- Pointer
- Posts
- Issue #611
Issue #611
Essential Reading For Engineering Leaders
Tuesday 29th April’s issue is presented by Nango
Nango offers everything you need to connect your AI agent with 400+ external APIs
Hundreds of pre-configured integrations: Quickly ship 2-way integrations with 400+ APIs
Built-in developer tooling: Excellent observability, pre-built OAuth, retries & rate-limit handling, E2E type safety, etc.
AI-native: Integrates directly with OpenAI tool calling and MCP
— Allen Cheung
tl;dr: “I started this ritual about 7–8 years ago when I first started managing managers and wanted a better format for bringing together the more senior, busier, more expensive people on my teams. Across teams and companies, I’ve iterated on the meeting, modifying the format to accommodate my team’s suggestions and what I saw were the gaps in our organization.”
Leadership Management
— Claire Lew
tl;dr: “The core of strategy work is always the same: discovering the critical factors in a situation and designing a way of coordinating and focusing actions to deal with those factors.”Strategy is: (1) Identifying the most critical challenge in a situation. (2) Addressing this challenge with some kind of coherent approach. (3) Having coordinating actions focused on overcoming this challenge."
Leadership Management
— Bri Cho
tl;dr: Unified APIs feel like a cheat code—one API, dozens of integrations. But as your product scales, what was once a shortcut starts slowing you down. Teams hit limits they didn’t expect. Customers ask for things the unified layer can’t support. Eventually, most B2B SaaS companies need a more flexible solution.
Promoted by Nango
Management Tools
— Chris Kiehl
tl;dr: Things I now believe, which past me would've squabbled with: (1) Simple is not given. It takes constant work. (2) There is no pride in managing or understanding complexity. (3) Typed languages are essential on teams with mixed experience levels. (4) Java is a great language because it's boring. (5) REPLs are not useful design tools (though, they are useful exploratory tools). And more.
CareerAdvice ThoughtPiece
“Be curious. Read widely. Try new things. I think a lot of what people call intelligence boils down to curiosity.”
— Phil Eaton
tl;dr: “I’ve been a developer, a manager, a cofounder, and now I'm a developer again. I ran away from each position until being a founder because I felt like I was limited by what I was allowed to do. But I reached an enlightment of sorts during my career progression: that everyone around me was dying for someone to pick things up, for employees to show engagement and agency. We think of our titles as our limits. We're quick to say and believe, "that isn't my job". While in reality titles reflect the minimum expected of us, not the maximum that is open to us.”
CareerAdvice
tl;dr: Stop discovering API contract breaks post merge. Signadot's AI-powered testing automatically detects API integration issues between microservices before merging code. No brittle mocks needed — test with real dependencies in lightweight Kubernetes sandboxes. Signadot’s AI testing agent identifies only meaningful API changes, eliminating false alarms. Join forward-thinking teams and try it for free today.
Promoted by Signadot
Tools Tests API
— Vicki Niu
tl;dr: “As guilds on Discord grow larger with longer histories, more and more of them bump up against Lucene’s MAX_DOC limit of ~2 billion messages. We needed a solution to scale search for these special cases, which we call BFGs, or Big Freaking Guilds. We wanted to retain the performance gains from storing all messages for a given guild on the same Elasticsearch shard, since that still works for the vast majority of guilds, but we needed a solution to scale search for BFGs as well.”
Architecture Database
— Sean Goedecke
tl;dr: “The principle here is something like the psychological trick door-to-door evangelists use on new converts - encouraging them to knock on doors knowing that many people will be rude, driving the converts back into the comforting arms of the church. It’s even possible to imagine AI models deliberately doing this exact thing: setting users up for failure in the real world in order to optimize time spent chatting to the model.”
AI ThoughtPiece
— Erik Steiger
tl;dr: “Generating 1 million PDFs in 10 minutes is no joke. That’s 1,667 PDFs per second, or ~0.6ms per PDF. With common PDF generators that take around 1 second each, we would need 11.5 compute days. Or 10 minutes times 1667 vCPUs—assuming they scale nicely.”
Most Popular From Last Issue
How I Use LLMs As A Staff Engineer — Sean Goedecke
Notable Links
Graphiti: Build real-time knowledge graphs for AI agents.
Hatchet: Run background tasks at scale.
Magnitude: OS AI-native testing framework for web apps.
Repomix: Pack your codebase into AI-friendly formats.
Tilt: Define your dev environment as code.
How did you like this issue of Pointer?1 = Didn't enjoy it all // 5 = Really enjoyed it |