Michael Sadahiro

Former GitHub CEO - Thomas Dohmke just came out of stealth with Entire : first open-source CLI called Checkpoints that captures agent context alongside code every time you commit Thomas Dohmke, former GitHub CEO, raise $60M in seed funding for his new startup Entire — now valued at $300M — to rethink how we build software in the age of AI agents. This isn’t just another dev tool — Entire is launching its first open-source CLI called Checkpoints that captures agent context alongside code every time you commit. That means the why, how, and what behind AI-generated code won’t be lost in chat logs or ephemeral sessions anymore. Why this matters for the AI-driven future of software development: Context as first-class data: Entire CLI hooks into your Git workflow and stores session history (prompts, responses, files changed, reasoning) as versioned checkpoints alongside code — making agent work traceable, reviewable, and auditable. Bridging human + agent workflows: Traditional tools assume humans write code — but with AI agents writing more every day, we need tooling that treats agent reasoning as part of the development lifecycle. Entire is setting that foundation. Open ecosystem & transparency: By launching this as open source, Entire is giving the community a way to build standards for agent context tracking rather than locking this capability in proprietary platforms. This is a huge milestone in the evolution of developer experiences — turning AI agent output from opaque artifact into traceable, collaborative process that humans can understand and improve. If you’re working with AI assistants, multi-agent systems, or just thinking about how to scale developer productivity in 2026 and beyond, this is worth a close look. https://lnkd.in/gZ8rEm9s #AI #DeveloperTools #OpenSource #SoftwareEngineering #CLI #AIagents #SoftwareContext

1

“Developers are a resource. They don’t talk to product people.” I overheard that on Seattle’s light rail, and it stuck with me. I make it a point to talk directly with the people who use the software — delivery drivers, paralegals, and others doing the real work. Every conversation saves me hours of coding the wrong thing. When developers are treated as “resources,” they’re cut off from the feedback that makes their work valuable. Talking to users isn’t a luxury — it’s how you build the right thing faster. You don’t need a big product team or AI tools for that. Just curiosity and a conversation.

246

42 Comments

I always tell devs: if you're stuck on what to build with AI, reverse-engineer real tools. Try this: - Look at GitHub Copilot—what does it automate? - Use Grammarly—how does it rewrite? - Play with Notion AI—what prompts power it? Then, recreate a tiny version of one. That’s the best way to learn.

3

GitHub just moved AI coding from autocomplete to architecture. Copilot Workspace now handles the entire development cycle—from issue to pull request. Early adopters report 55% faster feature completion. That's not incremental. That's structural. What this means for technical hiring: The value hierarchy is shifting. Junior developers who primarily translate requirements into boilerplate code face compression. Mid-level engineers who understand system design and can validate AI-generated solutions become more valuable. Senior architects who can decompose complex problems into AI-actionable tasks are suddenly force multipliers. This isn't about AI replacing developers. It's about redefining what "coding skill" means. The bottleneck is moving from syntax to judgment. For hiring managers: Start evaluating candidates on problem decomposition and code review capabilities, not just greenfield implementation speed. The engineers who thrive will be those who can effectively collaborate with AI tools as thought partners. For developers: If you're spending most of your time on implementation patterns that Copilot can generate, invest in the layers above and below—requirements analysis and architectural decision-making. The technical interview that asks candidates to write a sorting algorithm from scratch is becoming less predictive. The one that asks them to critique an AI-generated solution and identify edge cases? That's your signal. We're not losing coding jobs. We're compressing the implementation layer and expanding the strategy layer. Adapt your hiring lens accordingly.

4

𝗜 𝗶𝗻𝘁𝗲𝗿𝘃𝗶𝗲𝘄𝗲𝗱 𝗮 𝗯𝗿𝗶𝗹𝗹𝗶𝗮𝗻𝘁 𝗰𝗼𝗱𝗲𝗿. No degree. No credentials. Just raw talent, GitHub receipts, and more grit than half the team 𝘤𝘰𝘮𝘣𝘪𝘯𝘦𝘥. The hiring manager said: 🧊 “We only take CS grads.” I could’ve moved on. Instead, I pushed back. ⚔️ “He’s the best technical mind we’ve seen so far. If a diploma matters more than delivery, what are we even hiring for?” They took the chance. He got the job. 🔥 Today? He’s leading the team that once wouldn’t have considered him. That was the day I stopped being a gatekeeper. And became a bridge. Let’s kill the myth: A degree ≠ competence. A lack of one ≠ risk. What actually matters? ✅ Curiosity ✅ Grit ✅ Output ✅ Coachability ✅ The actual work We’ve hired too many “qualified” candidates who couldn’t ship a thing. And passed on brilliant minds because they didn’t follow the traditional path. If you’re still filtering by degree in 2025… You’re not hiring the best. You’re hiring the most conventional. 🎓 Degrees are optional. 🛠️ Grit and skill are not. 👉 If you care about hiring beyond resumes and seeing talent for what it really is, follow Natalia Kufel. I share stories, lessons, and a more human approach to recruiting. Proactive Global #HiredByNatalia #ProactiveGlobal #HiringDifferently #RecruiterLife #TalentOverTitles #CSNotRequired #RealTalkRecruiting #TechHiring #HiringTruths #NonTraditionalTalent #FutureOfWork #SkillsOverDegrees #CareerPaths #InclusiveHiring #HumanFirstHiring #GritOverPedigree #LinkedInRecruiter

28

9 Comments

Amazon you know you had it coming I vowed to troll you for ALL BS layoffs. Because the best thing that Andy Jassy could ship is the Kindle in color and mass firings. 🫏🫏🫏🫏 #softwareengineers #tpms #pms #qa #engineeringmanagers: I’m going to say this many times but this time I’m going to yell it 🎤🎤🎤🎤🎤 IF A COMPANY EVER ASKS YOUR ORG TO DOCUMENT YOUR ENTIRE WORK FLOW YOUR JOB IS TO DOCUMENT IT SO BADLY THAT THE COMPANY WILL GO OUT OF BUSINESS IN 2 WEEKS. I do not understand why people still simp over AI. 🫥🫥🫥 While they tell you to your face: We are LITERALLY taking your work AND then your whole JOB and CAREER. This. Is. Not. A. Drill. 🚨🚨🚨 I don’t care what you do. Put it all in Perl. Vibe code something so bass ackwards that every customer will literally get a picture of the CEO laughing and saying “I fired the person that made this to make myself more money if they push a button.” Make every API call sign your customers up for “More information about Scientology.” Sabotage it. Make plans with your fellow engineers to make sure it is the worst decision that they have ever made. Turn this around and make AI as incompetent as the CEO who decided to fire you after stealing your entire personal IP. I believe in you. You are an engineer. You can F up anything. And yes, this specific stink eye is for you, Prime Video & Amazon MGM Studios - wow. 🫡🫡🫡🫡🫡🫡🫡🫡🫡🫡🫡🫡🫡🫡🫡🫡

21

4 Comments

If you're a Senior Engineer looking to move up, the next role will likely be a Staff Engineer. And in this guide, Shruti shares tips from her own experience of getting promoted to Staff Engineer at PayPal and Slack. She talks about what Staff Engineers do (and how it's different from Seniors), why you might not be getting promoted, and how to take that next step. https://lnkd.in/g58dnFEG

103

3 Comments

This week I shipped two micro-frontends from prototype to production as part of Upwork's platform decomposition into composable smart surfaces. The job search browser went from zero to merged with real pagination, sidebar filtering, accessible scroll-to-top, and reactive search state. The membership manager was the heavier lift: 4 parallel GraphQL queries with graceful degradation, a full API transform layer normalizing raw schema into clean store state, reworked modals, alerts, tracking, and i18n — roughly 1,750 lines across 36 files. Both run as independent SPAs injected into the platform shell, each with its own Pinia store, error boundaries, and analytics instrumentation. What I keep learning at this scale is that the hard part isn't building the feature — it's designing the boundary. Where does the microapp end and the shell begin? Which queries are critical enough to block render, and which degrade silently? How do you instrument analytics without coupling to the host? These are the decisions that determine whether a micro-frontend architecture actually scales or just moves the monolith into smaller boxes.

3