I Broke Up with JSON (And My APIs Are 5× Happier Now)
A good run down of API data exchange formats that are faster to serialize/parse than JSON.
Bookmarks RSS feed
-
-
Nobody Cares How Hard You Work
Treating your career in terms of product-market fit is an intriguing way of improving your indispensability at work. You match your skills to a pain point, not a role. Only then hard work starts to matter. A good reinterpretation of the Japanese concept of ikigai, IMO.
-
How Salesforce Used AI To Reduce Test Failure Resolution Time By 30%
Not sure how the described Test Failure Triage Agent is different from RAG search. It’d be more interesting to learn how Salesforce used RAG at scale. Nonetheless, using vector-based similarity search is definitely an interesting solution for automating triaging and suggesting solutions for a massive amount of test failures in CI/CD pipelines.
-
The Productive Discomfort Zone: Why Real Career Growth Feels Awful
What a brutally honest advice to owning your career growth. Interesting overview about “the Dip” and its nuances. The tip about high vs. low agency at the end is just 🫡
-
How Flipkart Built a Highly Available MySQL Cluster for 150+ Million Users
A practical look at how primary database election works (in a replication setup) during failovers. What a simple and clever design! Love how monitor and orchestrator components are kept separate to reduce false positives.
-
State of the software engineering jobs market, 2025: what hiring managers see
Pretty much what I imagined about why cold-applying to job isn’t working anymore.
-
How Airbnb Runs Distributed Databases on Kubernetes at Scale
Distributed SQL databases are not magic boxes that will handle availability and reliability challenges without thoughtful engineering. Airbnb’s case study is a great example of such engineering: don’t let more than one node replacement run at a time (serialized replacement controlled by custom k8s controller logic), spread a database across data centers (AWS availability zones), and don’t store data on the same node as DB server (detachable AWS EBS volumes).
-
- YouTube
As a long-time developer, I can completely related to CJ’s rants against AI coding tools. Are they really making us 10x productive when we have to spend a non-trivial amount of time understanding, reviewing, re-prompting, and fixing AI generated code?
-
An Engineer's Guide To Cultivating Confidence
If there’s only one thing I am taking away from this article, it is to reverse the feeling-action order. Instead of waiting to feel confident before taking action, I’m going to take action to feel confident. Also love the 10% technique—overprepare for the critical 10% part for a confidence boost early on.
-
Python, Go, Rust, TypeScript and AI with Armin Ronacher
Choice of a language doesn’t always have to be based on performance or familiarity. Often it’s influenced by problem space (aka the core task) - suitability and ecosystem. Also, it’s always intriguing to hear a pragmatic take on AI coding from a once hater of the paradigm. AI agents are great at bootstrapping and doing the legwork, but we’ll always need humans to be fully in control (and for the emotions).