Michael Novati

You might be paying a full engineer's salary without getting anything in return. This will benefit the Tech leaders and the executors. Problem Weight = Problem Effect * Volume of Effected As technologists, we're constantly solving problems. But it becomes a "Grind" when we're never pausing to make the judgment if a specific problem is worth solving, if it's the highest priority to use our labour on. The equation on top is a simple way to understand the weight of a problem you're solving, and helps to guide which problem-opportunities are worth pursuing out of the infinite realms of problems captured by your mind's vision system. Here's a simple example: Your developer environments' data queries are slow; let's say it takes 30 seconds. Is this an impactful problem that you should dive headfirst? It depends. Let's look at it: The impact of the problem: 30 seconds lost on a data query. If, on average, you run the query 10 times, you're losing 3 minutes on the query. And every 30 seconds increases the probability to come out of the zone of focus, let's say the cost per 30 seconds of distraction is 5 minutes on average. We often make the mistake of calculating the weight of the problem without an understanding of human psychology. So, the cost of this delay on the workflow is around 5 * 10 = ~50minutes If we round up for easy calculation, it's 1 hour per day. For a realistic range, it'll be around 30mins - 60mins. So, problem impact = ~45mins ( avg: (30 + 60)/2 = 45mins) Now, if you have 10 engineers working on the project, you have the problem weight of: Problem Weight = 45mins * 10 = 450 mins. That query delay is costing you 7.5 lost engineering hours per day. So, due to this simple 30-second delay, you don't have 10 team members, you actually have 11 team members (you're paying ~8hours/day), and one of them is doing nothing. Would you pay someone the engineer's salary for doing nothing?

3

Why is Anthropic hiring engineers like crazy, even though Claude is writing all the code? Boris is the head of Claude Code, this was his response: They still need engineers to prompt, talk to customers, prioritize work, and coordinate with other teams To most people, that sounds more like the scope of a PM But of course, a high-judgment engineer is great at these things as well The lines between job descriptions are blurring

382

32 Comments

The strongest senior engineers rarely wait around in a slow hiring process. They pull out. Quietly, and usually for someone who moved faster. Managers tend to underrate this. A long, careful process feels like protecting quality. To a strong senior engineer it reads as a signal: slow decisions, unclear ownership, a place where things drag. And the better they are, the faster they draw that conclusion, because the better they are, the more options they have. In the Netherlands, scarce IT roles already sit open 56 days and longer. Every extra interview round, every week of silence, is not neutral. It is time your best candidate spends saying yes to someone else. The hard part is that you rarely see who you lost. The mediocre applicants wait, because they have to. The ones you actually wanted are simply gone, the role stays open, and it looks like a supply problem. Often it is a process problem in a supply problem's clothes. This is not about rushing or lowering your bar. It is about a process that signals, at every step, that a good decision lives here. When you look back at a hire that dragged, who went quiet first, and what was your process doing that week? Recognise this in your own pipeline? Happy to think it through with you, you know where to find me.

26

23 Comments

This post made me think about AI & podcasting - why do we use the word AI as a catch-all term? Who's it easier for? And what are we sacrificing without the specificity? Instead of being a podcaster who uses: - a local shaping algorithm for audio improvement - an LLM-based transcription tool - an LLM for processing and analysis of transcript, and; - a machine learning algorithm for social media discovery ...we say, "I use AI in my podcast." AI is a marketing term to entice people to use products. AI is a catch-all term to ensure a program or process is understood. Our post-production process doesn't need to be simple to understand. And in response to the question, "Do you use AI in your podcast?", why should the answer be Yes or No? I give you permission to be specific. Specificity is hard when you're doing text-based editing in Descript, but if you use a variety of tools with varying functions & algorithmic capacities, be specific. And to Stephanie Fuccio's point, lets start categorising and explaining what we use to assess its impact. If I make your Zoom call sound better, its not the same as a 3,000 episode per week production schedule with no human oversight. And if your apps are starting to stuff their packages with AI-based tools, explore what, why and how to figure out what's happening in there.

3

1 Comment

Hot Take or Unpopular Opinion? There is no real difference in Creators using AI. No, seriously. Do you want to philosophically compare the difference between a Creator sitting at a conference table discussing content calendars and directing teams of employees to crank out their bits and get them all posted in the right place at the right time with a lone Creator doing the exact same thing, but instead of a Team they have AI? What's the difference between the two? Cashflow? Privilege? I love listening to a podcast in which a creator is being interviewed for going from Nothing to Something in an amazing timeline, and getting to the part wherein they are going nowhere real fast and then the lightbulb lit and somehow suddenly had a dozen employees carrying them to success. For creators without that ability to just suddenly dump loads of cash into monthly wages, etc. I think AI is a great equalizer. For those with jobs and families and other obligations, it might be the one doable, achievable, realistic way to break into those lofty ranks of the pros. Sure, we risk all kinds of slop and disorder and chaos, but we already have that problem, thanks to Socials that reward such behavior and teams that are only too happy to keep the cash rolling in regardless. Is it up to us to honorably use such power? Yes, but the same could be said of the huge teams carrying their Creator figureheads. This isn't particular to videos either, with the same going on in spades for Kindle and I'm sure other media that I'm not participating in so have no idea. I could probably vent for hours, but suffice it to say, I'm in general in favor of AI for Creators. Pic Disclaimer: Made by Gemini, ironically enough.

3

1 Comment

🔥 Curiosity about useState drove me into the world of Fiber and reconciliation I started learning about the useState hook and how the React team actually built this thing. It looks like everything is tightly connected with Fiber. The hook logic lives inside Fiber, and the whole idea of custom hooks is just to reuse logic anywhere you want. The only rule is the function name must start with use*, but inside we just write normal JavaScript logic. When I checked how useState is implemented, I noticed they use Fiber heavily. Then I fell into the rabbit hole of Fiber and reconciliation. Very interesting stuff. Before React 16, they used something called stack reconciliation. It was fully synchronous, no priorities, no way to pause, and everything executed through the JavaScript call stack (LIFO), which caused a lot of problems for big apps. Then Fiber came and completely changed this. Instead of relying on the JS call stack, React now has its own Fiber nodes. Whenever you update something, React creates a separate work-in-progress Fiber tree in the background (a detached copy of the current one). After all the work is done, React commits it, and the Fiber root starts pointing to the newly committed tree. That’s basically how modern React avoids blocking the UI and handles priorities. People still say React “Virtual DOM” because it’s easy and they don’t understand Fiber architecture or how React really works internally after React 16 btw latest version is (19.x). Pretty interesting right ? #ReactArchitecture #DeepDive #TechExplained #ReactFiber #ReactReconciliation

3

1 Comment

I just published a new piece on RunnerQ: Durable Execution for Background Work in Rust (Without a New Platform). I’ve lost background jobs to restarts before, and I got tired of “fire-and-forget” being the default reliability model. RunnerQ is my take on durable background activities: persisted work, leases + reaping, retries with backoff, dead-lettering, idempotency, and observability all embedded as a library. In the article, I walk through: - what “durable execution” means in practice - RunnerQ’s architecture (engine, handlers, storage traits) - the activity lifecycle: pending → processing → retrying → completed / DLQ - how to adopt it with minimal code changes If you’re building Rust services where background work must not disappear, this will resonate. And even if you aren’t, you’ll still find it useful. #Rust #DistributedSystems #DurableExecution #ReliabilityEngineering #Backend

30

4 Comments

The most important investment in agentic coding velocity may have nothing to do with agents. Amplitude tripled its PR volume in six months. PR cycle time went from 5.2 hours to 44 minutes. Almost all of that came from fixing unglamorous infrastructure. 𝐂𝐈 𝐟𝐫𝐨𝐦 𝟑𝟎 𝐦𝐢𝐧𝐮𝐭𝐞𝐬 𝐭𝐨 𝟑: they swapped the JavaScript toolchain for native equivalents (tsc to tsgo, ESLint to oxlint, Prettier to oxfmt), moved to affected-only builds in Turborepo, and ran independent stages concurrently. Most tools got about 10x faster. 𝐅𝐮𝐥𝐥 𝐬𝐭𝐚𝐜𝐤 𝐮𝐩 𝐢𝐧 𝟐𝟖 𝐬𝐞𝐜𝐨𝐧𝐝𝐬: a custom CLI built on devbox starts every local service with one command, down from 3 minutes.  𝐕𝐚𝐥𝐢𝐝𝐚𝐭𝐢𝐨𝐧 𝐰𝐢𝐭𝐡 𝐧𝐨 𝐥𝐨𝐜𝐚𝐥 𝐬𝐞𝐭𝐮𝐩: add a label to a PR and get a frontend preview in under five minutes, or an ephemeral copy of the backend stack in the cloud. Once you have a full app stood up in 30 sec on every PR, testing and verification become much faster which decreases time to merge.  𝐑𝐞𝐯𝐢𝐞𝐰 𝐬𝐜𝐨𝐫𝐞𝐝 𝐛𝐲 𝐫𝐢𝐬𝐤: a model rates each code change on size, scope, test coverage, and API surface, and low-risk changes merge with no human. Cursor's Bugbot reviews every PR, and bug reports fell 55% even as PR volume tripled. Once CI dropped under five minutes, engineers stopped working locally. Remote is the new dev environment.

27

2 Comments