Rudolf Olah

Pull Requests shouldn’t feel like this. We’ve all opened a large PR review and thought: “Well... this is going to hurt.” Sure, small PRs are great, but sometimes you need the full picture. That’s where stacked PRs help (been using Graphite for that, a tool for managing PRs and code reviews). Stacked PRs treat your feature development as a chain of commits / PRs layered on top of each other. Think of it like a Jenga tower: each block (PR) builds on top of the previous one. pr1 → main pr2 → pr1 pr3 → pr2 Now it’s getting even better. I just got early access to Graphite Chat 👇 - I can literally ask it “what’s going on in this PR?” and it gives me the full story, codebase history, past PRs, even outside context. - I can just highlight it and ask “explain this”, instant breakdown. - Flags risks before reviewers even show up. - When someone leaves feedback, it helps me make sense of it and suggests fixes. - And the best part? When CI blows up, I can just ask “why did this break?” and actually get an answer. I’ll even be able to pop open the Graphite Editor right inside Chat to view and edit the codebase. Reviews feel lighter, more collaborative. Exactly how they should.

1,573

141 Comments

If you're looking for a new docs framework for your dev tool, I can't recommend Docusaurus enough. 1. You can change the appearance of any component—no matter how big or small—by ejecting its raw React code into your source tree. I've ejected almost all of its major components and restyled them with Tailwind. 2. It supports content written with MDX, which is like Markdown but better. MDX lets you write JSX tags so you can render React components (like badges and banners) directly inline your content. No more messy templating. 3. It bundles Prism for code snippet highlighting. So you if you want to change how each and every token of your language gets styled—which I do—you can override it with a simple centralized map—which I have. 4. You can swap out its on-site search engine with Algolia, now widely considered the best option for developer sites. It is, in other words, the perfect tool for people like me who are obsessed with every pixel.

118

16 Comments

Here are 5 NEW companies paying $100k-$400k for FULLY-REMOTE devs: 1. Circle 2. The Linux Foundation 3. Form3 4. Assured 5. Qualio They get it. - They don't have the FAANG brand. - They even pay less than FAANGs in the US. Yet, they're able to easily hire and retain FAANG-level talent. Want to know about other companies in this category? EuroTopTech.com lists thousand jobs from 100+ such companies. It also tracks thousand top-paying onsite, hybrid and country-remote jobs. In Europe. It also provides data regarding compensation and saving rates. Check it out if it sounds interesting 🙂 ♻️ Like/Comment/Share if you want to help people in your network.

232

18 Comments

Nobody talks about it, but long PR review times is a huge cause of slow engineering teams: teams that review PRs in <3 hours are 2.1x more productive than 8+ hours teams. Looking at 500K PRs opened by 3,200+ engineers, here's the data: - 25% get reviewed in <1 hour (amazing) - 31% in 1-3 hours (great) - 34% in 3-8 hours (getting slow) - 10% wait 24+ hours (what are you even doing?) Your team should stop treating PR reviews like something you'll "get to later." I love Google's guidelines: "If you are not in the middle of a focused task, you should do a code review shortly after it comes in." P.S.: If you are curious about the output - those are based on our own calculation of the complexity of the PRs (0.95 correlation with what engineers think - much higher than lines of code or any other method).

109

46 Comments

"My developers feel like they're flying blind." A CTO of a fast-growing fintech company told me recently. CTO: "To test a simple change to one microservice, a developer has to file a ticket with the Infra team to get a test environment configured. It takes 30-45 mins. They lose all momentum." Me: "So they're blocked, waiting on someone else just to do their core job?" CTO: "Yes. And the infra team is drowning in tickets. It's a lose-lose. The alternative is letting developers run the 40-service stack on their laptops, which is impossible." This dependency is a silent killer of productivity and innovation. When it's hard to test, developers take fewer risks, batch changes into massive PRs, and avoid refactoring. The strategic shift is to decouple developers from the underlying infrastructure. Give them a self-service platform where they can create their own isolated sandboxes with a single command or click. By leveraging request isolation in a shared environment: - Developers can instantly create an environment to test their specific microservice. - They get the benefit of the entire backend stack without the complexity of managing it. - The infrastructure team is no longer a blocker and can focus on building the platform. The result is developer autonomy. - Environment provisioning time: From 30+ mins to under 1 minute. - Developer dependency on infra team: Eliminated. - PR size: Decreased by 30% as developers could iterate faster. The CTO’s insight: "We've been trying to solve a developer problem with infrastructure tools. We need to give them a developer-first workflow." How are you empowering your developers to test their changes independently and quickly? #DeveloperExperience #PlatformEngineering #DevEx #EngineeringCulture #Microservices #CloudNative #DevProductivity #EngineeringLeadership

72

24 Comments

People are starting to figure it out. Vibe coding isn’t a magic bullet that turns billions of users into expert-level software engineers. Still, companies that figure out how to make vibe coding work will be far more successful than companies that stay focused on what it can’t do. A vibe coding startup called Dfinity claims it’s ushering in the “self-writing internet.” As this poster found out, vibe coding is great for building demos, but those don’t meet the standards of a production environment. When defects emerge or requirements evolve, vibe coding isn’t up to the challenges of iterating, fixing, and maintaining user-facing apps. Experts are still required, so keep them in the loop to start seeing value from vibe coding. Vibe coding apps make tiny mistakes, and those add up over time. Expert developers are required to fix mistakes as they go, so they don’t compound into cascading failures. Context length is another challenge, so the more code you have, the worse vibe coding apps get at generating code that integrates well with the existing code base. The more AI changes to resolve integration issues, the worse things get. Expert developers are required to integrate generated code into complex apps without breaking more than they’re improving. We should definitely be adopting vibe coding tools and reengineering workflows to maximize the amount of code they create for us. However, their capabilities only support an augmentation layer for expert developers. Keeping up with the leading edge of AI and gaining first-principles expertise is simple. #HighROIAI - https://lnkd.in/ghBwMVGj #UnhypedAI - https://lnkd.in/gDDJhXDJ Turn on notifications for my content by ringing the bell in my profile.

146

31 Comments

A big customer asked how to measure the business impact of Amp vs. other coding agents. Their Amp usage is growing crazy fast, and they want to make an informed decision. My answer was a bit surprising: This is all brand new. Amp and Claude Code (which are the latest generation of coding agents) have been widely used for ~60 days. Things are changing so fast and the impact has been so big that there has not been time or need for a set of measurement best practices to emerge. When orgs are adopting these tools, and they are adopting these rapidly, they are doing so because of very strong signals, usage, and feedback from devs who are trusted by the decision-makers. Amp’s team pages show you who’s using it, how much, and how often. And as the only coding agent with team thread sharing, only Amp lets you see what people are actually using it for (not just how much they're using it). Unlike the previous generation of seat-pricing-based coding assistants, Amp is purely usage-based pricing and doesn’t require you to make a big up-front $$/seat commit, so the decision here need not be as big as in the past. In fact, we are seeing a lot of customers who are already switching away from other seat-based coding tools they prepaid for a whole year for just ~2 months ago after a heavy impact measurement process that in hindsight yielded the wrong decision. Naturally they are nervous about making another big 12-month commitment. What we’ve seen work the best is to give devs access to a few tools, see what they like and use the most (based on # of threads, tokens, and daily usage), and then make those well-loved tool(s) more broadly available quickly. If that means staying pay-as-you-go or making a smaller enterprise commitment (such as for only 1-3 months of estimated usage, not a whole year), that is fine and all of the agentic coding tools that I’m aware of, Amp included, will allow that.

110

3 Comments

Proposal: move RubyGems (the gem and bundler CLI tools) to the same Ruby org that governs the language itself. It's an accident of history that Ruby, its dependency tools, and its dependency hosting are managed by three separate entities. (And it hasn't gone great.)

54

3 Comments

Hard Truth: Most "real-time" systems aren't actually real-time Here’s a hard truth: 90% of systems claiming to be "real-time" are lying to themselves and their users. If your worst-case execution time (WCET) isn’t measured, documented, and guaranteed, you’re not building real-time systems, you’re building hope-and-prayer systems. We’ve all seen it, RTOS-based projects where developers claim "real-time" capability but have never actually measured interrupt latency or task jitter. They’ll proudly show you their 1ms task scheduling but can’t tell you what happens when a cache miss coincides with a DMA transfer. I’ve debugged "real-time" systems that missed deadlines by 100ms under heavy load because someone used malloc() in an ISR. I’ve seen medical devices that claimed 10ms response times but actually had 200ms spikes when garbage collection kicked in. And I’ve watched aerospace systems fail certification because their "real-time" claims were based on average-case timing, not worst-case. Real-time isn’t about speed, it’s about determinism. It means knowing exactly how long every operation will take in the worst possible scenario. It means measuring, not assuming. It means understanding your hardware’s timing characteristics at the cycle level. 🔥 What's the most ridiculous "real-time" claim you've ever seen? Share your WCET horror stories below, bonus points for systems that failed spectacularly in the field! #EmbeddedSystems #RealTime #RTOS #WCET #Firmware #EmbeddedC #Determinism #TechTruth #EmbeddedEngineering

251

71 Comments

Most employees don’t get to sit in on the performance calibration meetings that shape careers behind closed doors. If they did, I think a lot of the mystery—and fear—around stack ranking would go away. In my experience, most managers genuinely try to do the right thing. But when you’re asked to identify the “bottom 10%” across an organization of hundreds, the process gets messy. You don’t just look at individual teams—you zoom out, compare ratios across groups, look at tenure, stalled growth, repeated coaching efforts, and yes, hear anecdotes from peers and HR. I wrote a detailed piece on how this works, including myths, common questions, and what to do if you find yourself on the wrong side of the process. It’s not light reading, but it’s probably one of the more useful things I’ve written for anyone navigating corporate performance management.

38

1 Comment

Slashing on EigenLayer isn’t just about penalties anymore, it’s becoming programmable. With Othentic framework, builders can now define custom, onchain slashing logic tailored to their app’s trust model. It’s a whole new design space for permissionless accountability. Read this detailed blog to dive into the framework, the mechanics & what it unlocks ↓ https://lnkd.in/g5tGiKq8 #Othentic #EigenCloud #EigenLayer #Slashing #Restaking #Ethereum #Web3 #TechnicalBlog #Blockchain #Web3Simplified

40

2 Comments

CSV exports of issues now include all custom issue fields. If you've added custom fields like... → Product area → Bug severity → Question type ... those same fields will now appear in your CSV export. This means no more manual reformatting when you're pulling data for leadership reports or quarterly reviews. Enjoy!

28

5 Comments

Every quarterly update has asks, for intros to candidates, customers, partners. Yet 99% of people write: "We're looking for senior engineers" and wonder why their network doesn't respond. Here's the fix: Include copy-paste descriptive blurbs for every ask. Instead of: "We're looking for senior engineers" Write: "We're looking for senior engineers. Here's a blurb you can forward: 'Hi [Name], I'm an advisor to [Company], a Series A startup building [thing]. They're looking for a senior engineer to lead their infrastructure team. Remote-friendly, $180-220k + equity, working directly with the CTO. Happy to make an intro if you're curious—the founder is exceptional.'" Your supporters WANT to help. But turning "we need engineers" into an actual intro email takes 15 minutes they don't have and you don't want them to butcher how they pitch your business anyways. Give them something they can copy, paste, and send in 30 seconds. Many of the founders I work with get texts from me to the same effect, "can you send me a blurb?" because the reality is that I don't know how you want to position yourself with a potential partner or candidate better than you do. Try it in your next update and let me know what happens! I'll certainly be copy-pasting! 🍝

61

11 Comments