the rights</h3><p>A contract gives you the right to the code; the repository gives you the code itself. Create the GitHub, GitLab or Bitbucket organisation under your company from day one and invite the provider. If a repository already lives in the provider’s account, ask for a transfer: on GitHub, <a href="https://docs.github.com/en/repositories/creating-and-managing-repositories/transferring-a-repository">issues, pull requests, wiki, stars and watchers are transferred with it</a>, and links to the old location redirect automatically.</p><p>Keep in mind what an unlicensed repository means. GitHub notes that <a href="https://docs.github.com/en/repositories/managing-your-repositorys-settings-and-features/customizing-your-repository/licensing-a-repository">without a licence, default copyright law applies and nobody may reproduce, distribute or create derivative works</a>. That is what you want for a private product. If you ever open-source part of it, choose a licence deliberately.</p><p>Bpifrance Création also recommends keeping proof of the date and content of your code, for example a deposit with the Agence pour la protection des programmes (APP) or an INPI e-Soleau envelope. A dated, company-owned Git history helps as well. Neither creates rights; they help you prove them.</p><h3>Open-source licences: what to check</h3><p>Almost every product depends on open-source packages, and most licences (MIT, BSD, Apache 2.0) ask little more than keeping copyright notices. Copyleft licences need more attention. According to the Free Software Foundation’s <a href="https://www.gnu.org/licenses/gpl-faq.html">GPL FAQ</a>:</p><ul><li>You may modify GPL software and use it privately without releasing your changes; the obligation to share source code applies when you distribute it.</li><li>The GNU Affero GPL (AGPL) goes further: if you run a modified version on a server, users interacting with it over a network must be able to get the source.</li><li>If your program links to a library under the GPL (not the LGPL), the work as a whole must be licensed under the GPL.</li></ul><p>For a SaaS product, AGPL components deserve a close look; for a mobile or desktop app you distribute, GPL ones do. Ask your developer for a dependency list with licences before each major release; package managers can produce one (for example composer licenses for PHP projects).</p><h3>A short checklist</h3><ul><li>Signed contract with an assignment of rights (French formalities if French law applies).</li><li>Licence for the provider’s pre-existing tools.</li><li>Repository, hosting and accounts owned by your company.</li><li>Dependency list with licences, no unexpected GPL or AGPL.</li><li>Handover clause: source, documentation, credentials.</li></ul><p>The same points apply whatever the model; see <a href="https://shipfastdev.com/blog/freelancer-vs-agency-vs-dev-retainer">freelancer vs agency vs dev retainer</a> and the access plan in onboarding a remote developer.</p><h3>Frequently asked questions</h3><h3>Who owns the code written by a freelance developer?</h3><p>By default in the US, the UK and France, the freelancer does. You become the owner only through a written assignment of rights. Employees are different: their employer generally owns the code they write in their job.</p><h3>Is “work for hire” enough for software in the US?</h3><p>Usually not for contractors. Software is not among the nine categories of commissioned works that can be “made for hire”, so contracts with contractors should include an explicit copyright assignment.</p><h3>Can I use GPL code in a commercial product?</h3><p>Yes, but if you distribute the product, the GPL requires you to provide the source code under the GPL. AGPL extends that to software used over a network. Check licences before you ship.</p><h3>What should I get from a developer at the end of a project?</h3><p>The full source code in your repository, documentation, build and deployment instructions, credentials for every account created for you, and confirmation that rights have been assigned.</p><h3>Sources</h3><ol><li><a href="https://www.copyright.gov/circs/circ30.pdf">Works Made for Hire (Circular 30)</a>, U.S. Copyright Office, 2024–08</li><li><a href="https://www.gov.uk/guidance/ownership-of-copyright-works">Ownership of copyright works</a>, Intellectual Property Office, GOV.UK, 2014–08</li><li><a href="https://www.legifrance.gouv.fr/codes/article_lc/LEGIARTI000006278890">Code de la propriété intellectuelle — Article L113–9</a>, Légifrance, n.d.</li><li><a href="https://www.legifrance.gouv.fr/codes/article_lc/LEGIARTI000006278958">Code de la propriété intellectuelle — Article L131–3</a>, Légifrance, n.d.</li><li><a href="https://bpifrance-creation.fr/encyclopedie/trouver-proteger-tester-son-idee/proteger-son-idee/protection-juridique-dun-logiciel">Protection juridique d’un logiciel</a>, Bpifrance Création, 2026–06</li><li><a href="https://www.gnu.org/licenses/gpl-faq.html">Frequently Asked Questions about the GNU Licenses</a>, Free Software Foundation (gnu.org), n.d.</li><li><a href="https://docs.github.com/en/repositories/managing-your-repositorys-settings-and-features/customizing-your-repository/licensing-a-repository">Licensing a repository</a>, GitHub Docs, n.d.</li><li><a href="https://docs.github.com/en/repositories/creating-and-managing-repositories/transferring-a-repository">Transferring a repository</a>, GitHub Docs, n.d.</li></ol><p><em>Originally published at </em><a href="https://shipfastdev.com/blog/code-ownership-and-ip-when-outsourcing"><em>https://shipfastdev.com</em></a><em> on November 4, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e67bceccae5c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[React Native vs Flutter in 2026: which one should your startup use?]]></title>
            <link>https://medium.com/@adilsoukaini/react-native-vs-flutter-in-2026-which-one-should-your-startup-use-e3cf16ac954b?source=rss-7f614efc1b10------2</link>
            <guid isPermaLink="false">https://medium.com/p/e3cf16ac954b</guid>
            <category><![CDATA[react-native]]></category>
            <category><![CDATA[mobile-app-development]]></category>
            <category><![CDATA[flutter]]></category>
            <category><![CDATA[startup]]></category>
            <dc:creator><![CDATA[Adil Soukaini | Full-Stack Developer]]></dc:creator>
            <pubDate>Mon, 02 Nov 2026 08:00:47 GMT</pubDate>
            <atom:updated>2026-11-11T16:06:18.595Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*uAdD44xi4ktX0Vn8.png" /></figure><p>In 2026, React Native (with Expo) and Flutter are both mature, production-ready ways to ship one app to iOS and Android. Choose React Native with Expo if your team already writes React and TypeScript, or if you want to share people and code with a web app. Choose Flutter if you want a highly custom interface that looks identical on every platform and your team is comfortable learning Dart.</p><p>Below: what the adoption data says, how they differ under the hood, hiring, and a short decision guide.</p><h3>Adoption: close to a tie among developers</h3><p>The 2024 Stack Overflow Developer Survey asked about both. Among respondents to the “other frameworks and libraries” question, <a href="https://survey.stackoverflow.co/2024/technology">9.4% used Flutter and 8.4% used React Native</a>; among professional developers it was 9.4% and 9.0%. Neither is a niche choice, and neither dominates.</p><p>The bigger gap is in the languages behind them. In the <a href="https://survey.stackoverflow.co/2025/technology">2025 survey</a>, 66% of respondents used JavaScript and 43.6% TypeScript, against 5.9% for Dart, the language Flutter uses. That matters more for hiring than framework popularity, as we’ll see.</p><h3>How they work, and why it matters for performance</h3><p>JetBrains’ overview of cross-platform frameworks summarises the core difference well:</p><ul><li><strong>Flutter</strong>: “has its own rendering engine for drawing widgets”. It doesn’t use the platform’s native buttons and lists; it draws its own.</li><li><strong>React Native</strong>: “React primitives render to native platform UI components”. A button is a real iOS or Android button.</li></ul><p>Consequences for you:</p><ul><li>Flutter apps look the same everywhere by default, which is good for strong custom branding and less good if you want each platform’s native feel.</li><li>React Native apps pick up platform conventions (scrolling, text inputs, accessibility behaviour) because they use native components.</li></ul><p>Both teams have invested heavily in performance. Flutter’s newer renderer, Impeller, <a href="https://docs.flutter.dev/perf/impeller">precompiles shaders at engine-build time</a> rather than at runtime; it is the only renderer on iOS and is enabled by default on Android API 29 and above. React Native’s New Architecture, a rewrite of its internals, has been <a href="https://reactnative.dev/blog/2024/10/23/release-0.76-new-architecture">enabled by default since React Native 0.76</a> (October 2024).</p><p>For the kind of app most startups build (forms, lists, maps, payments, notifications) either framework is fast enough. Performance problems usually come from how the app is built: heavy screens, unoptimised images, too much work on the main thread. Games and graphics-heavy apps are a separate discussion.</p><h3>Ecosystem and tooling</h3><h3>React Native and Expo</h3><p>The React Native documentation now says: <a href="https://reactnative.dev/docs/environment-setup">“if you’re building a new app with React Native, we recommend using a Framework”</a>, and names Expo as a production-grade framework with file-based routing and a standard library of native modules. In practice, “React Native in 2026” mostly means “Expo”.</p><p>Expo’s most useful feature for a startup is EAS Update: you can push changes to the JavaScript, styling and images <a href="https://docs.expo.dev/eas-update/introduction/">over the air</a>, without waiting for a store review. The limit is clear in the docs: anything that changes native code, permissions or the Expo SDK version still needs a new build.</p><p>The rest of the ecosystem is npm: most JavaScript libraries that don’t depend on the browser work, and a React web developer can be productive quickly.</p><h3>Flutter</h3><p>Flutter gives you a single toolkit with a large catalogue of built-in widgets, its own package repository and strong tooling. Because it draws everything itself, a design looks the same on every device you test, which reduces platform-specific UI bugs. The trade-off is that integrating platform features without an existing package means writing platform code, as with React Native.</p><h3>Hiring and team</h3><p>This is where the two differ most for a founder:</p><ul><li><strong>React Native</strong> draws on the JavaScript and TypeScript pool, the largest in the survey. A developer who knows React for the web can move to React Native in weeks, and one person can work across web and mobile.</li><li><strong>Flutter</strong> needs Dart. Dart tends to feel familiar to developers with Java, Kotlin, Swift or C# experience, but fewer developers already use it (5.9% in 2025, against 66% for JavaScript).</li></ul><p>If your backend is Node.js or your web frontend is React, React Native keeps everything in one language. If your backend is Laravel, the language argument is neutral; see <a href="https://shipfastdev.com/blog/laravel-vs-nodejs-for-startups">Laravel vs Node.js for startups</a>.</p><h3>From our work: an Expo app beside a Laravel web platform</h3><p>The <a href="https://shipfastdev.com/work/driveway-morocco">Driveway Morocco</a> mobile app is built with Expo and React Native and talks to the same Laravel backend as the web platform, through a JSON API with token authentication. The project’s CI uses EAS Build to produce a production Android App Bundle on each push to the main branch, and an EAS Submit step sends it to the Play Store’s internal testing track. The steps to get from there to a public listing are in our <a href="https://shipfastdev.com/blog/publish-expo-app-google-play-checklist">Google Play checklist for Expo apps</a>.</p><p>The reason for choosing Expo there was practical: the web platform was already React, so screens, validation rules and API types followed the same patterns on both sides.</p><h3>Decision guide</h3><p><strong>Choose React Native with Expo if:</strong></p><ul><li>your team or partner writes React and TypeScript;</li><li>you have, or will have, a React web app;</li><li>you want to fix bugs and change copy without a store review (EAS Update);</li><li>you want the platform’s native look and behaviour.</li></ul><p><strong>Choose Flutter if:</strong></p><ul><li>the interface is highly custom and must look identical on all platforms;</li><li>your developers already know Dart, or come from Kotlin, Swift or Java;</li><li>you have no web app to share code or people with.</li></ul><p><strong>Consider fully native (Swift and Kotlin) if</strong> the app depends on platform features at the edge of what cross-platform tools support, and you can fund two codebases.</p><p>Platform choice is one of the largest cost drivers in an app budget; see <a href="https://shipfastdev.com/blog/how-much-does-an-mvp-cost">how much an MVP costs</a>.</p><h3>Frequently asked questions</h3><h3>Is React Native or Flutter better in 2026?</h3><p>Neither is better in general. Usage is close (9.4% Flutter, 8.4% React Native in the 2024 Stack Overflow survey). Pick the one that matches your team’s language and your design goals.</p><h3>Is Flutter faster than React Native?</h3><p>Flutter draws its own UI with the Impeller renderer; React Native renders native components and has used its New Architecture by default since version 0.76. For typical business apps both are fast enough; how the app is built matters more.</p><h3>Can I update a React Native app without the App Store?</h3><p>With Expo’s EAS Update you can ship JavaScript, styling and image changes over the air. Changes to native code, permissions or the SDK version still need a new store build.</p><h3>Can one codebase cover web, iOS and Android?</h3><p>Both frameworks can target the web as well, but most startups keep a dedicated web app and share the backend, business rules and, with React, the developers.</p><h3>Sources</h3><ol><li><a href="https://survey.stackoverflow.co/2024/technology">2024 Stack Overflow Developer Survey: Technology</a>, Stack Overflow, July 2024</li><li><a href="https://survey.stackoverflow.co/2025/technology">2025 Stack Overflow Developer Survey: Technology</a>, Stack Overflow, July 2025</li><li><a href="https://kotlinlang.org/docs/multiplatform/cross-platform-frameworks.html">Popular cross-platform app development frameworks</a>, JetBrains / Kotlin documentation, June 2026</li><li><a href="https://reactnative.dev/blog/2024/10/23/release-0.76-new-architecture">React Native 0.76: New Architecture by default</a>, React Native blog, October 2024</li><li><a href="https://reactnative.dev/docs/environment-setup">Get Started with React Native</a>, React Native documentation, accessed October 2026</li><li><a href="https://docs.expo.dev/eas-update/introduction/">EAS Update: introduction</a>, Expo documentation, accessed October 2026</li><li><a href="https://docs.flutter.dev/perf/impeller">Impeller rendering engine</a>, Flutter documentation, accessed October 2026</li></ol><p><em>Originally published at </em><a href="https://shipfastdev.com/blog/react-native-vs-flutter"><em>https://shipfastdev.com</em></a><em> on November 2, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e3cf16ac954b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Security Risks of AI-Generated Code: Research and a Review Checklist]]></title>
            <link>https://medium.com/@adilsoukaini/security-risks-of-ai-generated-code-research-and-a-review-checklist-fd87b946c18d?source=rss-7f614efc1b10------2</link>
            <guid isPermaLink="false">https://medium.com/p/fd87b946c18d</guid>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Adil Soukaini | Full-Stack Developer]]></dc:creator>
            <pubDate>Fri, 30 Oct 2026 08:00:04 GMT</pubDate>
            <atom:updated>2026-11-06T11:13:56.587Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*EmqMun_rWipFK72J.png" /></figure><p>AI-generated code usually works, but it is often not secure. In Veracode’s 2026 tests, models produced code that compiled almost 100% of the time, yet failed security checks on nearly 44% of tasks, a pass rate that has not improved since 2025. The fix is not to ban AI tools; it is to treat their output like code from a contractor you have never worked with: review it, scan it, and test it before it reaches production.</p><p>Below is a summary of the research, the failure patterns that come up again and again, and a review checklist you can hand to your developer or agency.</p><h3>What the research found</h3><h3>Veracode: about half of security-relevant tasks fail, and it is not improving</h3><p>In its <a href="https://www.veracode.com/press-release/ai-generated-code-poses-major-security-risks-in-nearly-half-of-all-development-tasks-veracode-research-reveals/">2025 GenAI Code Security Report</a>, Veracode gave 80 curated coding tasks to more than 100 large language models. When a secure and an insecure approach were both possible, models chose the insecure one 45% of the time. Java was the riskiest language, with a security failure rate above 70%; Python, C# and JavaScript failed between 38% and 45% of the time. Models failed to defend against cross-site scripting in 86% of cases and log injection in 88%. Larger models did not perform significantly better than smaller ones.</p><p>The <a href="https://www.veracode.com/news/llms-are-getting-smarter-but-not-safer-veracode-2026-genai-code-security-report-finds-ai-generated-code-security-has-stalled-at-56%25-pass-rate/">2026 edition</a> found the average security pass rate “sits at 56 percent — virtually unchanged since last year’s report.” The best model on its summer 2026 leaderboard reached 68%, which still means failing nearly one security task in three. Python code passed 63% of the time, Java only 30%. Coding-specialised models (51%) did no better than general-purpose ones (52%), and model size made no clear difference, although reasoning models kept an edge (56% against 51%).</p><h3>Academic studies: vulnerable suggestions and overconfident users</h3><ul><li>An early study of GitHub Copilot, <a href="https://arxiv.org/abs/2108.09293">“Asleep at the Keyboard?”</a> (IEEE Symposium on Security and Privacy 2022), generated 1,689 programs across 89 security-relevant scenarios and found approximately 40% of them vulnerable.</li><li>A user study published at ACM CCS 2023 found that <a href="https://arxiv.org/abs/2211.03622">participants with an AI assistant wrote significantly less secure code</a> than those without, and were more likely to believe their code was secure.</li><li>A USENIX Security 2025 paper analysed 576,000 code samples from 16 models and found that <a href="https://arxiv.org/abs/2406.10279">at least 5.2% of packages suggested by commercial models and 21.7% by open-source models did not exist</a>, with 205,474 unique invented package names. An attacker can publish a malicious package under a name that models tend to invent, and anyone who installs it without checking is compromised.</li></ul><h3>Team level: more changes, less stability</h3><p>Google’s <a href="https://cloud.google.com/blog/products/ai-machine-learning/announcing-the-2025-dora-report">2025 DORA report</a> found that AI adoption is linked to higher delivery throughput but still has a negative relationship with delivery stability. Its explanation applies directly to security: without strong automated testing and fast feedback loops, more change volume leads to more incidents.</p><h3>Why AI-generated code fails security checks</h3><p>Security usually depends on context the model does not have: which input comes from an untrusted user, who is allowed to see which record, which secrets exist in your environment. Cross-site scripting, which depends on knowing where a piece of data will be displayed, had one of the worst results in Veracode’s 2025 tests (86% failure). The model writes code that does what you asked; whether it is safe depends on what you did not say.</p><h3>The failure patterns to look for</h3><ul><li><strong>Untrusted input reaching output or queries</strong>: cross-site scripting, SQL injection, log injection.</li><li><strong>Missing authorisation checks</strong>: an endpoint that checks the user is logged in, but not that the record belongs to them.</li><li><strong>Secrets in code</strong>: API keys pasted into source files or front-end bundles.</li><li><strong>Invented or outdated dependencies</strong>: packages that do not exist, or old versions with known vulnerabilities.</li><li><strong>Insecure defaults</strong>: permissive CORS, debug mode on, verbose error messages in production.</li><li><strong>Weak cryptography or home-made auth</strong> instead of the framework’s built-in tools.</li></ul><h3>A security review checklist for AI-generated code</h3><p>You don’t need to read code to use this list. Ask your developer to confirm each point for every feature that touches users, payments or data. Items 1 to 4 apply before any change is merged, 5 to 7 cover dependencies and secrets, and 8 to 12 cover testing and process.</p><ol><li>A human who understands the code has reviewed every AI-generated change, line by line.</li><li>Every endpoint checks authorisation (who can access this record), not just authentication.</li><li>All user input is validated on the server, and output is escaped by the framework (no raw HTML rendering of user content).</li><li>Database access uses the framework’s query builder or parameterised queries, never string-built SQL.</li><li>Every new package was checked: it exists, is maintained, and is the one the developer meant to install.</li><li>A dependency audit runs automatically (for example npm audit, composer audit or GitHub Dependabot).</li><li>No secret is committed to the repository or shipped to the browser; secret scanning is enabled.</li><li>Automated tests cover login, permissions and payment flows, and run on every change. End-to-end tests such as the Playwright setup we use catch broken access rules early.</li><li>A static analysis (SAST) tool runs in CI on every pull request.</li><li>Production runs with debug off, restrictive CORS and generic error pages.</li><li>AI agents cannot touch production data or credentials directly.</li><li>Someone owns security decisions, and you know who.</li></ol><h3>Questions to ask your developer or agency</h3><ul><li>Which AI tools do you use, and what is your review rule for their output?</li><li>What runs automatically on every pull request: tests, dependency audit, static analysis, secret scanning?</li><li>How are authorisation rules tested?</li><li>Who would notice if a new package did not exist yesterday?</li></ul><p>These questions belong in any <a href="https://shipfastdev.com/blog/how-to-choose-a-remote-development-partner">evaluation of a development partner</a>. If your app was built mainly with an AI app builder, read where vibe-coded MVPs break as well: missing data-access rules are the most common problem there.</p><h3>Frequently asked questions</h3><h3>Is AI-generated code secure?</h3><p>Not by default. Veracode’s 2026 tests found an average security pass rate of 56%, unchanged from 2025, even though almost all generated code compiled. It needs the same review and testing as any other code, and arguably more.</p><h3>Which programming language is riskiest for AI-generated code?</h3><p>In Veracode’s tests, Java: a 30% security pass rate in 2026, against 63% for Python. Language matters less than review, though: every language failed a significant share of tasks.</p><h3>Can an AI tool review its own code for security?</h3><p>It can help spot issues, but the same models chose insecure options in a large share of Veracode’s tasks. Use AI review as one layer, alongside static analysis, dependency audits, tests and a human reviewer.</p><h3>Do I need a security audit for my MVP?</h3><p>If it handles personal data, payments or logins, at least run through the checklist above before launch. A full external audit makes sense once you have real users or sensitive data at scale.</p><h3>Sources</h3><ol><li><a href="https://www.veracode.com/news/llms-are-getting-smarter-but-not-safer-veracode-2026-genai-code-security-report-finds-ai-generated-code-security-has-stalled-at-56%25-pass-rate/">LLMs Are Getting Smarter but Not Safer: Veracode 2026 GenAI Code Security Report</a>, Veracode, 2026–07</li><li><a href="https://www.veracode.com/press-release/ai-generated-code-poses-major-security-risks-in-nearly-half-of-all-development-tasks-veracode-research-reveals/">AI-Generated Code Poses Major Security Risks in Nearly Half of All Development Tasks, Veracode Research Reveals</a>, Veracode, 2025–07</li><li><a href="https://arxiv.org/abs/2108.09293">Asleep at the Keyboard? Assessing the Security of GitHub Copilot’s Code Contributions (IEEE S&amp;P 2022)</a>, Pearce, Ahmad, Tan, Dolan-Gavitt, Karri (arXiv), 2021–12</li><li><a href="https://arxiv.org/abs/2211.03622">Do Users Write More Insecure Code with AI Assistants? (ACM CCS 2023)</a>, Perry, Srivastava, Kumar, Boneh (arXiv), 2023–12</li><li><a href="https://arxiv.org/abs/2406.10279">We Have a Package for You! A Comprehensive Analysis of Package Hallucinations by Code Generating LLMs (USENIX Security 2025)</a>, Spracklen et al. (arXiv), 2025–03</li><li><a href="https://cloud.google.com/blog/products/ai-machine-learning/announcing-the-2025-dora-report">Announcing the 2025 DORA Report</a>, Google Cloud, 2025–09</li></ol><p><em>Originally published at </em><a href="https://shipfastdev.com/blog/ai-generated-code-security-risks"><em>https://shipfastdev.com</em></a><em> on October 30, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fd87b946c18d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Deploying a Laravel app on Google Cloud Run: the pitfalls and the fixes]]></title>
            <link>https://medium.com/@adilsoukaini/deploying-a-laravel-app-on-google-cloud-run-the-pitfalls-and-the-fixes-18daf9c8bfb6?source=rss-7f614efc1b10------2</link>
            <guid isPermaLink="false">https://medium.com/p/18daf9c8bfb6</guid>
            <category><![CDATA[laravel]]></category>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[google-cloud-platform]]></category>
            <category><![CDATA[docker]]></category>
            <dc:creator><![CDATA[Adil Soukaini | Full-Stack Developer]]></dc:creator>
            <pubDate>Wed, 28 Oct 2026 08:00:26 GMT</pubDate>
            <atom:updated>2026-11-04T18:31:47.945Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*NK8qUJODZfjtnPua.png" /></figure><p>To deploy a Laravel app on Google Cloud Run, you put PHP-FPM and a web server in one container that listens on the port Cloud Run gives it, build your frontend assets before the image, keep configuration on the service rather than in your CI file, and tell Laravel it runs behind an HTTPS proxy. Most of the work is in a handful of Laravel-specific traps; here are the ones we hit putting Driveway Morocco into production, and the fixes.</p><p>Driveway Morocco is ShipFastDev’s car-rental platform ( <a href="https://shipfastdev.com/work/driveway-morocco">case study</a>): Laravel 13 with an Inertia/React storefront, a Filament admin panel, Stripe deposit holds, and plugins (fleet, booking, payments) installed as local Composer packages. It went live on Cloud Run in August 2026, with a managed Postgres database on Supabase and GitHub Actions deploying on every push to the main branch. Every pitfall below comes from that deployment’s commit history.</p><h3>What Cloud Run expects from your container</h3><p>Read the <a href="https://docs.cloud.google.com/run/docs/container-contract">container runtime contract</a> before you write the Dockerfile. Four rules matter most for a Laravel app:</p><ul><li><strong>Port.</strong> The container must listen on 0.0.0.0 on the port in the PORT environment variable. Requests are sent to 8080 by default.</li><li><strong>No TLS inside the container.</strong> Cloud Run terminates HTTPS and proxies requests to your container as plain HTTP. Laravel has to be told about it, or it builds http:// links.</li><li><strong>The filesystem is memory.</strong> Writing to disk uses the instance’s RAM and disappears when the instance stops. Sessions, uploads and logs belong in the database, object storage or stdout.</li><li><strong>Startup and shutdown limits.</strong> An instance must listen for requests within 4 minutes of starting. On shutdown it receives SIGTERM, then SIGKILL 10 seconds later.</li></ul><p>Driveway runs a single image with PHP 8.4 FPM, Nginx and Supervisor. Nginx listens on the port, serves the built assets and hands PHP requests to FPM.</p><h3>Pitfall 1: frontend assets missing from the build</h3><p>With gcloud run deploy --source ., your directory is uploaded and <a href="https://docs.cloud.google.com/run/docs/deploying-source-code">built by Cloud Build using your Dockerfile</a> if one is present. What gets uploaded is decided by .gcloudignore, and the <a href="https://docs.cloud.google.com/sdk/gcloud/reference/topic/gcloudignore">default file pulls in your </a><a href="https://docs.cloud.google.com/sdk/gcloud/reference/topic/gcloudignore">.gitignore</a>. Vite writes its output to public/build, which is gitignored, so the first CI deploys shipped an image with no compiled CSS or JavaScript.</p><p>Running npm ci &amp;&amp; npm run build inside the image did not fix it cleanly: the build regenerated manifest.json with new hashes but no matching JS files. The working setup builds the assets in the CI job before the deploy step, and the Dockerfile only copies public/build. Removing Node and npm from the image also made it about 200 MB smaller.</p><h3>Pitfall 2: your local vendor/ overwriting the real one</h3><p>The Dockerfile runs composer install and then copies the source. Because vendor/ was not listed in .gcloudignore, COPY . replaced the freshly installed dependencies with the developer&#39;s local vendor/, and the Filament admin routes broke. Exclude vendor/ and node_modules/ from the upload; the image should install its own dependencies from the lockfile.</p><h3>Pitfall 3: Composer path repositories and symlinks</h3><p>Driveway’s plugins live in a plugins/ folder and are installed as Composer path repositories. <a href="https://getcomposer.org/doc/05-repositories.md">Composer symlinks path packages by default</a>. Inside the Docker build, the copy turned those symlinks into real directories, and the autoloader stopped finding plugin classes. Vehicle detail and checkout pages, which depended on plugin routes, rendered blank.</p><p>The fixes in the history were pragmatic: a build step that adds the plugins’ PSR-4 namespaces to Composer’s autoload map, a core-owned fleet controller, and core fallback routes for the booking flow. The image also needed the intl, zip and gd extensions for composer install to succeed. If you are starting fresh, try Composer&#39;s &quot;symlink&quot;: false option first: the Composer documentation notes that forcing mirroring &quot;can be useful when deploying&quot;.</p><h3>Pitfall 4: layer caching that breaks the build</h3><p>Docker’s advice is to <a href="https://docs.docker.com/build/cache/optimize/">copy the package files, install dependencies, then copy the source</a>. A commit applied that order (copy composer.json and plugins/, install, then copy the rest) and estimated typical deploys would drop from about 8 minutes to about 2. It was reverted the same day because the new order broke CI builds: with path repositories, the dependency layer depends on your own plugin code too. Get a working image first, then optimise, and prove the change with a real CI build.</p><h3>Pitfall 5: http:// links behind the HTTPS proxy</h3><p>This one took the whole storefront down. PHP-FPM saw plain HTTP, so Laravel’s url() helper produced http:// URLs for assets and API calls. The page was served over HTTPS, so the Content Security Policy blocked those requests: navigation, login, search, filters and booking all stopped working, with no server error to point at.</p><p>The fix was to make Laravel generate HTTPS URLs: set APP_URL and ASSET_URL to the public domain, force HTTPS in the startup script, and call URL::forceScheme(&#39;https&#39;) in production. The <a href="https://laravel.com/docs/requests#configuring-trusted-proxies">Laravel documentation</a> describes the standard route: enable the TrustProxies middleware so Laravel reads the forwarded protocol, and use * when you don&#39;t know your load balancer&#39;s addresses.</p><pre>// bootstrap/app.php<br>-&gt;withMiddleware(function (Middleware $middleware): void {<br>    $middleware-&gt;trustProxies(at: &#39;*&#39;);<br>})</pre><p>Either way, check it in a browser with the console open.</p><h3>Pitfall 6: secrets in the CI file</h3><p>The first workflow passed environment variables in the deploy step, including the database password and the app key. They were removed within the day. Cloud Run carries a revision’s configuration into the next revisions <a href="https://docs.cloud.google.com/run/docs/configuring/services/environment-variables">unless you change it explicitly</a>, so the pipeline only needs to ship a new image. Two details from the same page: --set-env-vars deletes every variable you don&#39;t list, while --update-env-vars only adds or changes; and Google does not recommend environment variables for secrets such as database credentials or API keys, pointing to Secret Manager instead.</p><p>For authentication, the pipeline first used a service account key stored as a GitHub secret. In September 2026 it moved to Workload Identity Federation with Google’s deploy action. The <a href="https://github.com/google-github-actions/auth">google-github-actions/auth documentation</a> recommends federation because it “obviates the need to export a long-lived credential”; a key file never expires and must be treated like a password. The workflow needs the id-token: write permission.</p><h3>Pitfall 7: cold starts and memory</h3><p>A new instance starts from the image, so anything cached at build time can be stale. Driveway’s startup script clears the config and route caches so the plugin manager reads which plugins are enabled from the database, and runs php artisan package:discover so Laravel registers installed packages. Keep this script short: it runs on every cold start, and the instance has 4 minutes to start listening.</p><p>Memory is the other limit. The <a href="https://docs.cloud.google.com/run/docs/configuring/services/memory-limits">default is 512 MiB per instance</a>, instances that exceed their limit are terminated, and files written to disk count against it. Driveway’s service was raised to 1 GiB for stability.</p><h3>Pitfall 8: one admin feature taking down every page</h3><p>In September 2026, a plugin registered an admin-only Filament/Livewire component in its boot() method. In the Cloud Run boot context that call threw, and because the exception happened during application boot, every page failed, storefront included. The error pages then called route(&#39;home&#39;), which wasn&#39;t in the route list after the failed boot, so React unmounted and visitors got a white screen.</p><p>Two fixes: admin-only registration is wrapped in a try/catch so it is best-effort, and the error pages fall back to plain paths when a named route is missing. Boot code runs on every request; an error page must not depend on anything that can fail.</p><h3>Queues and scheduled tasks</h3><p>The project’s production checklist lists a queue worker and the Laravel scheduler. A backend audit had already found that, without a worker, queued emails, push notifications and the deposit-hold expiry command silently never ran in the local Docker setup. On Cloud Run, plan this explicitly: under the default request-based billing, the container contract says CPU is allocated while instances process requests, so a queue:work process in the web container is not a reliable worker. The documented options are instance-based billing, which keeps CPU allocated for the instance&#39;s lifetime, or a Cloud Run job <a href="https://docs.cloud.google.com/run/docs/execute/jobs-on-schedule">triggered by Cloud Scheduler</a> for periodic commands.</p><h3>The checklist</h3><ul><li>Nginx and PHP-FPM in one image, listening on $PORT.</li><li>Assets built in CI; no Node in the runtime image.</li><li>vendor/ and node_modules/ excluded from the upload.</li><li>Autoloading tested inside the built image.</li><li>HTTPS URLs verified in a real browser.</li><li>Secrets in Secret Manager, CI on Workload Identity Federation.</li><li>Short startup script, memory sized for the app.</li><li>Guarded boot code, error pages that always render.</li><li>A plan for queues and cron.</li><li>Browser tests before deploy: see end-to-end tests with Playwright in CI.</li></ul><p>Most of these problems appeared only in the deployed container. Small, frequent deploys made each one easy to trace, as argued in <a href="https://shipfastdev.com/blog/weekly-releases-with-a-remote-developer">weekly releases with a remote developer</a>. The same project’s search setup, with a fallback when the search engine is down, is covered in Laravel search with Meilisearch and a database fallback.</p><h3>Frequently asked questions</h3><h3>Can Laravel run on Cloud Run without code changes?</h3><p>Mostly. You need a container that listens on $PORT, Laravel configured to generate HTTPS URLs behind the proxy, configuration from environment variables or Secret Manager, and no reliance on the local disk for anything that must persist.</p><h3>Do I need Nginx inside the container?</h3><p>Cloud Run forwards HTTP to your container, and PHP-FPM speaks FastCGI, not HTTP, so a web server has to sit in front of it. Driveway Morocco uses Nginx and PHP-FPM under Supervisor.</p><h3>How do I run Laravel queues and cron jobs on Cloud Run?</h3><p>Not as a background process in the web container under request-based billing, because CPU is only allocated while requests are processed. Use instance-based billing for a worker, or Cloud Run jobs triggered by Cloud Scheduler for scheduled commands.</p><h3>Why does my Laravel app on Cloud Run load assets over <a href="http://?">http://?</a></h3><p>Cloud Run terminates TLS and sends plain HTTP to the container, so Laravel thinks the request is insecure. Trust the proxy with the TrustProxies middleware or force the HTTPS scheme in production, and set APP_URL to your https:// domain.</p><h3>Sources</h3><ol><li><a href="https://docs.cloud.google.com/run/docs/container-contract">Container runtime contract</a>, Google Cloud, 2026–09</li><li><a href="https://docs.cloud.google.com/run/docs/deploying-source-code">Deploy services from source code</a>, Google Cloud, 2026–09</li><li><a href="https://docs.cloud.google.com/sdk/gcloud/reference/topic/gcloudignore">gcloud topic gcloudignore</a>, Google Cloud, 2026–10</li><li><a href="https://docs.cloud.google.com/run/docs/configuring/services/environment-variables">Configure environment variables for services</a>, Google Cloud, 2026–09</li><li><a href="https://docs.cloud.google.com/run/docs/configuring/services/memory-limits">Configure memory limits for services</a>, Google Cloud, 2026–09</li><li><a href="https://docs.cloud.google.com/run/docs/execute/jobs-on-schedule">Execute jobs on a schedule</a>, Google Cloud, 2026–09</li><li><a href="https://github.com/google-github-actions/auth">google-github-actions/auth</a>, Google (GitHub), 2026–10</li><li><a href="https://laravel.com/docs/requests#configuring-trusted-proxies">HTTP Requests: Configuring Trusted Proxies</a>, Laravel 13.x documentation, 2026–10</li><li><a href="https://getcomposer.org/doc/05-repositories.md">Repositories: Path</a>, Composer documentation, 2026–10</li><li><a href="https://docs.docker.com/build/cache/optimize/">Optimize cache usage in builds</a>, Docker Docs, 2026–10</li></ol><p><em>Originally published at </em><a href="https://shipfastdev.com/blog/deploy-laravel-google-cloud-run"><em>https://shipfastdev.com</em></a><em> on October 28, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=18daf9c8bfb6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Fixed price vs time and materials vs retainer: which software contract should you choose?]]></title>
            <link>https://medium.com/@adilsoukaini/fixed-price-vs-time-and-materials-vs-retainer-which-software-contract-should-you-choose-b1f48f6e28bd?source=rss-7f614efc1b10------2</link>
            <guid isPermaLink="false">https://medium.com/p/b1f48f6e28bd</guid>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[freelancing]]></category>
            <dc:creator><![CDATA[Adil Soukaini | Full-Stack Developer]]></dc:creator>
            <pubDate>Mon, 26 Oct 2026 08:00:06 GMT</pubDate>
            <atom:updated>2026-11-04T18:23:31.664Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*kef2H86iUMP_K1Xf.png" /></figure><p>Choose <strong>fixed price</strong> when you can write the scope down completely and it will not change; choose <strong>time and materials</strong> when the work is uncertain and you can follow it closely, always with a spending cap; choose a <strong>retainer</strong> when the product needs steady work every month and priorities shift as you learn. Each model moves risk to a different party: fixed price moves cost risk to the provider and change risk to you, time and materials moves cost risk to you, and a retainer fixes spend while leaving scope flexible.</p><h3>Fixed price: the provider carries the cost risk</h3><p>US federal procurement rules give a precise definition: a firm-fixed-price contract <a href="https://www.acquisition.gov/far/16.202-1">“provides for a price that is not subject to any adjustment on the basis of the contractor’s cost experience”</a>. If the work takes twice as long as estimated, the provider absorbs it. That is attractive, and it has consequences:</p><ul><li><strong>The estimate includes a risk margin.</strong> A provider who carries the risk prices it in. On well-understood work the margin is small; on vague work it is large or the provider loses money.</li><li><strong>Scope becomes a negotiation.</strong> Anything not written in the specification is a change request, with its own quote and delay. In product work, where you learn from users every week, that friction adds up.</li><li><strong>Quality is the hidden variable.</strong> When the budget runs short, the easiest things to cut are tests, documentation and edge cases. Protect them with written acceptance criteria.</li></ul><p><strong>Good fit:</strong> a marketing site from an approved design, a defined integration with a documented API, a migration with a clear before and after, a discovery or prototyping phase with a fixed length.</p><h3>Time and materials: you carry the cost risk</h3><p>Under the same rules, a time-and-materials contract pays for <a href="https://www.acquisition.gov/far/16.601">“direct labor hours at specified fixed hourly rates”</a>, and may be used only when it is not possible at the outset to estimate the extent or duration of the work with confidence. Two safeguards are mandatory there, and they are good practice for any startup:</p><ul><li><strong>A ceiling price</strong> that the provider exceeds at its own risk. For you: a monthly or per-feature cap that requires your approval to exceed.</li><li><strong>Active oversight</strong>, because time-based pay gives no built-in incentive to control cost. For you: weekly timesheets tied to tasks, and regular releases you can check.</li></ul><p>Time and materials is honest about uncertainty: you pay for the work actually done, and you can change direction at any time. The risk is drift: without priorities and caps, hours expand.</p><p><strong>Good fit:</strong> an MVP whose scope will move, debugging or performance work on an unfamiliar codebase, research spikes, extending a team with in-house leadership.</p><h3>Retainer: fixed spend, flexible scope</h3><p>A retainer pays a fixed monthly amount for an agreed capacity and set of responsibilities: for example a number of development hours, code reviews, deployments and hosting care. You decide each month what that capacity works on.</p><ul><li><strong>Budget is predictable</strong>, like fixed price.</li><li><strong>Scope is flexible</strong>, like time and materials: you reprioritise without a change request.</li><li><strong>Continuity is built in</strong>: the same people keep context, which supports small, frequent releases (see <a href="https://shipfastdev.com/blog/weekly-releases-with-a-remote-developer">weekly releases with a remote developer</a>).</li></ul><p>The risks are paying for idle capacity if you have no backlog, and disagreement about what “included” means. Write down the hours, what happens to unused hours, response times, what counts as out of scope and the notice period.</p><p><strong>Good fit:</strong> a live product that needs regular features and fixes, post-launch iteration on an MVP, a small company without in-house developers.</p><h3>How each model handles a change</h3><p>Imagine that two weeks into the work, user feedback shows the onboarding flow should be rebuilt.</p><ol><li><strong>Fixed price:</strong> the provider writes a change request with a new price and date; you approve, negotiate or drop the change. The original deadline may move.</li><li><strong>Time and materials:</strong> you reprioritise; the hours go to the new flow. Check the cap: if the change pushes you over it, decide what to cut.</li><li><strong>Retainer:</strong> you move the onboarding rebuild to the top of next month’s priorities; something else waits. Spend does not change.</li></ol><p>The <a href="https://agilemanifesto.org/">Agile Manifesto</a> put this tension into one line in 2001: “customer collaboration over contract negotiation”. A contract cannot remove uncertainty; it can only decide who pays for it and how quickly you can react.</p><h3>Hybrids that work</h3><ul><li><strong>Fixed-price discovery, then retainer or time and materials.</strong> Pay a fixed amount for a short scoping phase (specification, design, estimate), then build with a flexible model. See <a href="https://shipfastdev.com/blog/how-to-scope-an-mvp">how to scope an MVP</a>.</li><li><strong>Fixed price per iteration.</strong> Agree a fixed price for each two- to four-week cycle with a short list of goals, and re-plan between cycles.</li><li><strong>Time and materials with a cap per feature.</strong> Flexible, but every feature has a budget that triggers a conversation when 80% is used.</li></ul><h3>A note for France: régie and staff lending</h3><p>In France, time-based work is often called <em>régie</em>. French labour law prohibits <a href="https://code.travail.gouv.fr/code-du-travail/l8241-1">any for-profit operation whose sole purpose is lending staff</a> (outside regulated temporary work), so a régie contract must remain a service: the provider brings its own expertise and keeps authority over its people. With an individual freelancer, a client who sets hours, gives detailed orders and integrates the person like an employee risks <a href="https://bpifrance-creation.fr/encyclopedie/micro-entreprise-regime-auto-entrepreneur/lessentiel-ce-regime/independance-juridique">requalification as employment</a>. Define deliverables and leave the provider in charge of how the work is done.</p><h3>How to choose</h3><ol><li><strong>Can you write the scope completely today?</strong> Yes: fixed price is an option. No: time and materials or a retainer.</li><li><strong>Is the work a one-off or ongoing?</strong> One-off: fixed price or capped time and materials. Ongoing: retainer.</li><li><strong>Who will follow the work weekly?</strong> If nobody can, avoid uncapped time and materials.</li><li><strong>How will you judge progress?</strong> Whatever the model, look at delivered results. DORA’s <a href="https://dora.dev/guides/dora-metrics/">delivery metrics</a>, such as deployment frequency and change lead time, work for all three.</li></ol><p>For budget figures behind each model, see <a href="https://shipfastdev.com/blog/how-much-does-an-mvp-cost">how much an MVP costs</a> and <a href="https://shipfastdev.com/blog/cost-to-hire-a-software-developer">what a software developer costs</a>; for the provider side, <a href="https://shipfastdev.com/blog/freelancer-vs-agency-vs-dev-retainer">freelancer vs agency vs dev retainer</a>.</p><h3>Frequently asked questions</h3><h3>What is the difference between fixed price and time and materials?</h3><p>Fixed price sets the total price in advance, so the provider carries the cost risk and changes need a new quote. Time and materials pays for hours actually worked at agreed rates, so you carry the cost risk but can change direction freely.</p><h3>Is fixed price good for an MVP?</h3><p>Only if the scope is truly settled. Most MVPs change as users respond, and every change becomes a change request. A fixed-price discovery phase followed by a flexible model usually works better.</p><h3>What is a development retainer contract?</h3><p>A monthly contract for a fixed amount of development capacity and agreed responsibilities. You set priorities each month; spend stays the same.</p><h3>How do I control costs in a time and materials contract?</h3><p>Set a cap per month or per feature that requires your approval to exceed, ask for weekly time reports linked to tasks, and insist on regular releases you can check.</p><h3>Sources</h3><ol><li><a href="https://www.acquisition.gov/far/16.202-1">FAR 16.202–1 Firm-fixed-price contracts</a>, Federal Acquisition Regulation (Acquisition.gov), n.d.</li><li><a href="https://www.acquisition.gov/far/16.601">FAR 16.601 Time-and-materials contracts</a>, Federal Acquisition Regulation (Acquisition.gov), n.d.</li><li><a href="https://agilemanifesto.org/">Manifesto for Agile Software Development</a>, agilemanifesto.org, 2001</li><li><a href="https://code.travail.gouv.fr/code-du-travail/l8241-1">Article L8241–1 du Code du travail</a>, Code du travail numérique (Ministère du Travail), n.d.</li><li><a href="https://bpifrance-creation.fr/encyclopedie/micro-entreprise-regime-auto-entrepreneur/lessentiel-ce-regime/independance-juridique">Indépendance juridique : risque de requalification du contrat de mission en contrat de travail</a>, Bpifrance Création, 2026–06</li><li><a href="https://dora.dev/guides/dora-metrics/">DORA’s software delivery performance metrics</a>, DORA, 2026–01</li></ol><p><em>Originally published at </em><a href="https://shipfastdev.com/blog/fixed-price-vs-time-and-materials-vs-retainer"><em>https://shipfastdev.com</em></a><em> on October 26, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b1f48f6e28bd" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Laravel vs Node.js for a startup backend: how to choose]]></title>
            <link>https://medium.com/@adilsoukaini/laravel-vs-node-js-for-a-startup-backend-how-to-choose-211d32eca67a?source=rss-7f614efc1b10------2</link>
            <guid isPermaLink="false">https://medium.com/p/211d32eca67a</guid>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[laravel]]></category>
            <category><![CDATA[nodejs]]></category>
            <category><![CDATA[startup]]></category>
            <dc:creator><![CDATA[Adil Soukaini | Full-Stack Developer]]></dc:creator>
            <pubDate>Fri, 23 Oct 2026 08:00:37 GMT</pubDate>
            <atom:updated>2026-10-30T13:03:39.418Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*z2WOelMAOr1h6IEy.png" /></figure><p>For most startup backends, both Laravel and Node.js will do the job; the better choice depends on your product and your team. Laravel gets a typical SaaS or marketplace backend (accounts, admin, payments, emails, background jobs) built fastest because those pieces ship with the framework. Node.js, usually with NestJS or Express, fits best when your team writes TypeScript everywhere, when the product is heavy on real-time connections, or when you want the largest hiring pool.</p><p>Here is how they compare on the four things a founder should care about: ecosystem, hiring, speed of building and hosting.</p><h3>First, what you are comparing</h3><p>So the fair comparisons are Laravel vs NestJS (both opinionated, batteries-included) and Laravel vs Express (opinionated vs minimal).</p><h3>Ecosystem: included vs assembled</h3><p>Laravel’s advantage is that the common parts of a business application are first-party and designed to work together. Authentication scaffolding comes from its starter kits, API tokens from Sanctum, background work from queues, and the official package list includes billing (Cashier), search (Scout), queue monitoring (Horizon) and real-time broadcasting. One set of docs, one upgrade cycle.</p><p>In Node.js, the npm ecosystem is enormous, but you assemble the stack: an ORM, a validation library, a queue, an auth solution. NestJS reduces the assembly by giving you modules, dependency injection and conventions. The NestJS docs are explicit about why it exists: many Node.js libraries exist, but <a href="https://docs.nestjs.com/">“none of them effectively solves the main problem of architecture”</a>.</p><p>For a founder, the practical question is: how many decisions does your developer have to make before building your first feature? With Laravel, fewer. With Express, many. NestJS sits in between.</p><h3>Hiring pool: what the survey data says</h3><p>The 2025 Stack Overflow Developer Survey (23,678 respondents to the web frameworks question) reports these usage shares among all respondents:</p><p>In the <a href="https://survey.stackoverflow.co/2024/technology">2024 survey</a>, the same frameworks stood at Node.js 40.8%, Express 17.8%, NestJS 5.8% and Laravel 7.9%. The samples differ each year, so treat this as direction rather than growth rates, but neither side is shrinking.</p><p>Two readings matter for hiring:</p><ol><li><strong>JavaScript/TypeScript developers are the larger pool.</strong> If your frontend is React or React Native, one language across the stack lets the same people work on both sides.</li><li><strong>Framework-specific pools are closer than the language numbers suggest.</strong> NestJS (6.7%) and Laravel (8.9%) are in the same range. Hiring “a Node.js developer” does not mean hiring someone who knows how to structure a large backend.</li></ol><p>PHP also has a very large installed base: W3Techs reports that PHP is used by <a href="https://w3techs.com/technologies/details/pl-php">69.8% of websites whose server-side language it knows</a> (October 2026). That is mostly existing sites, including WordPress, but it means PHP skills are common among agencies and freelancers. For more on today’s hiring market, see hiring developers in 2026.</p><h3>Speed of building</h3><p>For a CRUD-heavy product (users, listings, bookings, invoices, an admin area) Laravel is usually quicker to a working first version, because the conventions answer most structural questions. Official starter kits provide the authentication screens and backend logic, and a new project uses SQLite by default, so you can start without a database server.</p><p>Node.js catches up, or wins, when:</p><ul><li>the frontend and backend share TypeScript types, so API changes are checked at compile time on both sides;</li><li>the product keeps many connections open (chat, live dashboards, multiplayer features), which suits Node.js’s non-blocking design: the Node.js site notes that <a href="https://nodejs.org/en/about">“many connections can be handled concurrently”</a>;</li><li>you are building small services or serverless functions rather than one main application.</li></ul><p>Node.js is weaker for CPU-heavy work in the request path; its documentation points to child processes and the cluster module to use multiple cores. Laravel pushes heavy work to queues. In both cases, long tasks belong in background jobs.</p><h3>Hosting and running costs</h3><p>Both run anywhere containers run. Laravel has its own managed options (Laravel Cloud, Forge, Vapor), and it runs well on container platforms; we describe one setup in <a href="https://shipfastdev.com/blog/deploy-laravel-google-cloud-run">deploying Laravel on Google Cloud Run</a>. Node.js has the broadest choice of serverless and edge platforms, which helps if your frontend is Next.js.</p><p>At MVP scale, the hosting bill rarely decides the choice. The database, queues, search and file storage cost the same whatever language talks to them.</p><h3>From our work</h3><p>ShipFastDev builds with both Laravel and NestJS. The <a href="https://shipfastdev.com/work/driveway-morocco">Driveway Morocco</a> platform is a single Laravel application that serves the React web interface through Inertia and, since the mobile app was added, a JSON API with Sanctum token authentication for the Expo app. Search runs on Laravel Scout with Meilisearch (described in Laravel search with a Meilisearch fallback). One backend serves two clients, which kept the mobile app from needing its own server.</p><h3>Decision checklist</h3><p><strong>Choose Laravel if:</strong></p><ul><li>your product is a classic web application or marketplace with an admin side;</li><li>you want auth, queues, emails, scheduled jobs and testing set up the same way on day one;</li><li>your developer or partner already knows PHP well.</li></ul><p><strong>Choose Node.js with NestJS if:</strong></p><ul><li>your team writes TypeScript across web, mobile and backend;</li><li>real-time features are central to the product;</li><li>you expect to hire several backend developers and want the largest pool.</li></ul><p><strong>Choose plain Express if</strong> the backend is small (a few endpoints, webhooks, a proxy to other services) and you don’t need a framework’s structure.</p><p>Whatever you choose, the bigger risk is an unstructured codebase, not the language; see technical debt explained for founders.</p><h3>Frequently asked questions</h3><h3>Is Laravel still a good choice in 2026?</h3><p>Yes. Laravel’s share in the Stack Overflow survey went from 7.9% (2024) to 8.9% (2025), and the current documentation covers Laravel 13.</p><h3>Is Node.js faster than Laravel?</h3><p>For many concurrent, I/O-bound connections Node.js’s event-driven model is a good fit. For a typical CRUD API the database and caching usually matter more than the language. Measure with your own workload before deciding on speed alone.</p><h3>Laravel or NestJS for a mobile app backend?</h3><p>Both work: each can expose a JSON API with token authentication. Pick NestJS if your mobile team writes TypeScript and wants shared types; pick Laravel if you also need a web admin and want it in the same app.</p><h3>Can I switch from Laravel to Node.js later?</h3><p>Yes, but it is a rewrite. A more common path is to keep the main backend and add a separate Node.js service for the part that needs it, such as real-time features.</p><h3>Sources</h3><ol><li><a href="https://survey.stackoverflow.co/2025/technology">2025 Stack Overflow Developer Survey: Technology</a>, Stack Overflow, July 2025</li><li><a href="https://survey.stackoverflow.co/2024/technology">2024 Stack Overflow Developer Survey: Technology</a>, Stack Overflow, July 2024</li><li><a href="https://laravel.com/docs/13.x/installation">Laravel 13.x documentation: Installation</a>, Laravel, accessed October 2026</li><li><a href="https://docs.nestjs.com/">NestJS documentation: Introduction</a>, NestJS, accessed October 2026</li><li><a href="https://nodejs.org/en/about">About Node.js</a>, OpenJS Foundation / Node.js, accessed October 2026</li><li><a href="https://w3techs.com/technologies/details/pl-php">Usage statistics of PHP for websites</a>, W3Techs, October 2026</li></ol><p><em>Originally published at </em><a href="https://shipfastdev.com/blog/laravel-vs-nodejs-for-startups"><em>https://shipfastdev.com</em></a><em> on October 23, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=211d32eca67a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Hiring Junior Developers in the AI Era: What the Data Shows]]></title>
            <link>https://medium.com/@adilsoukaini/hiring-junior-developers-in-the-ai-era-what-the-data-shows-651c5af671b6?source=rss-7f614efc1b10------2</link>
            <guid isPermaLink="false">https://medium.com/p/651c5af671b6</guid>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[careers]]></category>
            <category><![CDATA[hiring]]></category>
            <dc:creator><![CDATA[Adil Soukaini | Full-Stack Developer]]></dc:creator>
            <pubDate>Wed, 21 Oct 2026 08:00:10 GMT</pubDate>
            <atom:updated>2026-10-28T15:22:13.946Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*dho0gtHFcgB800Ei.png" /></figure><p>Hiring of junior developers has fallen further than any other part of the software job market since 2022. Payroll data shows employment of US software developers aged 22 to 25 down nearly 20% from its late-2022 peak by September 2025, and new graduates now make up a small share of hires at both large tech companies and startups. Researchers disagree on how much of this is AI and how much is interest rates, but for a small team the practical question is the same: a junior hire works only if someone senior has time to review their work.</p><h3>What the data shows</h3><ul><li><strong>Employment of young developers is down.</strong> In the Stanford Digital Economy Lab’s <a href="https://digitaleconomy.stanford.edu/wp-content/uploads/2025/08/Canaries_BrynjolfssonChandarChen.pdf">payroll study</a> (ADP data), employment for software developers aged 22 to 25 had declined nearly 20% from its late-2022 peak by September 2025, while older age groups kept growing.</li><li><strong>The gap keeps widening.</strong> In the <a href="https://digitaleconomy.stanford.edu/news/canariesaug26/">August 2026 update</a>, employment of 22–25-year-olds in highly AI-exposed occupations stands about 19% below where it would be had it kept pace with less-exposed peers, up from a 15% gap in July 2025. The adjustment happens “primarily through reduced hiring of young workers rather than increased separations.”</li><li><strong>New graduates are a small share of hires.</strong> <a href="https://www.signalfire.com/blog/signalfire-state-of-talent-report-2025">SignalFire’s 2025 report</a> found new grads made up just 7% of hires at the 15 largest tech companies, down 25% from 2023 and over 50% from 2019. At VC-backed startups, they were under 6% of hires, down 11% from 2023 and over 30% from 2019.</li><li><strong>The rebound is senior.</strong> US software development postings have grown almost 15% since early 2025, but <a href="https://hiringlab.indeed.com/2026/07/08/ai-and-job-postings-from-destruction-to-creation/">71% of the increase</a> between May 2025 and May 2026 came from senior roles (Indeed Hiring Lab).</li><li><strong>Graduates are taking other paths.</strong> According to <a href="https://economicgraph.linkedin.com/content/dam/me/economicgraph/en-us/PDF/us-software-engineer-talent-landscape-2026.pdf">LinkedIn’s Economic Graph</a>, 55% of computer science graduates started in non-software-engineering jobs in 2023 and 2024, against 49% in 2016, and entry-level software engineering hiring did not rebound at the end of 2025.</li></ul><h3>Is AI the cause? Two readings of the same trend</h3><p>The Stanford authors argue their results are consistent with generative AI affecting entry-level work. Declines are concentrated in occupations where AI tends to automate tasks rather than augment workers. In a <a href="https://digitaleconomy.stanford.edu/news/canaries-interest-rates-and-timinga-more-on-recent-drivers-of-employment-changes-for-young-workers">February 2026 note</a>, they also tested the interest-rate explanation and found that “more AI-exposed jobs are actually less exposed to interest rates on average.” With stricter controls, they find the decline is significant only after 2024, and they acknowledge that the fall in young developer employment has “various factors,” not AI alone.</p><p>LinkedIn’s economists read their own data differently. They find that software engineering hiring, junior and overall, has tracked the broader tech sector and national hiring, and that changes in real interest rates explain most of the variation over the past decade. They call the lack of an entry-level rebound “concerning for job seekers” but say it is not enough to blame AI.</p><p>Both camps agree on the fact that matters for you: the entry level is weak, and the recovery so far favours experienced people. For the wider market picture, see <a href="https://shipfastdev.com/blog/ai-impact-on-software-developer-jobs">how AI is changing the developer job market</a>.</p><h3>The paradox: juniors gain the most from AI</h3><p>The same tools that may be reducing junior hiring help juniors the most. In randomised field experiments with <a href="https://www.microsoft.com/en-us/research/?p=1148213">4,867 developers</a> at Microsoft, Accenture and a Fortune 100 company, an AI coding assistant increased completed tasks by 26.08% on average, and less experienced developers both adopted it more and gained more.</p><p>There is a catch. In a user study published at the ACM CCS 2023 security conference, <a href="https://arxiv.org/abs/2211.03622">participants with an AI assistant wrote significantly less secure code</a> than those without, and were more likely to believe their code was secure. Participants who trusted the AI less and engaged more with their prompts produced fewer vulnerabilities. A junior with AI can produce a lot of plausible code quickly; without review, that is a risk, not a gain. Our checklist on AI-generated code security lists what to check.</p><h3>Should your startup hire a junior developer?</h3><h3>It can work if</h3><ul><li>You already have a senior developer or technical lead who can review every change and has scheduled time for it.</li><li>The product’s architecture is settled, and the backlog contains well-defined tasks: UI components, tests, internal tools, small integrations.</li><li>You have automated tests and continuous integration, so mistakes show up before production.</li><li>You can wait a few months for the hire to become fully productive.</li></ul><h3>Think twice if</h3><ul><li>You are pre-MVP with no technical lead. A junior should not be the person deciding architecture, authentication or data models.</li><li>Your product handles payments, health data or other sensitive information and no one can review security.</li><li>You need speed now. The research on <a href="https://shipfastdev.com/blog/ai-coding-assistants-productivity-research">AI coding assistant productivity</a> shows gains depend on review and testing, which a junior alone cannot supply.</li></ul><p>If you lack senior capacity, you can rent it: a fractional CTO or development partner can provide the review and direction that make a junior productive.</p><h3>How to make a junior hire work with AI tools</h3><ol><li><strong>Make code review non-negotiable.</strong> Every pull request gets a senior review before merge.</li><li><strong>Ask juniors to explain AI-generated code.</strong> If they cannot explain a block line by line in review, it does not get merged. This keeps learning happening.</li><li><strong>Route sensitive work to seniors.</strong> Authentication, payments, permissions and database migrations stay with experienced people until the junior has shown they handle them well.</li><li><strong>Give them tests to write.</strong> Writing tests teaches the codebase and strengthens the safety net that AI-heavy teams need.</li><li><strong>Track progress.</strong> Watch the number of review rounds per pull request; it should fall over the first months.</li><li><strong>Document onboarding.</strong> A written setup guide and coding conventions help juniors and AI tools alike. See our guide to onboarding a remote developer.</li></ol><h3>The long-term cost of skipping juniors</h3><p>Every senior developer was once a junior. If companies stop hiring at the entry level, the pool of experienced people who can review AI-generated code will shrink over time. For a single startup that is not a reason to hire someone you cannot support, but it is a reason not to dismiss junior candidates outright once you have the review capacity in place.</p><h3>Frequently asked questions</h3><h3>Is it harder for junior developers to get hired in 2026?</h3><p>Yes. Employment of US developers aged 22 to 25 fell nearly 20% from late 2022 to September 2025, and most of the 2025–2026 recovery in postings has been for senior roles.</p><h3>Is AI replacing junior developers?</h3><p>Partly, according to Stanford’s payroll research, which links the decline to AI-exposed occupations. LinkedIn’s analysis points mainly to interest rates and the economic cycle. The cause is debated; the weak entry-level market is not.</p><h3>Should a startup hire a junior developer?</h3><p>Only if a senior person can review their work regularly and the tasks are well defined. Without that, a junior with AI tools can produce code faster than anyone can check it.</p><h3>How should junior developers use AI tools safely?</h3><p>Use them, but treat output as a draft: understand every line, write tests, and get a senior review. Research shows AI users can write less secure code while feeling more confident about it.</p><h3>Sources</h3><ol><li><a href="https://digitaleconomy.stanford.edu/wp-content/uploads/2025/08/Canaries_BrynjolfssonChandarChen.pdf">Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence (working paper, November 2025 version)</a>, Stanford Digital Economy Lab, 2025–11</li><li><a href="https://digitaleconomy.stanford.edu/news/canariesaug26/">Canaries in the Coal Mine: August 2026 update</a>, Stanford Digital Economy Lab, 2026–08</li><li><a href="https://digitaleconomy.stanford.edu/news/canaries-interest-rates-and-timinga-more-on-recent-drivers-of-employment-changes-for-young-workers">Canaries, Interest Rates, and Timing: More on Recent Drivers of Employment Changes for Young Workers</a>, Stanford Digital Economy Lab, 2026–02</li><li><a href="https://www.signalfire.com/blog/signalfire-state-of-talent-report-2025">SignalFire State of Talent Report 2025</a>, SignalFire, 2025–05</li><li><a href="https://economicgraph.linkedin.com/content/dam/me/economicgraph/en-us/PDF/us-software-engineer-talent-landscape-2026.pdf">U.S. Software Engineer Talent Landscape</a>, LinkedIn Economic Graph, 2026–02</li><li><a href="https://hiringlab.indeed.com/2026/07/08/ai-and-job-postings-from-destruction-to-creation/">AI and Job Postings: From Destruction to Creation?</a>, Indeed Hiring Lab, 2026–07</li><li><a href="https://www.microsoft.com/en-us/research/?p=1148213">The Effects of Generative AI on High-Skilled Work: Evidence from Three Field Experiments with Software Developers</a>, Microsoft Research (Cui, Demirer, Jaffe, Musolff, Peng, Salz), 2025–06</li><li><a href="https://arxiv.org/abs/2211.03622">Do Users Write More Insecure Code with AI Assistants? (ACM CCS 2023)</a>, Perry, Srivastava, Kumar, Boneh (arXiv), 2023–12</li></ol><p><em>Originally published at </em><a href="https://shipfastdev.com/blog/junior-developers-hiring-ai-era"><em>https://shipfastdev.com</em></a><em> on October 21, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=651c5af671b6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[OptiVision: a white-labeled Odoo ERP for opticians]]></title>
            <link>https://medium.com/@adilsoukaini/optivision-a-white-labeled-odoo-erp-for-opticians-c837eada9ee5?source=rss-7f614efc1b10------2</link>
            <guid isPermaLink="false">https://medium.com/p/c837eada9ee5</guid>
            <category><![CDATA[erp]]></category>
            <category><![CDATA[odoo]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[small-business]]></category>
            <dc:creator><![CDATA[Adil Soukaini | Full-Stack Developer]]></dc:creator>
            <pubDate>Mon, 19 Oct 2026 08:00:34 GMT</pubDate>
            <atom:updated>2026-10-26T14:45:42.020Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ede2-CxlLU5VdB3-.png" /></figure><p>OptiVision ERP is ShipFastDev’s white-labeled Odoo 16 setup for opticians and eyewear shops: an eyewear catalogue with frame, lens and contact-lens specifications, a point of sale, stock management and sales reporting by eyewear category, all under the shop’s own name and logo. It runs on Odoo Community with two custom modules. It is not public; a demo is available on a call.</p><h3>What an optician can do</h3><ul><li><strong>Keep an eyewear catalogue.</strong> Frames, prescription lenses, contact lenses, sunglasses, accessories and lens care, each product with an “Optician Details” tab next to Odoo’s usual sales, purchase and inventory tabs.</li><li><strong>Record frame specifications</strong>: material, style, colour, frame width, bridge width, temple length, and lens width and height.</li><li><strong>Record lens and contact-lens specifications</strong>: lens type (single vision, bifocal, progressive), material, index and prescription ranges; coatings (UV protection, blue-light filter, anti-reflective, photochromic); base curve, diameter, water content and oxygen permeability for contact lenses; and prescription values (SPH, CYL, AXIS, ADD, prism).</li><li><strong>Flag what a product needs</strong>: whether a prescription is required, whether it is a prescription lens, whether it needs in-house fitting, and its warranty.</li><li><strong>Start with real brands.</strong> The sample data includes frame brands (Ray-Ban, Oakley, Gucci, Tom Ford, Persol), lens makers (Essilor, ZEISS, Hoya) and contact-lens brands (Acuvue, CooperVision, Bausch + Lomb).</li><li><strong>Sell at the counter</strong> with Odoo’s point of sale, products grouped as frames, lenses, contact lenses, sunglasses, lens care and accessories. Odoo documents that its <a href="https://www.odoo.com/documentation/16.0/applications/sales/point_of_sale.html">Point of Sale</a> works on any device with a web browser, even if you are temporarily offline.</li><li><strong>See what sells.</strong> A sales analysis report shows revenue by eyewear category, for example prescription lenses against frames and sunglasses.</li><li><strong>Scan barcodes</strong> for stock adjustments, picking and receiving, with a USB scanner or a phone camera.</li></ul><h3>What was built</h3><ul><li><strong>Odoo 16 Community in Docker.</strong> Odoo and PostgreSQL run with Docker Compose, with a production Nginx configuration alongside. Stock and replenishment come from Odoo’s standard <a href="https://www.odoo.com/documentation/16.0/applications/inventory_and_mrp/inventory.html">Inventory</a> app.</li><li><strong>An optician module.</strong> A custom module adds the eyewear fields, a brand model, an API-key model, the optician views and sample data. Like any Odoo module, it declares its dependencies and data in a <a href="https://www.odoo.com/documentation/16.0/developer/reference/backend/module.html">manifest</a>, so it installs and upgrades on its own, with one documented command to upgrade it.</li><li><strong>A white-label module.</strong> A second module removes Odoo’s branding from the interface and gives the shop its own login page, colours, headers and footers, PDF report branding and email templates. A Branding settings page holds the brand name, tagline, support contact, colours and logos, and each company can set its own login logo, email header logo, PDF logo, favicon and footer text. Its latest version also hides the Enterprise upgrade prompts.</li><li><strong>A REST API for other apps.</strong> Endpoints list products (filtered by type, brand or in-stock status), return product details with images, list brands, check stock by SKU and update stock levels. They are built as Odoo <a href="https://www.odoo.com/documentation/16.0/developer/reference/backend/http.html">web controllers</a>. API keys are managed in Inventory settings, one per app, each read-only or read-write, with usage tracking (last used, request count) and the option to regenerate or deactivate a key.</li><li><strong>Community modules instead of Enterprise features.</strong> Modules from the Odoo Community Association (OCA) cover barcode scanning, barcode generation and <a href="https://github.com/OCA/stock-logistics-barcode">multiple barcodes per product</a>, SQL and visual BI reports, Excel and CSV exports, and approval workflows. For the till, OCA’s <a href="https://github.com/OCA/pos/tree/16.0">point-of-sale modules</a> add margin on POS orders, the discount amount on the cashier screen and ticket, order-line editing, customer comments, lot selection and a default customer.</li></ul><h3>How it is kept reliable</h3><ul><li><strong>Rate limits.</strong> The production Nginx configuration allows 10 API requests per second (burst of 20) and 5 login attempts per minute (burst of 5).</li><li><strong>Security headers and filtering.</strong> HSTS, X-Frame-Options, X-Content-Type-Options and a Content-Security-Policy, plus rules that block SQL-injection patterns, cross-site scripting attempts, vulnerability scanners and path traversal.</li><li><strong>Caching.</strong> Static files are cached for 7 days and images for 1 day.</li><li><strong>API access.</strong> Every endpoint except the health check requires an API key, and updating stock requires a key with write permission. The health check lets monitoring tools see whether the service is up.</li><li><strong>A production checklist.</strong> The README lists what to do before go-live: change the master and database passwords, enable HTTPS, configure the firewall, schedule backups, review API-key permissions and turn on two-factor authentication for admins. Backup, restore and module-upgrade commands are documented.</li><li><strong>Standard Odoo underneath.</strong> The changes live in separate modules rather than in Odoo’s code, so the base system stays standard Odoo 16 Community.</li></ul><h3>Recent releases</h3><p>The repository’s four commits are all dated 25 January 2026:</p><ul><li>The initial project: Docker setup, the optician and white-label modules, the OCA modules and the documentation.</li><li>Docker Compose exposes an additional Odoo port and includes a commented Nginx service for a local reverse proxy.</li><li>The white-label module moved to version 16.0.2, with a fuller debranding description and scripts and styles that hide Enterprise features and upgrade prompts.</li><li>Documentation on where to find the OCA barcode scanner, how to install the alternatives and which Enterprise features they replace.</li></ul><h3>When Odoo fits</h3><p>For the wider question of whether an ERP like this suits your business, see Odoo ERP for small businesses, and No-code vs custom development for when configuration is enough and when custom modules are worth it. More projects are in the <a href="https://shipfastdev.com/portfolio">portfolio</a>, including the <a href="https://shipfastdev.com/work/driveway-morocco">Driveway Morocco case study</a>.</p><h3>Frequently asked questions</h3><h3>Is OptiVision a separate product from Odoo?</h3><p>No. It is Odoo 16 Community with two custom modules (optician fields and white-label branding) and a set of OCA modules. Staff see the shop’s brand instead of Odoo’s in the back office, on the login page and in reports and emails.</p><h3>Can it connect to an online shop or other apps?</h3><p>Yes, through its REST API: products, brands, stock checks by SKU and stock updates, each app with its own key set to read-only or read-write.</p><h3>Does it need Odoo Enterprise?</h3><p>No. It runs on Odoo Community, and OCA modules provide the barcode scanning, reporting, exports and approval workflows that would otherwise come from Enterprise.</p><h3>Can I see it?</h3><p>OptiVision is not public. A demo is available on a call.</p><h3>Sources</h3><ol><li><a href="https://www.odoo.com/documentation/16.0/applications/sales/point_of_sale.html">Point of Sale (Odoo 16.0 documentation)</a>, Odoo S.A., accessed October 2026</li><li><a href="https://www.odoo.com/documentation/16.0/applications/inventory_and_mrp/inventory.html">Inventory (Odoo 16.0 documentation)</a>, Odoo S.A., accessed October 2026</li><li><a href="https://www.odoo.com/documentation/16.0/developer/reference/backend/module.html">Module manifests (Odoo 16.0 developer documentation)</a>, Odoo S.A., accessed October 2026</li><li><a href="https://www.odoo.com/documentation/16.0/developer/reference/backend/http.html">Web controllers (Odoo 16.0 developer documentation)</a>, Odoo S.A., accessed October 2026</li><li><a href="https://github.com/OCA/pos/tree/16.0">OCA/pos, branch 16.0</a>, Odoo Community Association (GitHub), accessed October 2026</li><li><a href="https://github.com/OCA/stock-logistics-barcode">OCA/stock-logistics-barcode</a>, Odoo Community Association (GitHub), accessed October 2026</li></ol><p><em>Originally published at </em><a href="https://shipfastdev.com/work/optivision-erp"><em>https://shipfastdev.com</em></a><em> on October 19, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c837eada9ee5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Stack Overflow Developer Survey 2025: Takeaways for Founders]]></title>
            <link>https://medium.com/@adilsoukaini/stack-overflow-developer-survey-2025-takeaways-for-founders-df34aac741f3?source=rss-7f614efc1b10------2</link>
            <guid isPermaLink="false">https://medium.com/p/df34aac741f3</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[programming]]></category>
            <dc:creator><![CDATA[Adil Soukaini | Full-Stack Developer]]></dc:creator>
            <pubDate>Fri, 09 Oct 2026 08:00:22 GMT</pubDate>
            <atom:updated>2026-11-11T16:08:38.836Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*MHfTWUkmje5K-IqM.png" /></figure><p>For founders, the 2025 Stack Overflow Developer Survey and GitHub’s Octoverse 2025 point the same way: pick mainstream, well-documented tools (JavaScript or TypeScript, Python, PostgreSQL, Docker), assume your developers use AI but check its output, and hire for autonomy rather than office presence. The <a href="https://survey.stackoverflow.co/2025/">survey</a> drew more than 49,000 responses from 177 countries and remains the latest full edition; Stack Overflow said on 30 September 2026 that the 2026 results would follow within days.</p><p>Here are the numbers that matter when you choose a stack or hire, and what to do with them.</p><h3>1. Mainstream languages keep hiring simple</h3><p>Among all respondents to the <a href="https://survey.stackoverflow.co/2025/technology">2025 survey</a>, the most used languages were JavaScript (66%), HTML/CSS (61.9%), SQL (58.6%), Python (57.9%) and TypeScript (43.6%). Python’s adoption jumped 7 percentage points from 2024, which Stack Overflow attributes to its role in AI, data science and back-end work.</p><p>GitHub’s data adds a twist. According to <a href="https://github.blog/news-insights/octoverse/octoverse-a-new-developer-joins-github-every-second-as-ai-leads-typescript-to-1/">Octoverse 2025</a>, TypeScript became the most used language on GitHub in August 2025, gaining about 1 million contributors (+66% year over year), with Python second at roughly 850,000 new contributors (+48%). GitHub calls it the most significant language shift in more than a decade.</p><p><strong>What to do:</strong> unless you have a specific reason, build on TypeScript (front end and often back end) or Python (data and AI-heavy products), plus SQL. You will find developers, libraries and documentation easily, and AI tools have seen plenty of code in these languages.</p><h3>2. Framework choice: popularity is one factor, not the only one</h3><p>The most used web frameworks and technologies in 2025 were Node.js (48.7%), React (44.7%), Next.js (20.8%), Express (19.9%), ASP.NET Core (19.7%), Angular (18.2%), Vue.js (17.6%), FastAPI (14.8%), Django (12.6%) and Laravel (8.9%).</p><p>A smaller share does not make a framework a bad choice. Laravel, for example, has a smaller but established community and a very complete toolkit for business applications. What matters more is that your framework is actively maintained, that you can find two or three developers who know it well, and that it fits your product. Our comparison of <a href="https://shipfastdev.com/blog/laravel-vs-nodejs-for-startups">Laravel vs Node.js for startups</a> walks through the trade-offs.</p><h3>3. PostgreSQL is the default database</h3><p>PostgreSQL was used by 55.6% of respondents, ahead of MySQL (40.5%), SQLite (37.5%), Microsoft SQL Server (30.1%), Redis (28%) and MongoDB (24%). For most startup products with users, orders, subscriptions and permissions, a relational database is the safe starting point, and PostgreSQL is the one most developers already know.</p><h3>4. Containers are standard, cloud choice is open</h3><p>Docker was used by 71.1% of respondents, a 17-point jump from 2024; Stack Overflow describes it as having moved “from a popular tool to a near-universal one.” For cloud platforms, AWS led with 43.3%, then Microsoft Azure (26.3%) and Google Cloud (24.6%).</p><p><strong>What to do:</strong> ask for your app to run in a container from day one. It makes moving between hosts and onboarding new developers much easier. Any of the big three clouds is a defensible choice; we describe one setup in <a href="https://shipfastdev.com/blog/deploy-laravel-google-cloud-run">deploying Laravel on Google Cloud Run</a>.</p><h3>5. AI is everywhere, trust is not</h3><p>In the <a href="https://survey.stackoverflow.co/2025/ai">AI section</a>, 84% of respondents were using or planning to use AI tools, and 50.6% of professional developers used them daily. But 45.7% distrusted the accuracy of AI output, against 32.3% who trusted it, and 66% named “AI solutions that are almost right, but not quite” as their top frustration. 72.2% said vibe coding was not part of their professional work.</p><p>Agents are spreading fast. In its <a href="https://stackoverflow.blog/2026/09/30/getting-ready-for-2026-results-a-look-back-on-developer-survey-findings/">September 2026 look-back</a>, Stack Overflow reports that AI agent usage rose from 31% in 2025 to 59% in an April 2026 pulse survey. Octoverse adds that nearly 80% of new developers on GitHub use Copilot within their first week.</p><p><strong>What to do:</strong> expect any developer you hire to use AI, and ask how they verify its output. The evidence on <a href="https://shipfastdev.com/blog/ai-coding-assistants-productivity-research">AI coding assistant productivity</a> and on <a href="https://shipfastdev.com/blog/ai-generated-code-security-risks">AI-generated code security</a> explains why review and tests matter more than the tool.</p><h3>6. What developers want from a job</h3><p>The <a href="https://survey.stackoverflow.co/2025/work">work section</a> shows where developers actually work: 32.4% fully remote, 17.2% hybrid leaning flexible, 19.9% hybrid leaning in-person, 17.9% in-person and 12.6% free to choose. Fully remote work has declined: the look-back puts it at 38.0% in 2024.</p><p>Only 24.5% of respondents said they were happy at work, 47.1% complacent and 28.4% not happy. The top factors for job satisfaction were autonomy and trust to manage their own tasks, competitive pay and benefits, and solving real-world problems. The share who see AI as a threat to their job rose from 12.1% in 2024 to 15.0% in 2025.</p><p>For US respondents, median salaries were $138,000 for full-stack developers, $175,000 for back-end developers and $145,000 for front-end developers.</p><p><strong>What to do:</strong> a small company can rarely compete on US salaries, but it can offer autonomy, clear ownership and real problems. If you are comparing hiring in-house with other models, see <a href="https://shipfastdev.com/blog/cost-to-hire-a-software-developer">the cost to hire a software developer</a> and our guide to hiring developers in 2026.</p><h3>7. The talent pool is increasingly global</h3><p>Octoverse counts more than 180 million developers on GitHub, with more than 36 million joining in 2025, about one every second. India alone added more than 5 million. The supply of capable developers outside the US and Western Europe keeps growing, which is one reason founders look at <a href="https://shipfastdev.com/blog/nearshore-software-development-morocco">nearshore development in Morocco</a> and similar options.</p><h3>A short stack checklist based on the survey</h3><ul><li>Is each main technology used by a large share of developers, or at least actively maintained with a healthy community?</li><li>Can you name three ways to find developers for it?</li><li>Does it run in Docker and deploy to a mainstream cloud?</li><li>Is your database relational unless you have a clear reason otherwise?</li><li>Does your team have tests and code review in place before relying on AI-generated code?</li></ul><h3>Frequently asked questions</h3><h3>What was the most popular programming language in the 2025 Stack Overflow survey?</h3><p>JavaScript, used by 66% of all respondents, followed by HTML/CSS, SQL and Python. On GitHub, TypeScript overtook all other languages by contributors in August 2025.</p><h3>Which database should a startup use?</h3><p>For most products, a relational database. PostgreSQL was the most used database in 2025 (55.6% of respondents), which also makes it easy to hire for.</p><h3>How many developers use AI tools?</h3><p>84% of 2025 respondents were using or planning to use AI tools, and about half of professional developers used them daily. Trust is lower: 45.7% distrusted the accuracy of AI output.</p><h3>When are the 2026 Stack Overflow survey results released?</h3><p>On 30 September 2026, Stack Overflow said the 2026 results would be published within days. Until then, the 2025 survey is the most recent full dataset.</p><h3>Sources</h3><ol><li><a href="https://survey.stackoverflow.co/2025/">2025 Stack Overflow Developer Survey</a>, Stack Overflow, 2025</li><li><a href="https://survey.stackoverflow.co/2025/technology">2025 Developer Survey: Technology</a>, Stack Overflow, 2025</li><li><a href="https://survey.stackoverflow.co/2025/ai">2025 Developer Survey: AI</a>, Stack Overflow, 2025</li><li><a href="https://survey.stackoverflow.co/2025/work">2025 Developer Survey: Work</a>, Stack Overflow, 2025</li><li><a href="https://stackoverflow.blog/2026/09/30/getting-ready-for-2026-results-a-look-back-on-developer-survey-findings/">Getting ready for 2026 results: a look back on Developer Survey findings</a>, Stack Overflow Blog, 2026–09</li><li><a href="https://github.blog/news-insights/octoverse/octoverse-a-new-developer-joins-github-every-second-as-ai-leads-typescript-to-1/">Octoverse: A new developer joins GitHub every second as AI leads TypeScript to #1</a>, GitHub, 2025–10</li></ol><p><em>Originally published at </em><a href="https://shipfastdev.com/blog/developer-survey-2025-takeaways-for-founders"><em>https://shipfastdev.com</em></a><em> on October 9, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=df34aac741f3" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI Coding Assistants and Productivity: What the Research Actually Says]]></title>
            <link>https://medium.com/@adilsoukaini/ai-coding-assistants-and-productivity-what-the-research-actually-says-12c7d0bd21d6?source=rss-7f614efc1b10------2</link>
            <guid isPermaLink="false">https://medium.com/p/12c7d0bd21d6</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[productivity]]></category>
            <dc:creator><![CDATA[Adil Soukaini | Full-Stack Developer]]></dc:creator>
            <pubDate>Wed, 07 Oct 2026 08:00:14 GMT</pubDate>
            <atom:updated>2026-11-07T11:41:15.022Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*d10HEhJqjPDf0BiQ.png" /></figure><p>The research on AI coding assistants does not give one number. Controlled experiments on well-defined tasks show large speedups (55% faster in GitHub’s 2022 study), field experiments in companies show a smaller but real gain (about 26% more completed tasks), and the best-known randomised trial on experienced developers working in their own large codebases found they were 19% slower with AI. At team level, Google’s DORA research finds more throughput but less stability.</p><p>Here is what each major study measured, why they disagree, and how to use the evidence when you plan a roadmap or judge a developer’s estimates.</p><h3>The main studies, side by side</h3><h3>GitHub’s controlled experiment (2022)</h3><p><a href="https://github.blog/news-insights/research/research-quantifying-github-copilots-impact-on-developer-productivity-and-happiness/">GitHub</a> asked 95 professional developers to write an HTTP server in JavaScript. Those using Copilot finished in 1 hour 11 minutes on average, against 2 hours 41 minutes without it: 55% faster. 78% of the Copilot group completed the task, against 70% without. This is the source of the often-quoted “55%” figure. It measured one greenfield, clearly specified task, which is close to the best case for an assistant.</p><h3>Field experiments at Microsoft, Accenture and a Fortune 100 company</h3><p>A team of economists (Cui, Demirer, Jaffe, Musolff, Peng and Salz) randomly gave some developers access to an AI coding assistant in their normal jobs. Across the three experiments and <a href="https://www.microsoft.com/en-us/research/?p=1148213">4,867 developers</a>, the tool led to a 26.08% increase (standard error 10.3%) in completed tasks. Less experienced developers adopted the tool more and gained more.</p><h3>METR’s randomised trial with experienced open-source developers (2025)</h3><p><a href="https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/">METR</a> recruited 16 experienced developers from large open-source repositories they already maintained. Each of 246 real issues (about two hours each) was randomly assigned to “AI allowed” or “AI not allowed”. The main tool was Cursor Pro with Claude 3.5/3.7 Sonnet. With AI, tasks took 19% longer. Before starting, the developers expected a 24% speedup; after finishing, they still believed AI had sped them up by 20%. METR is careful to say the result does not show AI fails everywhere, only that it slowed these developers on this kind of work at that time.</p><h3>METR’s follow-up (February 2026)</h3><p>METR ran a larger study from August 2025 with <a href="https://metr.org/blog/2026-02-24-uplift-update/">57 developers, 143 repositories and more than 800 tasks</a>. The estimated speedup was 18% for returning developers (confidence interval from a 38% speedup to a 9% slowdown) and 4% for newly recruited ones (from 15% faster to 9% slower). METR itself calls this an unreliable signal: many developers refused to take part because they did not want to work without AI, 30% to 50% avoided submitting tasks they expected AI to speed up a lot, and pay dropped from $150 to $50 an hour. METR is changing its study design rather than publishing a headline number.</p><h3>Google’s DORA report (2025)</h3><p>The <a href="https://blog.google/innovation-and-ai/technology/developers-tools/dora-report-2025/">2025 DORA report</a> surveyed nearly 5,000 technology professionals. AI adoption reached 90%, with a median of two hours a day spent working with AI. Over 80% said AI improved their productivity and 59% said it improved code quality. On delivery metrics, <a href="https://cloud.google.com/blog/products/ai-machine-learning/announcing-the-2025-dora-report">Google’s DORA team</a> found a positive relationship between AI adoption and software delivery throughput, but AI adoption “does continue to have a negative relationship with software delivery stability.”</p><h3>Where the gains show up</h3><ul><li><strong>Self-contained, well-specified tasks</strong>, like the HTTP server in the GitHub experiment.</li><li><strong>Less experienced developers</strong>, who gained the most in the field experiments.</li><li><strong>Team throughput</strong>: DORA links AI adoption to teams shipping more.</li><li><strong>Developer sentiment</strong>: in the <a href="https://survey.stackoverflow.co/2025/ai">2025 Stack Overflow survey</a>, 84% of respondents were using or planning to use AI tools, and 60.7% viewed them favourably.</li></ul><h3>Where they don’t</h3><ul><li><strong>Experts in large codebases they know well</strong>, the METR 2025 setting. Context, conventions and quality standards that live in a maintainer’s head are hard to hand to a model.</li><li><strong>Stability.</strong> DORA’s explanation is that AI increases change volume, and “without robust control systems, like strong automated testing, mature version control practices, and fast feedback loops, an increase in change volume leads to instability.”</li><li><strong>“Almost right” code.</strong> 66% of Stack Overflow respondents named “AI solutions that are almost right, but not quite” as their top frustration, and 45.7% distrust the accuracy of AI output against 32.3% who trust it.</li><li><strong>Self-assessment.</strong> METR’s developers felt 20% faster while measuring 19% slower. Opinions about AI speed are not evidence.</li></ul><h3>Why the studies disagree</h3><p>They measure different things. GitHub timed one task. The field experiments counted completed tasks in developers’ day-to-day work. METR timed real issues in mature projects, with screen recordings. DORA relies on surveys linked to delivery metrics. The populations differ too (professionals on a test task, employees at large companies, open-source maintainers), as do the tools, which have improved quickly between 2022 and 2026. None of these studies is wrong; they answer different questions.</p><h3>What this means for your product team</h3><ol><li><strong>Measure delivery, not impressions.</strong> Track how long changes take to reach production, how often you deploy, how often a release breaks something, and how much work gets redone. If you work with a remote developer, a <a href="https://shipfastdev.com/blog/weekly-releases-with-a-remote-developer">weekly release rhythm</a> makes these numbers visible.</li><li><strong>Put tests and CI in place before pushing more AI-generated code.</strong> DORA’s finding on stability is a direct warning. Automated end-to-end tests, such as the Playwright setup we describe here, are the kind of control system it refers to.</li><li><strong>Don’t cut estimates in half because of a vendor figure.</strong> A realistic planning assumption sits somewhere between the field-experiment gain and no gain at all, depending on the codebase and the developer.</li><li><strong>Budget review time.</strong> More code means more to review, and AI-generated code has its own <a href="https://shipfastdev.com/blog/ai-generated-code-security-risks">security failure patterns</a>.</li><li><strong>Watch for debt.</strong> Code that is “almost right” and merged anyway becomes technical debt.</li></ol><h3>Frequently asked questions</h3><h3>Do AI coding assistants make developers faster?</h3><p>Often, but not always. Studies range from 55% faster on a single controlled task to 26% more completed tasks in company field experiments, while METR’s 2025 trial found experienced developers 19% slower in their own large codebases.</p><h3>Did the METR study show AI slows developers down?</h3><p>For 16 experienced open-source maintainers working on their own repositories in early 2025, yes: tasks took 19% longer. METR does not generalise this to all developers, and its 2026 follow-up was too affected by selection bias to give a reliable new estimate.</p><h3>How much productivity gain should a startup plan for?</h3><p>Plan conservatively and measure. The company field experiments found about 26% more completed tasks on average, with bigger gains for less experienced developers. Your result depends on task type, codebase and testing practices.</p><h3>Should I require my developers to use AI tools?</h3><p>Adoption is already near universal (90% in DORA 2025). The more useful requirement is on outcomes: tests, code review and stable releases, whatever tools are used.</p><h3>Sources</h3><ol><li><a href="https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/">Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity</a>, METR, 2025–07</li><li><a href="https://metr.org/blog/2026-02-24-uplift-update/">We are Changing our Developer Productivity Experiment Design</a>, METR, 2026–02</li><li><a href="https://github.blog/news-insights/research/research-quantifying-github-copilots-impact-on-developer-productivity-and-happiness/">Research: quantifying GitHub Copilot’s impact on developer productivity and happiness</a>, GitHub, 2022–09 (updated 2024–05)</li><li><a href="https://www.microsoft.com/en-us/research/?p=1148213">The Effects of Generative AI on High-Skilled Work: Evidence from Three Field Experiments with Software Developers</a>, Microsoft Research (Cui, Demirer, Jaffe, Musolff, Peng, Salz), 2025–06</li><li><a href="https://cloud.google.com/blog/products/ai-machine-learning/announcing-the-2025-dora-report">Announcing the 2025 DORA Report</a>, Google Cloud, 2025–09</li><li><a href="https://blog.google/innovation-and-ai/technology/developers-tools/dora-report-2025/">How are developers using AI? Inside our 2025 DORA report</a>, Google, 2025–09</li><li><a href="https://survey.stackoverflow.co/2025/ai">2025 Developer Survey: AI</a>, Stack Overflow, 2025</li></ol><p><em>Originally published at </em><a href="https://shipfastdev.com/blog/ai-coding-assistants-productivity-research"><em>https://shipfastdev.com</em></a><em> on October 7, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=12c7d0bd21d6" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>