Elektrine lite

← Feed

Pedro Piñera

pedro@mastodon.pepicrft.me

<p>I’m a Berliner. I enjoy building developer tools and learning new technologies.</p><p>I co-founded <span class="h-card" translate="no"><a href="https://fosstodon.org/@tuist" class="u-url mention">@<span>tuist</span></a></span>, a platform team offered as a service</p><p><a href="https://mastodon.pepicrft.me/tags/Berlin" class="mention hashtag" rel="tag">#<span>Berlin</span></a> <a href="https://mastodon.pepicrft.me/tags/Software" class="mention hashtag" rel="tag">#<span>Software</span></a> <a href="https://mastodon.pepicrft.me/tags/Elixir" class="mention hashtag" rel="tag">#<span>Elixir</span></a> <a href="https://mastodon.pepicrft.me/tags/Openness" class="mention hashtag" rel="tag">#<span>Openness</span></a> <a href="https://mastodon.pepicrft.me/tags/Communities" class="mention hashtag

Posts

  • Post #2575811

    Imagine the day executives of some big tech companies have workforces entirely made of AI agents… will they tell those agents that it’s not about people but empowerment? Will they outsource cheap labour to fix the mistakes of those agents? I have so many questions…

  • Post #1087090

    Not quite the same, but I’ve been observing other businesses in @tuist space and their models, and the way they are run gave me this sad feeling of the people behind them not really caring about the problem space. It’s purely about maximising value extraction as any cost to approach an exit or going public. And this lack of care percolates to the extreme where they even publish case studies of customers they don’t have. https://www.joanwestenberg.com/the-passive-income-trap-ate-a-generation-of...

  • Post #775428

    Coding agents are shipping code faster than ever. When models converge in capabilities, speed is what will set teams apart. Plug your agent into @tuist and get visibility into your Xcode and Gradle builds, tests, and artifacts. Via CLI, MCP, or API. #Xcode #Gradle

  • Post #647259

    We started building Tuist from the &amp;quot;roof&amp;quot; - deep in the toolchain, contributing to tools we don&amp;#39;t own, investing weeks into problems with no direct financial return. Most companies won&amp;#39;t do this. They&amp;#39;d rather abstract everything away and sell you compute. But trust is the real currency in developer tooling. You earn it by showing up and doing the work, not by growth-hacking your way in. https://pepicrft.me/blog/2026/03/02/building-the-house-from-the-...

  • Post #647258

    We published an RFC for Build Reports in Tuist. The idea is to capture build data across Xcode, Bazel, and Gradle so teams can understand why builds got slower or caches stopped hitting, instead of guessing. Each build system keeps its own data model so you don&amp;#39;t lose toolchain-specific signals that matter for debugging. We&amp;#39;re also exploring MCP tools for progressive build analysis. https://community.tuist.dev/t/rfc-build-reports-toolchain-agnostic-build-observability-and-comp...

  • Post #647257

    Published a new post on how we think about developer productivity at Tuist. Three layers: toolchain, compute, and cache. Toolchain is where the real impact is and the most neglected. Compute is the easy sell. Cache breaks the physical ceiling but only works when colocated with compute and accessible from everywhere. Without a framework like this, you end up with a product that is a collection of loosely related tools. https://tuist.dev/blog/2026/03/04/three-layers-of-developer-productivity

  • Post #647256

    This year we decided to do some work on our empty balcony and bought some plants for the spring season.

  • Post #647255

    I’m looking for feedback to expose a new interface to @tuist cache infrastructure https://community.tuist.dev/t/expanding-tuists-cache-infrastructure-beyond-build-systems/935/3

  • Post #647254

    “It&amp;#39;s not that the people building them don&amp;#39;t care. I think most founders do care. But the pressure to grow, to compete, to ship what the market expects, can quietly push the product toward a kind of sameness.” https://pepicrft.me/blog/the-obsession-gap/

  • Post #647253

    We are trying something new at @tuist that I haven’t seen anywhere else but thought, why not? We are moving our discussions around our upcoming marketing videos to the marketing forum: https://community.tuist.dev/t/marketing-video-introducing-mcp/938/1

  • Post #647252

    “A chef who has spent years working with ingredients, who understands the subtleties of heat and timing and texture, who loves the act of cooking itself, will always produce something different from someone following the same instructions for the first time.” https://pepicrft.me/blog/the-beauty-of-being-open/

  • Post #647250

    Made some progress today in my balcony garden. Planted basil, cherry tomatoes, camomile, and mint. Fingers crossed 😊

  • Post #647249

    Really liked this blog post about OpenClaw. I have to say there was something that felt magic from seeing a bunch of pieces together interfaceable from Telegram, but my usage decreased over time to the point that my only use case for it is something I can already do with Claude and/or ChatGPT: https://entropytown.com/articles/2026-03-12-openclaw-sandbox/

  • Post #638404

    With so many improvements and new features landing on @tuist, you might want to subscribe to our product newsletter. We send a curated update every 6 weeks, right on your inbox: https://tuist.dev/newsletter #Swift #iOSDev

  • Post #638401

    Kicking off the last dev cycle of 2025 at @tuist 🎬 Big focus on making cache nodes self-hostable so you can run them close to your CI, office, or even AI coding environments. Latency is going to matter a lot in the Xcode 26 CAS world. Also working on test flakiness detection and some nice improvements to Previews. https://community.tuist.dev/t/plans-towards-the-end-of-the-year-in-the-cycle-2025c6/833

  • Post #638399

    Just published our first public case study. monday.com has been using @tuist since 2022. They went from a monolithic Xcode project with constant merge conflicts to 200 modules that any engineer can create confidently. CI dropped from 20+ minutes to 9. This is exactly why we built Tuist. First of many stories to come. https://tuist.dev/case-studies/monday

  • Post #638393

    We used Codex to migrate Mastodon iOS to @tuist Clean builds went from 110.8s to 22.3s (80% faster). The agent handled the full feedback loop: extracted build settings, mapped targets, fixed errors, debugged runtime crashes. Manual migrations take days. This took hours. The best output? A reusable skill file so other teams can migrate without starting from scratch. Try it on your project: https://tuist.dev/blog/2026/02/02/migrating-mastodon-ios-to-tuist-with-codex

  • Post #500277

    https://nesbitt.io/2026/02/26/git-in-postgres.html