Today’s issue of Hungry Minds is brought to you by:

Happy Monday! ☀️

Welcome to the 245 new hungry minds who have joined us since last Monday!

If you aren't subscribed yet, join smart, curious, and hungry folks by subscribing here.

📚 Software Engineering Articles

🗞️ Tech and AI Trends

👨🏻‍💻 Coding Tip

  • Use PerformanceObserver to track Core Web Vitals and catch early bottlenecks

Time-to-digest: 5 minutes

Big thanks to our partners for keeping this newsletter free.

If you have a second, clicking the ad below helps us a ton—and who knows, you might find something you love. 💚

  • GitLab co-founder. 400k downloads in 6 months. Retweeted by Elon. Developers are choosing the open-source path.

  • 400+ models, self-hosted, transparent pricing, centralized team billing. Zero vendor lock-in. Real control, no compromises.

  • 20+ releases in 30 days. This is actual shipping, not marketing theater. The agents (Orchestrator, Architect, Coder, Debugger) actually work together.

  • Lives in VS Code without forcing a new workflow. Your code stays local. Your preferences stay intact. Your teams stay aligned.

  • Built for developers managing real codebases who are done with vibe coding. Built by developers who actually understand what that means.

Datadog faced the challenge of building a metrics platform that could handle billions of data points per second while maintaining sub-second query response times. Their solution? A custom database engine called Monocle, built from scratch in Rust, that revolutionized their real-time metrics storage.

The challenge: Design a system that can ingest billions of metrics per second while serving queries in milliseconds and keeping operational costs manageable at scale.

Implementation highlights:

  • Kafka-centric architecture: Used as central write-ahead log, eliminating node-to-node communication

  • Thread-per-core design: Each CPU core operates in isolation with dedicated workers and no shared state

  • Custom LSM storage: Optimized for write-heavy workloads with arena allocation and time-based pruning

  • Two-tier data model: Separate index database for tags and RTDB for time series data

  • Smart query scheduling: Implemented CoDel algorithm to prevent system overload during query bursts

Results and learnings:

  • Massive throughput: Successfully processes billions of metrics per second

  • Sub-second queries: 99% of queries complete in milliseconds

  • Cost-effective: Achieved 10x better resource utilization compared to previous RocksDB solution

The success of Monocle demonstrates that sometimes building a custom solution, while challenging, can deliver exponentially better results than adapting existing tools. It's like they say in engineering: if you want something done right (and really, really fast), sometimes you have to build it yourself!

ARTICLE (typescript's revenge)
TypeScript's rise in the AI era

GITHUB REPO (transaction-ninja)
DBOS Transact

ARTICLE (framework-graveyard)
Dead framework theory

ARTICLE (good-vibes-only)
Vibe coding needs a spec, too

ESSENTIAL (list-master-101)
Linked List: Explained in 4 minutes

Want to reach 200,000+ engineers?

Let’s work together! Whether it’s your product, service, or event, we’d love to help you connect with this awesome community.

Brief: Turing Award winner and Meta's chief AI scientist Yann LeCun plans to depart to launch his own startup focused on world models, amid Meta's broader AI reorganization and shift towards superintelligence development.

Brief: FFmpeg, a critical open-source multimedia framework used by major tech companies, sparks debate after telling Google to either help fund the project or stop using AI to report bugs, highlighting the growing tension between volunteer-maintained projects and corporate demands.

Brief: Former Apple creative director reveals how Tim Cook's decision to introduce ads in Apple Maps contrasts with Steve Jobs' past rejection of OS advertising, highlighting a shift from prioritizing user experience to pursuing new revenue streams.

Brief: Moonshot AI releases Kimi K2, a 1 trillion parameter AI model with 32 billion active parameters that achieves SOTA performance in coding and math tasks while featuring innovative agentic capabilities that allow it to autonomously use tools and execute complex tasks.

Brief: OpenAI releases GPT-5.1 with warmer, more conversational responses after CEO Sam Altman admitted GPT-5 was made "pretty restrictive" for safety concerns, marking a shift from cautious AI guardrails toward treating adult users with more freedom.

This week’s coding challenge:

This week’s tip:

Use PerformanceObserver with detailed buffering to catch performance bottlenecks that occur before your main script loads. The observer can capture pre-existing entries and provide granular timing data for navigation, resources, and user interactions.

Wen?

  • SPA performance monitoring: Track Core Web Vitals during client-side route transitions

  • E-commerce optimization: Monitor LCP shifts when product images load dynamically

  • Dashboard applications: Detect layout thrashing during data visualization updates

"Don't judge each day by the harvest you reap but by the seeds that you plant."
Robert Louis Stevenson

That’s it for today! ☀️

Enjoyed this issue? Send it to your friends here to sign up, or share it on Twitter!

If you want to submit a section to the newsletter or tell us what you think about today’s issue, reply to this email or DM me on Twitter! 🐦

Thanks for spending part of your Monday morning with Hungry Minds.
See you in a week — Alex.

Icons by Icons8.

*I may earn a commission if you get a subscription through the links marked with “aff.” (at no extra cost to you).